mirror of
https://github.com/logos-co/logos-cpp-sdk.git
synced 2026-08-27 15:51:10 +00:00
master
61
Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
 Dario LipicarandClaude Opus 5
|
937f17ed01 |
feat(cdylib): define logos_module_accept_inbound_token (#151)
* feat(cdylib): define logos_module_accept_inbound_token logos-protocol only DECLARES the module-impl C ABI; every language backend owes each definition. A missing one links clean and dies at dlopen, on Linux only — macOS links plugins -undefined dynamic_lookup and hides it entirely. Also fixes a misrouted diagnostic in the ABI check. The at-nextmaj MAJOR probe ran before the export diff, so an export that NOTHING defines was reported as "a version guard testing MINOR without MAJOR" — the wrong lesson, sending the reader to fix a guard that is not there. The probe is now gated on the symbol being present at the current MINOR, with a self-test over synthetic sets because both diagnostics are inline shell and otherwise untestable. Requires logos-protocol fix/token-direction-key-namespace (59b27ef). Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> * chore(deps): relock logos-protocol to 42460e5b (0.8.0), which is what makes the ABI check non-vacuous WHAT MOVED logos-protocol 480f40ff (0.5.0) -> 42460e5b (0.8.0) Nothing else in the lock changed. The input tracks a branch rather than a rev, so the update moved it on its own. WHY IT HAD TO This PR's definition of logos_module_accept_inbound_token is guarded on protocol >= 0.8, so at the locked 0.5 the emitter wrote NOTHING and checks.<sys>.module-impl-abi passed with the feature entirely absent. That check diffs the emitted export set against the list the PINNED logos-protocol declares, and 0.5 declares ten exports, none of them the inbound door. The PR was green because the check could not see the thing the PR adds. Measured rather than argued. Deleting the whole emitter block from cpp-generator/experimental/lidl_gen_cdylib.cpp: at 480f40ff (0.5.0) -- GREEN, i.e. the check was vacuous logos-protocol 0.5.0 declares 10 module-impl exports [A: --from-header, header-first] defines all 10 declared module-impl exports. /nix/store/mlqs29va7lx4m49cclni6pf01yg7n2js-logos-cpp-sdk-module-impl-abi-tests at 42460e5b (0.8.0) -- RED, naming the missing export FAIL: [A: --from-header, header-first] does not define every module-impl C ABI export. DECLARED by logos-protocol but NOT DEFINED by this backend: - logos_module_accept_inbound_token (version probe: 7 exports at MINOR=0, 11 at MINOR=8 -- one short) Restoring the block returns the check to green at the SAME store path it had before the deletion, so the red is attributable to the emitter and to nothing else in the tree. CHECKS Every check the flake exposes, built individually on x86_64-linux at the new lock. Substituters restricted to cache.nixos.org because cache.nix.logos.co is returning 502, so these are builds rather than cache hits. checks.x86_64-linux.generator-cli /nix/store/ndimz2vnkrqlms1pl6bi0jhci8snzh82-logos-cpp-sdk-generator-cli-tests checks.x86_64-linux.module-impl-abi /nix/store/dncpnlql0jlk1yhzygnfchcjrxh1ndf7-logos-cpp-sdk-module-impl-abi-tests checks.x86_64-linux.tests /nix/store/za8digicp97vd1aqis3x5ypnbwbnqz5r-logos-cpp-sdk-tests-0.2.0 module-impl-abi now reports, for all four generator configurations (--from-header, --lidl, zero-method, records+events): logos-protocol 0.8.0 declares 12 module-impl exports; resolving generated code at LOGOS_PROTOCOL_VERSION_MINOR=8 version probe: 7 exports at MINOR=0, 12 at MINOR=8 defines all 12 declared module-impl exports. The Darwin checks were not built; no macOS builder was available. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5 <noreply@anthropic.com> |
||
|
Dario LipicarandClaude Opus 5
|
692af756cc |
fix: bound the argument count above; and an lp reply it cannot read (#150)
* fix(cdylib): bound the argument count above, not just below
An EXTRA argument was dropped and the call succeeded. The generated dispatch
guarded `args.size() < minArgs` and nothing bounded the other direction, on
every method at every arity -- measured on both providers of two different
contracts, through three consumer surfaces, as conformance case family
`failure/B/arity/too-many`.
Arity is the one part of a contract a caller cannot verify for itself. A
method that gains or loses a parameter upstream answered a stale caller with a
plausible value instead of a refusal, and the caller had no way to tell which
contract it had just talked to.
Two arms, not one, which is why the conformance table registered them as two
defects:
* The ordinary arm gets `args.size() > maxArgs`, where maxArgs is the
DECLARED parameter count rather than the required one -- bounding at
minArgs would reject a caller who legitimately supplies a trailing
optional. When the two coincide the message keeps the exact-count wording;
when they differ it says `at most`, because claiming a count the method
does not require would be wrong in the other direction.
* A ZERO-parameter method had no gate at all -- not the same guard with
minArgs = 0, a different code path, since `args.size() < 0` is unsigned and
was skipped as dead. The upper bound is emitted unconditionally, so it lands
here too.
The guard sits ABOVE the `md.derived` branch, so the generated identity
dispatch inherits it: `version("junk")` used to answer "1.0.0" with status ok,
which is worse than answering nothing -- a correct-looking reply to a call that
should have been refused.
Two existing tests asserted the old behaviour and are updated, not deleted:
`WrongArgumentCountReportsInvalidArgs` forbade any `args.size() >` in the
output, and `ZeroArgumentMethodEmitsNoArityGate` asserted a zero-arg method
emitted no `invalid_args` at all. That second assertion WAS the defect, written
down as a guarantee. Both now pin the bound, and two cases are added for the
arms they did not reach: a trailing optional widening the upper bound, and the
derived identity dispatch.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* fix(lp): a reply it cannot read must not become a provider refusal
The lp half of the same defect logos-qt-sdk fixes on the Qt consumer surface.
`jsonToStdResult` returned a default-constructed StdLogosResult for any
non-object reply, and that default -- `success = false`, empty error -- is
byte-for-byte what a provider sends when it REFUSES a call.
Every other lenient decode in this file bottoms out at a value no provider
means as an answer: "" for a string, 0 for a number, {} for a map. This one
did not. So "I could not read this reply" and "you were rejected" were the same
StdLogosResult, and no caller could separate them.
Fixed here as well as on the Qt side deliberately, and in the same change:
logos-qt-sdk's bare_scalar_slots_stay_lenient warns that tightening a scalar
decode on one surface without the other makes them diverge, and it is right.
The bare scalars stay lenient on both.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
|
||
|
Dario LipicarandClaude Opus 5
|
acea0d24e2 |
feat(codegen): a lossless Qt type mapping, and LIDL types in getMethods (#149)
* feat(codegen): a lossless Qt type mapping — typed containers and optionals
`lidlTypeToQt` answered four different LIDL types with one Qt name. `[uint]`,
`[bstr]`, `[[uint]]` and `[any]` were all QVariantList; `{tstr: uint}` and
`{tstr: any}` were both QVariantMap; every `?T` was a bare QVariant. A Qt
consumer therefore lost, on the SAME contract, types that the std consumer next
door kept — it could not tell `?tstr` from `?uint`, and got no compile-time
check on any element.
The table is now recursive:
[T] QList<qtOf(T)> ([tstr] stays QStringList)
{tstr: V} QMap<QString, qtOf(V)>
?T std::optional<qtOf(T)> (through optionalValueType,
so ??T stays two-state)
any QVariant — KEPT, deliberately
`any` is the one row that must not widen: QVariant is the only Qt type that
holds bytes AND an exact uint64 AND arbitrary nesting at once, so every
narrower spelling would lose what it was chosen to carry. The rule is applied
at the LEAF, so anything whose element type bottoms out at `any` keeps the
QVariant-family spelling at every depth — `[any]` is QVariantList, `[[any]]`
still is, `{tstr: [any]}` is QVariantMap, `?any` is QVariant.
THE TRAP, and why this is not just a rename. A widened name must never reach
QVariant::fromValue / qvariant_cast / logos::qt::toWire as a WHOLE value.
logos-protocol's qvariantToNlohmann matches a CLOSED userType() set:
QList<qulonglong> is in none of it, so it serialises to JSON null. The decode
fails just as quietly — qvariant_cast<QList<qulonglong>> of a QVariantList
yields an EMPTY list. Neither direction warns. So every widened slot is encoded
and decoded by a generator-emitted ELEMENT LOOP, the shape the record cases
already used, and `lidlQtNeedsElementLoop` is the single predicate that decides
which slots need one.
The emitted loops take their source as a lambda PARAMETER, not a body-local
binding. They nest (`[[uint]]`), every level wants the same short names, and a
local — or a range-for over a name the loop itself declares — is then
self-referential: it compiles and reads uninitialised memory. Measured: three
round-trip tests died on SIGTRAP before the argument form.
THE STRING-KEYED EMITTER IS FROZEN, ON PURPOSE. generator_lib is keyed on flat
type NAMES (lidl_to_json flattens the contract before it gets there, because
that emitter also serves the metaobject-introspection path), so it cannot
derive the levels an element loop needs without parsing C++ type names back
into a tree. Every widened spelling is folded back to the name it produced
before (legacyQtBase), which keeps BOTH surfaces it feeds byte-for-byte
unchanged: the legacy Qt consumer, and the Qt-free lp one whose table is
DERIVED from it through mapParamTypeStd. Verified by generating a
28-method contract through both before and after: the diff is empty. The
widened types are spent in the TypeExpr-driven emitters instead
(lidl_gen_client.cpp here, lidl_gen_qt_consumer.cpp in logos-qt-sdk).
Also here, because both are consequences of the table becoming recursive:
* lidlTypeToQt gained a record-name HOOK. A wrapper nests its record structs
in the wrapper class, so a type written outside that scope must qualify
them — and the emitters used to do that by matching the three shapes that
could mention a record on the finished string. `?Point` and
`QList<QList<Point>>` are now spellable, so the qualification happens
during the walk, at the one place that knows a name is a record.
* lidlTypeToLidlText — the LIDL contract spelling of a type. Unused here; the
commit that follows puts getMethods() on it.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* feat(cdylib): getMethods publishes the LIDL contract vocabulary, not Qt names
A module's published metadata — `returnType`, `parameters[].type`, `signature`
— answered in Qt type names. Two things wrong with that, and the second is the
one that matters:
* a cdylib module is Qt-FREE. It described itself in the types of a language
it does not use, to readers (`lm`, logoscore's method listing, basecamp's
module inspector) that are showing a human what the module offers.
* it was LOSSY. `[uint]`, `[bstr]` and `[any]` are three different LIDL types
and all three published as the single word QVariantList, so the listing
could not be read back as a contract. That is now `[uint]`, `[bstr]`,
`[any]`; `{tstr: uint}`; `? tstr`; and a record publishes its declared
name.
WHY THIS IS SAFE — checked, not assumed. The historical objection is recorded
in the function this replaces: these strings are read as METATYPES, and
emitting a record's struct name here once made the host SIGSEGV. Nothing in the
current runtime does that. logos-plugin-qt's QtProviderObject dispatches on
`method.returnMetaType()` / `parameterMetaType(i)` — the QMetaObject, never
this JSON — and every remaining reader treats these fields as opaque text:
logos-module's `lm` prints them, logoscore's output.cpp prints them, basecamp's
CoreModuleManager forwards the JSON to QML, and the plain wire's json_mapping
only round-trips them. Nothing anywhere builds a QMetaObject from this
metadata.
The spelling comes from lidlTypeToLidlText, which mirrors logos-lidl's
serializeTypeExpr. It is a COPY, because that function is file-local to
logos-lidl's serializer.cpp and the public headers expose no type printer —
so instead of hoping, the pairing is ASSERTED: the test round-trips each shape
through `lidl::serialize` and reads the type text back out of the emitted
`.lidl`. When logos-lidl exports a printer, delete the copy and call it.
Not fixed by this, and not attempted: the Rust SDK's provider generator has its
own `qt_type_name` writing the same JSON, so the two languages now disagree
about how a module describes itself. That is a cross-repo change.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* test(doctests): the generator round-trip pins the lossless Qt spellings
`cpp-sdk-generator-roundtrip.test.yaml` is a CI gate
(.github/workflows/doctests.yml), and two of its `expect_contains` were
pinned to the type names the Qt consumer produced BEFORE the lossless
mapping:
QStringList labels(const QVariantList& ids
QVariant nearest(const Point& p, QVariant limit
The generator now emits `QList<qulonglong>` and
`std::optional<Point>` / `std::optional<qulonglong>` for those slots, so both
assertions failed. The `nearest` step's `run` grep was pinned the same way
(`QVariant nearest`), so the line it was supposed to assert on was not even
in the output being searched.
Verified by running the spec's own steps against the generator built from
this commit: 10 run-steps, 0 failures. The `[uint]` -> QList<qulonglong> and
`?T` -> std::optional<T> lines were read out of the real
`consumer/sensor_module_api.h` and `geometry/geometry_module_api.h`, not
written from the mapping table.
Prose too, in three places that described the old table: the Flow-3 type
mapping ("other arrays -> QVariantList"), the composite-types intro
("optionals ... stay QVariantMap / QVariant"), and the composite-signature
step. They now say what the mapping actually is — one LIDL type, one C++
spelling, with `any` the single deliberate exception — and `nearest` is
called out as the one signature carrying both halves of the optional
mapping.
`doctests/outputs/cpp-sdk-generator-roundtrip.md` carries the same prose
corrections. That tree is hand-pinned and CI never diffs it, which is
exactly why it must be corrected by hand.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* fix(generator): the consumer wrapper comes from the contract, not from getMethods
`logos-cpp-generator <plugin> --module-only` — the invocation
logos-plugin-qt's generate-module-headers.sh makes for every module's lp
wrapper — built that wrapper's whole type surface out of the plugin's
PUBLISHED `getMethods()` metadata. It now builds it out of the module's `.lidl`
contract, the file the same invocation already passes as `--events-from`.
WHY THIS IS A DEFECT AND NOT A PREFERENCE. generator_lib is keyed on flat type
NAMES, and mapParamType / mapReturnType fall back to QVariant for a name they
do not recognise (generator_lib.cpp:142 and :153). So the wrapper's types
depend on the VOCABULARY a module happens to publish its metadata in, and a
vocabulary this emitter has no row for degrades to QVariant — LogosMap on the
lp surface — with no diagnostic at any layer. It is a machine reader of a
listing that every other consumer treats as human-facing text, and it fails
silently.
It was measured, not theorised. 621772a made the cdylib backend publish the
LIDL contract vocabulary (`tstr`, `[uint]`, `result`, `? tstr`) in place of Qt
type names, because that listing is what `lm`, logoscore and basecamp show a
human and Qt names are the wrong answer for a Qt-free module. Every
`interface: "universal"` module's lp wrapper collapsed:
logos-test-modules' `checks.unit-tests-new-api` went PASS -> FAIL, and the
compiler said exactly why —
error: no viable conversion from 'LogosMap' to 'StdLogosResult'
StdLogosResult r = modules().test_basic_module.resultWithMap();
`result` is not a name mapReturnType knows, so it became QVariant, so it became
LogosMap. Bisected to exactly 621772a (5ffd90b passes, dd52d9d fails).
THE FIX IS TO STOP READING THAT VOCABULARY, not to learn a second one.
`int` means a 32-bit Qt int in one table and a 64-bit LIDL integer in the
other, and the reader cannot tell from the string which table it is holding —
a merged table would silently mistype every integer on every module. The
contract has no such ambiguity: it is a TypeExpr tree, and lidl_to_json is the
single place it is flattened. Taking methods from it makes this path emit the
same wrapper as `--general-only --dep <name>=<name>.lidl`, which is what
buildHeaders.nix already runs under cross-compilation and for the entire Qt
surface. Contract-first, on every platform, for every surface.
WHAT CHANGED, exactly:
* loadEventsFromLidl -> loadContractFromLidl. It already parsed the whole
contract and threw the methods away; it now returns them, after the same
lidlCheckRecords + lidlInjectIdentity + noteOptionalPositionalSlots that
main.cpp's --dep path applies. Identity is injected rather than read,
matching the provider side (main.cpp's --backend cdylib), so the two cannot
disagree about name() / version().
* A sidecar that is NAMED BUT MISSING is now refused (exit 2), and an
unreadable or malformed one is fatal (exit 4). Both used to be shrugged off
— which shipped a wrapper with no typed events, and would now ship one with
no typed methods, in the silently-empty shape generate-module-headers.sh
exists to refuse.
* The plugin is STILL LOADED. That load is the dlopen check this path
performs (exit 3 on an SDK/ABI skew) and it is unchanged; what the plugin
says about itself is now compared against the contract instead of believed,
and a divergence — a stale sidecar — is reported by name on stderr. Only
`isInvokable` entries are compared: a cdylib publishes its events into the
same array, tagged `"type": "event"`, and both emitters already skip those.
* A module with NO contract keeps introspection — a handcrafted Qt plugin's
QMetaObject is still the only description of its API that exists, and Qt
type names are the right vocabulary for it — but a listing spelled in the
LIDL vocabulary with no contract to go with it is now REFUSED (exit 7)
instead of silently producing the untyped wrapper. That combination is only
reachable by hand: buildHeaders.nix always passes the flag when the sidecar
exists, and it is the shape the developer guide used to suggest. The two
vocabularies are not distinguishable in general, which is the whole reason
this emitter must read only one — but they do not have to be: the words
they share (`int`, `bool`) are all in the known table and never reach the
fallback, so the check keys on the LIDL half Qt has no word for at all
(`tstr`, `bstr`, `uint`, `float64`, `result`, `any`, and anything starting
`[`, `{` or `?`). No Qt type is spelled that way, so it cannot false-fire;
a false negative is just the old behaviour.
THE ENUMERATION, because two previous ones missed this reader. Searching for
who greps `returnType` is what missed it; the question is what the data FLOWS
INTO. Every consumer of a published getMethods array in the workspace:
MACHINE (one, and it is this one)
logos-cpp-sdk cpp-generator/plugin_introspect.cpp, reached only through
logos-plugin-qt's generate-module-headers.sh / buildHeaders.nix.
HUMAN-READABLE OR OPAQUE PASSTHROUGH (all of them)
logos-module's `lm` (prints; --json re-emits verbatim), logoscore-cli's
client/output.cpp (prints) and core_service_dispatch.cpp (forwards),
logos-logoscore-tui (formats one line per method), logos-module-viewer
(reads the QMetaObject directly, not this JSON), basecamp's
CoreModuleManager / MainUIBackend (hands the JSON string to QML),
logos-protocol's json_mapping.cpp and qvariant_rpc_value.cpp (round-trip
the strings unread).
PRODUCERS, for completeness: lidl_gen_cdylib.cpp (LIDL vocabulary),
logos-plugin-qt's QtProviderObject (Qt names, from the QMetaObject) and
lidl_gen_cdylib_glue.cpp (forwards the cdylib's), logos-rust-sdk's
rustgen_provider.rs (still Qt names — the two languages disagree, as
621772a noted), and logos-protocol's ModuleProxy, which appends derived
name()/version() entries spelled `QString`. None of that reaches a type
decision any more, which is the point of the change.
Build-system paths checked and clear: `<plugin> --module-only` is invoked
from exactly one place in the workspace (generate-module-headers.sh:60);
LogosModule.cmake, buildPlugin.nix and mkLogosModuleTests.nix all use
`--general-only`, which is contract-driven already; the doctests' `--lidl
--module-only` is a different mode entirely.
VERIFIED.
`nix build path:./repos/logos-test-modules#checks.aarch64-darwin.unit-tests-new-api`
with this SDK overridden in (plus the logos-lidl overrides the branch needs at
the qt-sdk and plugin-qt nodes) — 32 passed, 0 failed. The same command against
this branch's HEAD fails to compile, as above. The build log shows the path
taken, per module:
Detected new-API plugin (LogosProviderPlugin), using getMethods() — 43 methods
Using the module's LIDL contract for the method surface — 41 methods
(the plugin's published listing is a description, not a type source)
The refusal, measured by hand against a real LIDL-publishing plugin
(test_basic_module, built from this branch) because no check exercises a
hand-run invocation:
no --events-from -> exit 7, nothing written, the message above naming
8 offending slots
with --events-from -> exit 0, 41 typed methods, 69 `std::string` in the
emitted lp header
a pre-621772a build of the SAME module (Qt-name listing), no --events-from
-> exit 0, still generates, still typed — the refusal does
not fire on the vocabulary this emitter can read
nix/tests-generator-cli.nix gains the two CLI-surface cases this adds: a
`--events-from` naming a file that does not exist is refused with that
sentence, and — the control that makes it mean something — the same command
with a READABLE contract gets past the flag and fails on the plugin instead. No
plugin is needed for either: the contract is loaded before the plugin is
opened.
logos-cpp-sdk's own checks (tests, generator-cli, module-impl-abi): 334 of 334.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
|
||
|
Dario Gabriel LipicarandClaude Opus 5
|
b505ee95eb |
fix(generator): widen BOTH rejection detectors to a CLOSED SET of codes
The two emitted detectors — logosDispatchRejection (QVariant, Qt surface) and
logosDispatchRejectionJson (nlohmann, lp surface) — each matched the single
literal "dispatch_failed". Providers have been answering a wrong argument COUNT
with "invalid_args" all along: this repo's own cdylib dispatch emits it
(experimental/lidl_gen_cdylib.cpp:805) and so does logos-rust-sdk's
args::invalid_args. Nothing detected it. The refusal therefore arrived as a
VALUE and the return table erased it — `_result.toList()` on that map is `[]`,
`.toString()` is "", `.toLongLong()` is 0 — so a caller could not tell "you sent
me the wrong number of arguments" from "the provider returned nothing".
Measured on the untyped surface, where the erasure is visible:
logosctl call test_basic_module isPositive (missing required argument)
-> exit 0, status:"ok", result {"code":"invalid_args", ...}
The set is now {dispatch_failed, invalid_args, unknown_method}, in ONE
kRejectionCodes array. Both emitters build their condition text from it, so the
Qt and Qt-free twins cannot drift apart — which is what two hand-written copies
of the same literal were always going to do.
"unknown_method" is listed before any provider emits it, on purpose. An unknown
method is currently answered with a bare null, byte-identical to a legitimate
null return (logos-protocol logos_protocol.h says so outright), and closing that
is a provider-contract change across the SDKs. Detectors go first because
widening one is backwards-compatible on its own — nothing emits the code, so
nothing changes — whereas a new provider code shipped against narrow detectors
would arrive at consumers as DATA: the same silent-success bug, freshly minted.
The set stays CLOSED. NOT "any three-key object with a code": a method may
legitimately return a three-string map, and an `any` return certainly can, so a
shape-only match would let user data impersonate a refusal. Every guard above
the compare — exactly three keys, all three present, all three strings — is
untouched.
WHY FIVE COPIES AND NOT ONE. The other four are logos-qt-sdk's byte-identical
lidl_gen_qt_consumer.cpp, logos-rust-sdk's args::as_dispatch_rejection, and
logos-logoscore-cli's core_service/call_envelope.cpp. Two shared homes were
considered, both rejected here and both recorded at kRejectionCodes: a shared
EMITTER in share/lidl-frontend (the channel exists — it already ships
lidl_emit_common to logos-qt-generator) collapses 2 of 5 and turns four
independently landable fixes into an ordered stack; a runtime predicate in
logos-protocol is the principled end state, and is how the analogous CONVERSION
duplication was actually solved, but it trades a text-level duplication for a
build-level version coupling — the emitted body is self-contained today, so a
wrapper compiles against whatever protocol its module pins. Each repo instead
holds the vocabulary in one named place, so a drift is visible.
Tests: each code asserted present in the emitted condition on both surfaces,
plus the negatives that keep the match closed — exactly three comparisons and
no more, the shape guards still emitted, and the compare still ahead of
`return true`. 316/316 ctest.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
|
||
|
Dario Gabriel LipicarandClaude Opus 5
|
dbe1d63677 |
feat(abi): define logos_module_set_call_caller, gated on protocol 0.6
The definition lands BEFORE the protocol declares the export. logos-protocol only DECLARES the module-impl C ABI and every backend owes the definition; that gap shipped twice, each time as an undefined symbol at dlopen, on Linux only, invisible on macOS. Declaring first would turn this repo, logos-rust-sdk and logos-module-builder red the night the bump merged. Defining first costs nothing: the guard is MAJOR-aware >= 0.6 and the current pin is 0.5, so nothing is emitted today and the ABI check sees declared == defined. This is the case #146 made possible. The next-MAJOR probe resolves the emitter at MAJOR+1, where a >= 6 guard IS true, so the emitted set there is legitimately a SUPERSET of the declared one. The probe used to demand equality and would have rejected this outright. cpp/logos_caller.h carries the LogosCaller type (std-typed, Qt-free) and logos::currentCaller(), reading a thread-local stack the generated export pushes to. Two things the audit corrected, both worth reading: * A present-but-unreadable `instance` is DROPPED and the module still identified. This backend already did that; Rust returned Unknown, and each had a passing test pinning its own answer, so neither suite could see the divergence. The protocol header now states the rule normatively and Rust is aligned to it. * The accessors are explicitly HIDDEN on ELF. The header argued this state must not be unified across images and then relied on being inline to achieve it — which is false: a function-local static in an inline function emits STB_GNU_UNIQUE at default visibility and the loader collapses every image's copy into one, even under RTLD_LOCAL. Measured across two dlopen'd images: default visibility let a push in A be read by B; hidden restored isolation. logos-module-builder sets no visibility anywhere, so real plugins were built the first way. An anonymous namespace would be worse — vague linkage is load-bearing WITHIN an image, since the generated TU pushes and the author's TU reads. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> |
||
|
Dario Gabriel LipicarandClaude Opus 5
|
51abc407df |
fix(cdylib): guard the conditional exports on MAJOR too, not MINOR alone
The three version guards this emitter writes tested
LOGOS_PROTOCOL_VERSION_MINOR and ignored LOGOS_PROTOCOL_VERSION_MAJOR. At
protocol 1.0.0 the MINOR resets to 0, so all three would have gone false
and the C++ backend would have dropped from ten module-impl exports to
seven — losing grant_host_services (0.3) and the teardown pair (0.5).
What makes this worth catching early is that it is NOT a link error and
NOT a dlopen failure. logos-plugin-qt's glue guards the matching CALLS the
same way, so the definitions and the calls disappear together: everything
builds, everything loads, and modules simply stop having teardown and stop
being grantable. There is no diagnostic anywhere. logos-rust-sdk would
meanwhile keep emitting all ten, because it compares the (major, minor)
tuple — so the two backends would silently disagree.
The arithmetic is emitted EXPANDED rather than behind a function-like
macro, because the generated sources are resolved by unifdef in the check
below, and unifdef evaluates nested integer arithmetic but silently
no-ops on what it cannot parse — which would be a green check over
unresolved text.
checks.<sys>.module-impl-abi gains the probe that would have caught it: it
already resolved the generated sources at the real MINOR and at MINOR=0,
and now also at one MAJOR up, where the full declared export set must
still be emitted. Nothing pins that to 1 in particular — it is "one past
whatever we are on", which is exactly where a MINOR-only guard breaks, and
it is the only probe here that does not resolve at the current major.
Proven to fail: reverting only the guard change turns it red naming all
three affected exports.
FAIL: [A: --from-header, header-first] the export set is not complete
at protocol 1.0.
DECLARED but NOT emitted once the MAJOR advances:
- logos_module_about_to_unload
- logos_module_grant_host_services
- logos_module_set_unload_done_callback
The MAJOR is read from the same logos-protocol output as the MINOR and the
export list, so no version literal enters this check.
logos-plugin-qt carries the same defect in its two glue guards; that is a
separate PR, and until both land the two sides still agree at every
version that exists today.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
|
||
|
Dario LipicarandClaude Opus 5
|
667990f28c |
feat(sdk): aboutToUnload — a module's chance to finish before teardown (#143)
* feat(sdk): aboutToUnload — a module's chance to finish before teardown
A module could not flush state or close handles on the way out: the host
stopped it, its destructors ran, and anything mid-flight was gone. This gives
LogosModuleContext the Qt Creator contract for exactly that problem.
enum class LogosShutdown { Synchronous, Asynchronous };
virtual LogosShutdown aboutToUnload() { return LogosShutdown::Synchronous; }
void unloadFinished() const; // any thread
Default Synchronous, so no existing module changes behaviour. A module with work
to finish returns Asynchronous and calls unloadFinished() when done; the host
waits, but only for a bounded grace period.
unloadFinished() is a NO-OP outside a framework context, and after the deadline
has passed. That matters more than it reads: a module needs no special case for
being torn down under a deadline it already missed.
Two SFINAE pairs mirror the existing maybeSet* helpers. maybeAboutToUnload
reports Synchronous for an impl that never inherited LogosModuleContext, which
is exactly right -- it has no hook, so there is nothing to wait for.
Both names join the reserved set beside onContextReady: an impl overriding
aboutToUnload is talking to the framework, not publishing API, and leaking
either would generate a consumer wrapper for a lifecycle hook (LogosShutdown
has no LIDL type to return anyway).
The cdylib backend emits the two optional C ABI exports. The completion
callback is installed BEFORE the impl is asked to unload, and that ordering is
the correctness of the whole async path: an impl that finishes INLINE would
otherwise signal into a slot that is still empty, and the host would wait out
its entire grace period for a module already done. There is a test for it,
because nothing about reading the code makes that failure visible.
294/294, 4 new. Requires logos-protocol#62; flake.lock pins that branch.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* chore(deps): track logos-protocol master now that the teardown ABI has landed
logos-co/logos-protocol#62 merged as 9664ae2. The lock pointed at the PR branch
while it was open.
The narHash is unchanged across the move (sha256-JTREoJn2kjQmYyYHg2RQb4...), so
the merged tree is byte-identical to the branch this was built and tested
against.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* fix(generator): guard the teardown emission on protocol 0.5
The emitted exports named logos_module_unload_done_cb unconditionally, so a
module built with this generator against an older logos-protocol failed to
compile on a typedef it never asked for:
error: 'logos_module_unload_done_cb' was not declared in this scope
in generated code the author never wrote and cannot see. That is what this PR's
doc-tests hit -- new generator, older protocol pin.
logos-protocol#63 gives the surface a MINOR (0.5) so it is detectable, and both
the statics and the exports now sit behind
#if defined(LOGOS_PROTOCOL_VERSION_MINOR) && LOGOS_PROTOCOL_VERSION_MINOR >= 5
the same way the 0.3 trust-root surface is guarded a few lines below. A module
built against 0.4 simply has no teardown entry point, which is the same state as
a module that never overrode the hook -- and the glue that would call it is
generated alongside, so nothing goes looking for the missing symbol.
Both halves need the guard, not just the exports: the typedef is what an older
header lacks, and it is the statics that name it. The test asserts both.
295/295. flake.lock tracks protocol master (0d2a3c0), where 0.5 landed.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
|
||
|
Dario LipicarandClaude Opus 5
|
9c05d4ad11 |
feat(lp): emit the AsyncResult twin, and fix the LpClient create race (#142)
* feat(lp): emit the AsyncResult twin, and fix the LpClient create race
Two related changes to the Qt-free (ApiStyle::Lp) consumer surface.
1. logos::LpClient::ensure() published its lazily-created lp_client through a
plain pointer with no synchronization. Two threads reach a dep's FIRST call
concurrently more often than the lazy-init shape suggests: a
concurrency:"multi" module dispatches handlers on concurrent QThreads, and
any module running a worker of its own (an HTTP handler, a chain-sync pump)
races that worker against the dispatch thread. So this was a data race, and
it leaked whichever client lost.
A mutex around the body is the obvious fix and the wrong one: for a Qt-affine
transport lp_client_create marshals construction onto the Qt main thread and
BLOCKS there, so a worker holding the lock would wait for the main thread
while the main thread, reaching the same ensure() from an inbound call, waits
for the lock — trading a data race for a deadlock. Construct outside any lock
and publish with a CAS instead; the loser destroys its own client, which
lp_client_destroy permits from any thread. A failed create is not latched.
2. `<name>AsyncResult` is now emitted for the lp surface, matching the Qt one.
It was withheld for a reason that belonged to the transport rather than the
emitter: lp_invoke_async used to hard-code `cb(1, ...)`, so an AsyncResult
over it would have reported ok() for a call to a module that was not even
loaded — an error channel that lies is worse than none. logos-protocol#40
fixed that, and the new logos::LpClient::invokeAsyncResult surfaces the
failure in C++.
The generated twin also folds a provider REJECTION: a provider that ran and
refused answers {"code": "dispatch_failed", ...} as its RESULT, not as a
transport error, so the decode would otherwise erase it into a default value.
The lp SYNC path folds it too now, as the Qt sync path already did — with no
qWarning fallback, since a Qt-free wrapper has no logger to fall back to.
lp `<name>Async` is deliberately unchanged: its callback takes the value
alone, exactly as on the Qt side.
Also unblocks logos-qt-sdk's LpBridge::invokeAsyncResult, which keeps a private
second lp_client only because logos::LpClient had no error-carrying async.
Verified: sdk tests 309/309 (incl. 9 new LpClient and 10 new generator tests),
generator-cli green, and the generated wrapper compiles under -Wall -Wextra
-Werror. The concurrency test was checked against the pre-fix header as a
control: 8 threads, 8 clients created, 7 leaked, threads disagreeing on which
client was the module's.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* fix(tests): stub lp_string_free, and keep the path that needs it live
The sdk_tests link failed on GCC/Linux with an undefined reference to
lp_string_free from logos::LpClient::getMethods(), while linking cleanly on
macOS/clang.
The stub set was genuinely missing lp_string_free. It went unnoticed because the
lp_get_methods stub returned NULL, which made getMethods()'s free call
unreachable: lp_get_methods is a non-inline extern "C" function DEFINED IN THE
SAME TU as the test, so clang may inline it, prove the pointer null and delete
the call — no reference, no link error. GCC kept the call, and the linker wanted
the symbol.
Adding the stub alone would fix the link and leave the trap: the free path would
still be dead, so the next compiler that keeps the call decides whether this
builds. So lp_get_methods now returns a real heap allocation, which is the ABI's
actual contract ("every char* RETURNED by this library is owned by the caller;
free it with lp_string_free"), and lp_string_free frees it and counts. The
single-threaded test asserts the count, so the ownership rule is pinned rather
than merely satisfied.
Verified by reproducing the failure on macOS with -O0 (which stops clang folding
the call away): the pre-fix file fails with exactly `"_lp_string_free",
referenced from: logos::LpClient::getMethods()`, and the fixed one links and
runs 9/9. Full check: 308/308.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
|
||
|
Dario LipicarandClaude Opus 5
|
1f69bca54a |
feat(generator): emit the derived module identity methods (#141)
* feat(generator): emit the derived module identity methods
Run logos-lidl's injectIdentityMethods() on every ModuleDecl this generator
emits code from, and give the cdylib dispatch a body for the two methods it
adds.
Injection happens at EMISSION points, never at artifact points:
* generateInterfaceWrappers -- one load point covering both --dep and
--interface, so a consumer sees name()/version() on every dependency and
bound interface;
* --from-header --backend cdylib and --lidl --backend cdylib, so the provider
answers them;
* NOT --header-to-lidl, which writes the published contract.
The distinction is belt-and-braces rather than load-bearing: the injected
methods are `derived` and lidlSerialize omits those, so the .lidl a
--from-header build writes stays byte-identical to what --header-to-lidl writes
for the same header.
The dispatch emits a literal for a derived identity method instead of the usual
lidlImpl().<name>(...) -- the author's impl class has no such member, so
delegating would not compile. The literal is the module's own name and version,
so it cannot drift from the metadata the module was built with. A module that
declares name() itself is not derived and still reaches its impl.
290/290 tests pass, 4 new: that the emitted literal is the module's OWN version
(a test at 1.0.0 could not tell a correct generator from one that fell back),
that identity is listed for introspection as well as dispatched, that an
author's own name() still reaches the impl, and that a versionless declaration
falls back rather than emitting "" -- which would read as a failed call.
Requires logos-lidl#10; flake.lock pins that branch until it merges.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* chore(deps): track logos-lidl master now that the identity pass has landed
logos-co/logos-lidl#10 merged as ae3ffe0. The lock pointed at the PR branch
while it was open; this re-points it at master.
The narHash is unchanged across the move (sha256-WHmisUvYA8DQ2ZxSF2mM2...),
so the merged tree is byte-identical to the branch this was built and tested
against.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
|
||
|
Dario LipicarandClaude Opus 5
|
95d7b3a9c5 |
feat: split the SDK by capability, retire the provider-header path, and harden the cdylib decode (#138)
* feat(generator): remove --provider-header (interface: "provider")
Every provider now goes through the module-impl C ABI, so the LOGOS_METHOD
dispatch path is gone: parseProviderHeader, generateProviderDispatch,
ParsedMethod and the joinDocLines helper only it used (~300 lines), plus the
test that covered it.
`toQVariantConversion` is NOT removed — it is shared with live emitters — and
its test stays.
The flag is REFUSED rather than dropped. Without that, `--provider-header x.h`
falls through to the plugin-path branch, which reads the flag itself as a
plugin path and reports "Plugin file does not exist: --provider-header" — a
missing-file error for what is really a retired mode. It now exits 2 with a
message pointing at interface: "universal".
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* feat(sdk): logos_host_services.h — the C++ veneer over the privileged surface
Phase C1. A Qt-free, header-only wrapper over the three trust-root lp_* calls
A3 added, so capability_module can become an ordinary universal module instead
of a hand-written Qt plugin reaching into TokenManager directly.
Deliberately FREE FUNCTIONS, not a LogosModuleContext seam as the plan
sketched. The grant is process-global per IMAGE (host binary and module cdylib
each link their own logos-protocol, so each has its own grant state and its own
TokenManager), so the gate lives in the caller's own image and there is nothing
per-instance to inject; a context seam would imply the privilege is a property
of one impl object, which it is not. It is also markedly cheaper: a seam would
need a new module-impl C ABI export plus lockstep changes in BOTH codegen paths
(the Qt provider glue and the cdylib wrapper).
constantTimeEquals lives here rather than in each caller: the natural spelling
(a == b) leaks the matching-prefix length through timing, and a trust root
comparing tokens with == is the exact bug this file exists to prevent. Ported
from capability_module's own implementation to std::string.
Two things the tests caught that reading had not:
* lp_inform_module_token_to takes SIX arguments (client, auth_token,
origin_module, module_name, token, timeout_ms), not the three I first wrote.
The wrapper now mirrors it exactly, with the protocol's own default-timeout
semantics documented.
* sdk_tests compiles against logos_headers alone, which carries no protocol
include path. It now resolves logos_protocol.h from LOGOS_PROTOCOL_ROOT,
accepting either the source layout (cpp/) or a package layout (include/) and
failing loudly on neither, rather than hard-coding the one in use today.
The suite deliberately does NOT link logos-protocol: the lp_*-calling wrappers
are `inline` and never ODR-used by these tests, so no protocol symbol is
referenced. That is itself the assertion — the veneer must not drag the
protocol library into a header-only consumer. A future test that calls one will
fail to LINK rather than silently pull it in.
Also documents a real gap found while writing it: lp_token_get performs NO
host-service check, so "token_registry" gates ENUMERATION only. The plan claims
that service covers `lp_token_get(any)`; it does not. Flagged at the call site
rather than papered over — if lookup should be gated, the gate belongs in
lp_token_get.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* feat(generator): emit logos_module_grant_host_services in the cdylib exports
Completes the C1 codegen half. A3 declared this entry point in
logos_module_impl.h and documented that the grant MUST cross the C ABI, but the
generator never emitted it, so nothing could open a module's gates.
The body forwards to lp_grant_host_services in the MODULE's own image, which is
the entire point. Verified that premise on a real built module rather than
taking it from the header comment: test_basic_module_cpp_plugin.dylib DEFINES
25 lp_* symbols and imports zero — logos-protocol is statically linked into
each plugin, so the module's gate state really is its own, and a grant recorded
only in the host would leave lp_token_keys() returning null forever. That
failure is silent: null is indistinguishable from an empty token store.
Emitted unconditionally rather than behind a codegen flag. Which modules are
privileged is the host's decision — it pushes nothing to an ordinary module —
and lp_grant_host_services validates the names and fails closed, so a per-module
flag would only add a second place for declaration and capability to disagree.
The comment states the boundary honestly: this is a declaration-and-audit
mechanism, NOT a defence against a hostile module. The cdylib links
logos-protocol, so its own code can call lp_grant_host_services() directly and
self-grant. What the gate buys is that the privilege is explicit, greppable and
off by default. Isolation between modules rests on process separation, the auth
token, and the target's allowedCallers.
Three tests, one per property that could regress independently: the export
exists; its body actually forwards (a stub returning 0 would make every host
push look successful while both gates stayed shut); and it is emitted for an
ordinary module too, not only for privileged ones.
Confirmed in the built artifact: nm on a real universal module lists
_logos_module_grant_host_services alongside the other seven module-impl exports.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* fix(generator): guard the grant export on the protocol MINOR that added it
The emitted logos_module_grant_host_services calls lp_grant_host_services,
which logos-protocol only gained at MINOR 3. A module built against an older
protocol therefore failed to compile in GENERATED code its author never wrote.
Found by giving logos-template-module a standard flake: its own lock resolves
protocol master, and the build died on `use of undeclared identifier`.
Guarded on LOGOS_PROTOCOL_VERSION_MINOR >= 3. A module built against 0.2 has no
grant entry point at all, which is the same fail-closed state as never being
granted.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* docs: point logos_async_result.h at the one LogosModule.cmake
The comment named "logos-plugin-qt/cmake/LogosModule.cmake and its
module-builder twin". There is no twin any more: logos-plugin-qt's copy is
deleted and the file exists once, in logos-module-builder.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* chore(deps): rev-pin logos-protocol at c8bab12 (the trust-root surface)
logos_host_services.h is a veneer over lp_token_keys /
lp_inform_module_token_to / lp_grant_host_services, which landed on
logos-protocol's feat/per-client-token-store branch and are NOT on its
master — master is still LOGOS_PROTOCOL_VERSION_MINOR 2, so the `tests`
check could not compile against the previously locked 03842db.
Rev-pinned in the URL rather than left master-tracking, because
`nix flake update` cannot reach a commit that is not on the tracked
branch. Re-point at master once that branch merges.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* refactor(generator): remove --module-dir, and the two dead files it documented
--module-dir walked a directory of BUILT plugins and generated one consumer
wrapper per dependency by dlopen'ing each and reading its QMetaObject. Every
wrapper now comes from a contract instead -- `--general-only` with one
`--dep <name>=<name>.lidl` per dependency -- which builds no dependency plugin
and, unlike introspection, works under cross-compilation.
It is REFUSED rather than ignored, mirroring --provider-header right below it.
Falling through to the dependency LISTING would have exited 0 having generated
nothing: the exact shape that lets a stale caller look green while shipping a
module with no typed API.
No nix build changes as a result -- the flag had no caller left in any of them,
so a store-path diff would be empty either way and would prove nothing. The
only observable difference is what the binary does when handed the flag, so
that is what the new `generator-cli` check asserts, by EXIT CODE:
OK: control - --metadata alone exits 0 and lists dependencies
OK: --module-dir exits non-zero (status=2)
OK: --module-dir fails with the removal diagnostic
OK: --general-only still emits the umbrella
The control matters: without it a non-zero exit could equally mean the binary
is broken. It runs against an EXISTING modules directory too, because the old
code only errored when that directory was missing.
Also deleted, both genuinely dead:
* cpp/compile.sh -- compiles logos_api.cpp, module_proxy.cpp, token_manager.cpp
and six headers, NONE of which exist in this repo any more (they moved to
logos-protocol / logos-qt-host in the host split). The script cannot run.
* docs/docs.md -- 789 lines with zero inbound references anywhere in the
workspace, documenting --module-dir and a cpp/ layout that is gone.
cpp-generator/compile.sh is KEPT: logos-module-builder's LogosModule.cmake:404
still invokes it (`add_custom_target(cpp_generator_build ...)`) on the
LOGOS_CPP_SDK_IS_SOURCE branch, and it builds into exactly the
LOGOS_DEPS_ROOT/build/cpp-generator path that file then reads.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* refactor(generator): the umbrella emitter leaves legacy/, and gets its own mode
`cpp-generator/legacy/` held four things and only one was legacy. The shared
emitter library was misfiled there: `generator_lib.{h,cpp}` is already consumed
by the MODERN `experimental/lidl_gen_client.h` and by all 11 tests under
tests/generator/. `lidl_to_json.{h,cpp}` likewise. Both are now
`cpp-generator/`; `legacy/` is down to `main.cpp` + `legacy_main.h`.
The logos_sdk umbrella (`struct LogosModules`) is not legacy either — it is the
CURRENT typed-dependency surface. `LogosModuleContext::modules()` returns it,
so every universal module that calls a declared dependency goes through it, and
LogosModule.cmake runs `--general-only` for every module build. Yet the only
code that could emit it lived inside the directory the plan wants deleted.
So `cpp-generator/main.cpp` gains `--umbrella`, with `--general-only` routed to
the same implementation and dispatched before the fall-through to legacy_main.
The deps-driven emission needed no rewriting: `makeUmbrella{Header,Source}
FromDeps` were already in generator_lib, and legacy/main.cpp merely wrapped
them in file I/O. -352 lines from legacy/main.cpp (827 -> 475), including the
interface-wrapper helpers that only that branch used.
`--general-only` keeps working identically, because LogosModule.cmake and
logos-basecamp both call it. The alias is guarded on `--metadata`, since
`--general-only` was never a standalone mode — without metadata it fell through
and reported the flag as a missing plugin path, and it still does.
The scraping `writeUmbrellaHeader`/`writeUmbrellaSource` are untouched: they
belong to `generateFromPlugin`, the QPluginLoader introspection path, and die
with it.
Verified byte-identical, which is the whole claim of a relocation. An
adversarial pass built its own pre- and post-change binaries and diffed the
emitted `logos_sdk.{h,cpp}` across 12 real metadata.json files x {qt,lp} x
{--general-only,--umbrella}: 48/48 identical, stdout/stderr/exit included, with
a positive control (qt vs lp) confirming the harness can see a difference. Real
modules then built through logos-module-builder against both binaries with
`diff -r` empty, including the compiled plugin. 266/266 tests pass, and the
pre-change tree also reports 266, so no test was silently dropped.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* feat(sdk): split the SDK by capability, and add the host façade
Host, module-consumer and module-provider are three distinct capabilities. The
SDK exposed them as one target, so a program linked all three regardless of
what it was. Each now gets its own INTERFACE target:
::common logos_json.h, logos_result.h
::consumer logos_lp_client.h, logos_async_result.h
::provider logos_module_context.h, logos_host_services.h
::host logos_host_core.h (new)
`logos_headers` stays as an umbrella over all four, so the ~70 existing
consumers are unaffected — logos_module_context.h alone has 66. Migrate to the
narrow targets when touching a repo; additive first, removal second.
NOTE the misnomer the split exposes: `logos_host_services.h` is MODULE-side
despite its name — it is the veneer a privileged module uses for services the
host granted it — so it belongs to ::provider, not ::host. Renaming it touches
9 files across 5 repos, so it is left for a change that can carry that cascade.
── logos_host_core.h ───────────────────────────────────────────────────────
`logos::host::LogosCore`, a plain RAII wrapper over liblogos' logos_core_* C
API, for the four programs that stand up a core (basecamp, logoscore-cli,
standalone-app, module-viewer). They currently open-code the same calls, and
basecamp had already grown a private wrapper for them.
It is deliberately an ORDINARY class — no codegen, no injection seam, no
void*. LogosModuleContext needs `_logosCoreSetContext_`, SFINAE `maybeSet*`
helpers and a void* round-trip because a module impl is user-authored but
FRAMEWORK-instantiated. A host is main(): it constructs this itself. For the
same reason there is no `modules()` here — the host holds its own LogosModules
from its own generated logos_sdk.h, so this header needs no generated type.
What it earns, each tied to a measured hazard:
* OWNERSHIP. liblogos allocates its char**/char* returns with new[], so
`delete[]` is correct and free() is undefined behaviour. That rule lived in
a comment in one repo's .cpp; it is now in one place.
* ORDERING. Three setters must precede logos_core_start(), stated only in
comments in logos_core.h. They are constructor arguments here, so the
illegal order is not expressible.
* SHAPE. logos_core_get_module_stats() takes no module name and returns one
blob for every module; stats(name) does that parse once.
The logos_core_* ABI is re-declared rather than included: logos-liblogos
depends on logos-cpp-sdk, so including its header would invert the graph. Every
host already hand-declares it; this makes it one declaration instead of four.
15 tests, 281/281 suite total. They define the extern "C" ABI themselves and
allocate exactly as liblogos does, so the ownership rules are exercised rather
than asserted; the ordering test records call order and pins start() as last.
Also fixes a real gap: nix/include.nix carries its own header list, separate
from cpp/CMakeLists.txt's install(FILES), so a new header silently did not ship
in the export layout.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* feat(generator): a Qt-typed umbrella that needs no LogosAPI
Splits a consumer's TYPE SURFACE from its TRANSPORT. Until now the qt umbrella
was `explicit LogosModules(LogosAPI* api)` while the lp one was default-
constructible, so "Qt types" implicitly meant "has a LogosAPI" — and a cdylib
module, whose provider surface is the std logos_module_impl.h C ABI and which
holds no LogosAPI anywhere, could not have Qt-typed dependency wrappers at all.
Its generated glue emits `new LogosModules()` unconditionally
(lidl_gen_cdylib.cpp:693), so the combination did not merely misbehave, it did
not compile.
That was a codegen choice, not a law: the wrapper bodies already run over lp_*.
`--binding api|origin` selects it, defaulting to `api`. A second enum rather
than a third ApiStyle value, deliberately: ApiStyle names the type surface and
is switched on by six emitters (makeHeader/makeSource/returnTypeFor/
paramTypeFor/toWireFor/fromWireFor); a "Qt types, explicit origin" member would
force all six to answer a transport question whose honest answer is "same as
Qt" every time. ApiStyle::Lp ignores the new axis — lp is origin-bound by
construction — and that is asserted rather than assumed.
The emitted umbrella bakes metadata.json#name as the origin literal:
LogosModules() : test_fullapi_cpp(QStringLiteral("test_fullapi_qtproxy")) {}
FullApi bind_full_api(const QString& moduleName) {
return FullApi(QStringLiteral("test_fullapi_qtproxy"), moduleName); }
Origin is the CONSUMER's own name and target is the dep — origin first in both
bind_ overloads. This is the load-bearing property: LpBridge::forTarget derives
origin from `api->moduleName()`, and reusing it silently gives a consumer the
caller's identity, which has already preserved a privilege escalation once in
this tree. An empty metadata name is refused at the CLI (exit 6, naming the
file) and emits `#error` in the header: a module that cannot state its identity
must not compile, and must never be handed a blank or borrowed one.
Verified additive on 172 real metadata.json x 2 api-styles = 344 runs, all
producing output, byte-identical old binary vs new. Mutation control: swapping
bind_<iface>'s (origin, moduleName) to (moduleName, origin) fails the suite at
MakeUmbrellaTest.QtExplicitOriginStatesTheConsumersOwnNameEverywhere. 281 -> 286
tests.
Framing worth keeping: the origin is SELF-ASSERTED from the module's own
metadata and is not attested by the transport. That is not a regression —
`api->moduleName()` is equally process-stated — but "explicit origin" means the
module names itself, not that the host vouches for the name.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* docs(generator): stop pointing callers at a flag that no longer exists
--backend qt is deleted from logos-qt-generator, and this repo was still
signposting it. main.cpp's refusal said "Use it for --backend qt", and the usage
text advertised --lidl … --backend qt and --from-header … --backend qt. Those
now point at nothing — the exact failure the deletion removes, one repo over.
The refusal names the real replacement chain instead: --backend cdylib here,
then logos-qt-host-generator --backend cdylib for Qt-plugin packaging.
docs/project.md's "Provider Generation" section documented three emitters that
no longer exist; rewritten to state the seam and the two-step pipeline.
docs/spec.md's dataflow diagram showed <name>_qt_glue.h / <name>_dispatch.cpp as
outputs; the diagram is corrected and the sections describing that shape are
marked historical rather than deleted, because the onInit wiring they document
still applies to the cdylib glue.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* refactor(cpp-generator): merge legacy/ into the generator, dropping its dead half
`legacy/` was never a library with an API surface. Two files, 475 lines, exactly
one exported symbol — `int legacy_main(int, char**)` — with every other
definition `static`, compiled INTO logos-cpp-generator and reached by fallthrough
at the end of main(). So there was nothing to keep separate: it is one mode of
this binary, and it now lives beside the others as plugin_introspect.{cpp,h}
behind `runPluginIntrospectMode()`, named for what it does.
Deleting it was never an option — that premise was checked and refused earlier.
`--general-only` alone has ~10 live callers across 7 repos including the central
module path (buildPlugin.nix:206,211, buildHeaders.nix:221,
LogosModule.cmake:423). The mode is load-bearing; only its packaging was wrong.
110 lines go with the move, all genuinely unreferenced:
* cppStringEscape — zero callers anywhere.
* writeUmbrellaHeader / writeUmbrellaSource and the `if (!moduleOnly)` block
that called them. This is the real prize: a SECOND, directory-SCRAPING
implementation of logos_sdk.{h,cpp}, unreachable in practice because
generate-module-headers.sh:60 always passes --module-only. generator_lib's
deps-driven makeUmbrella*FromDeps is now the only umbrella emitter, so the
two cannot drift.
* a dead `QJsonDocument doc(methods);` and its commented-out use.
With the block gone, `--module-only` suppresses nothing, so the parameter and
its plumbing go too. The FLAG stays tolerated rather than rejected, because
generate-module-headers.sh passes it unconditionally — a comment at the old
parse site says so.
Verified: #default builds, and both checks pass — `generator-cli` (which
exercises the CLI surface, including --general-only) and `tests`.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* feat(sdk): make the by-name call path a supported API
The dynamic (by-name) invoke path already existed and was already ungated at
every layer — lp_client_create / lp_invoke in the C ABI, logos::LpClient above
it, and the Qt client above that. Nothing checked a host service; there was no
gate to open. What was missing was the ERGONOMICS, which is what turned a
supported capability into something callers reached around the umbrella to get.
Three additive pieces, no gate touched:
1. LogosModuleContext::moduleName() — the module's own registry name, i.e. the
origin it authenticates as. The typed wrappers bake their origin in at
codegen time; a by-name call has to state one, and a wrong origin
authenticates as nobody and fails far from the call site. Set through a NEW
`_logosCoreSetModuleName_`, deliberately not a fourth parameter on
`_logosCoreSetContext_`: every generated provider calls that signature, so
widening it would break each one until regenerated, for a value the
generator knows statically. Set before the context, so moduleName() is live
inside onContextReady().
2. LogosModules::dynamic(target) on the origin-bound umbrella — the untyped
client, with the origin baked in exactly as the typed members' is, and
cached per target because LpClient owns a connection. The typed members over
metadata.json#dependencies stay the ordinary way to call another module;
this is for the cases whose target is a runtime value (a proxy, a router).
3. LpClient::getMethods() over the already-exported lp_get_methods. Invoke
without introspect is guessing — a caller that cannot ask what exists can
only hardcode, and a wrong guess fails at runtime like a typo.
Verified: #default builds, and both checks pass (tests, generator-cli).
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* fix(cdylib): shape-check [any]/{tstr:any} args, and fix the umbrella's includes
TWO fixes; the second is a regression I introduced two commits ago.
1. jsonArgToStd() returned the raw json for LogosList / LogosMap — "untyped JSON
passes through, as it always has". Arity was checked, type was not, so a
scalar reached a `[any]` parameter untouched. A proxy forwarding it through a
Qt-typed consumer then turned "notalist" into
["n","o","t","a","l","i","s","t"], because qvariant_cast reads a QString as a
sequential container — and the downstream provider saw a well-formed array
with nothing left to refuse. Now emits logos::jsonRequireArray /
jsonRequireObject; the throw lands in the dispatch's existing catch as
{"code":"dispatch_failed"}. Bare `any` stays raw, deliberately: it declares
nothing, so there is nothing to check it against.
Measured on the conformance matrix: closes all 8 failing cells (498/22/12/8
-> 500/18/12/2), and a whole-matrix per-cell diff shows exactly 14 status
changes, every one inside the two hostile cases. The other 518 cells are
byte-identical in status and value. The remaining 2 are the fix working — the
C++ cdylib provider now refuses the same input, which cases.json still pins
as lenient via expect_by_provider; that is a registry edit, and known.json's
Q1b already names this exact outcome as the intended fix.
2. The Lp umbrella emitted `logos::LpClient& dynamic(...)` and a
std::map<..., std::unique_ptr<logos::LpClient>> while <map>, <memory> and
logos_lp_client.h were conditional on interfaceNames. A module WITH
dependencies compiled by accident, because <dep>_api.h drags the header in
transitively. A module with NO dependencies and no interfaces includes
nothing else and failed outright with "no type named 'LpClient' in namespace
'logos'". test_fullapi_cpp is exactly that shape.
I shipped that in
|
||
|
Dario LipicarandClaude Opus 5
|
9d508292eb |
fix(generator): defer generated event subscriptions instead of acquiring a replica (#134)
* fix(generator): defer generated event subscriptions instead of acquiring a replica
Both C++ generators emitted, at all three subscription sites (the generic
on(QString, RawEventCallback), its EventCallback overload, and every typed
on<Event>):
LogosObject* origin = ensureReplica(); // blocking requestObject
if (!origin) return false; // PERMANENT -- never retried
m_client->onEvent(origin, eventName, callback);
That asks "is the module reachable right now" at the one moment the answer is
no. Every C++ consumer subscribes from init(), onContextReady() or a backend
constructor, all of which run while the dependency's host has been spawned but
has not called listen() yet. The guard inside requestObject was dead code for
years -- isConnected() returned a latch that was always true -- so the call fell
through to a blocking wait that usually succeeded, slowly. Making isConnected()
truthful turns the same code into an instant, permanent, silent failure: the
wrapper compiles, returns a bool, and never delivers.
All three sites now route through the deferred channel:
return m_client->onEventWhenAvailable(m_moduleName, eventName, callback) != 0;
and ensureReplica() / m_eventReplica are deleted from both generators. Keeping a
per-wrapper replica would reintroduce both halves at once -- a blocking acquire
on the subscriber's thread, and a permanent failure when the module had simply
not started yet.
The return becomes ACCEPTED rather than live, false only for errors no retry can
fix. That is stated in the emitted comment so it reaches every generated file
rather than only this message.
VERIFIED AT THREE LEVELS, because the first two prove less than they look:
emits -- 265/265 cpp-sdk tests. The goldens now pin the emitted CALL SITE and
EXPECT_FALSE the removed symbols; they are string comparisons and
would pass on code that does not compile, which is exactly how this
defect survived.
compiles-- both generators' output compiled against the local protocol branch
(EXIT=0), including a 15-event contract with a 3-parameter event.
defers -- real A/B on a live qt_remote transport with real generated code:
7/7 green on the migrated generator, 3/3 red in 0-8 ms on the
pristine one, with published-first controls green in both.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* chore(deps): bump logos-protocol to 0183e8c for onEventWhenAvailable
The generator change in this PR emits `m_client->onEventWhenAvailable(...)` at
all three subscription sites. This repo pinned logos-protocol 0f26ffd, which has
zero occurrences of that symbol -- compiling the emitted wrapper against it gave
EXIT=1 and 16 errors, every one "no member named 'onEventWhenAvailable' in
'LogosAPIClient'". That is why this PR was opened as a draft and why the bump has
to ride in the SAME commit range as the emission change: split them and cpp-sdk
master is red for every Qt-api-style consumer.
0183e8c is logos-protocol master with #47, #53 and #55 in. It is deliberately not
the first commit that introduces onEventWhenAvailable: #47's tip also carries the
use-after-free fix for tryAcquireNow (09f684f), without which a consumer that
subscribes more than once to a not-yet-reachable module frees a QtRO facade that
is still registered in a shared replica implementation's connect list. Generated
Qt consumers subscribe exactly that way -- one on<Event> per declared event, from
init() -- so pinning below that commit would make this change crash rather than
merely fail to compile.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
|
||
|
Dario LipicarandClaude Opus 5
|
f3369faca4 |
feat(generator): async callers can see the error, sync callers can set a deadline (#132)
The two consumer surfaces had complementary holes:
sync : T foo(params…, logos::CallError* err = nullptr) error yes, timeout NO
async: void fooAsync(params…, cb, Timeout = Timeout()) timeout yes, error NO
so an async caller could not tell a failed remote call from a provider that
legitimately returned 0 / "" / false — the exact ambiguity the sync path's
CallError* was added to resolve — and a sync caller could not say how long it
was willing to wait, even though the transport overload the generator already
calls takes both.
Both fixes are additive:
T foo(params…, logos::CallError* err = nullptr, Timeout timeout = Timeout());
void fooAsync(params…, std::function<void(T)> cb, Timeout timeout = Timeout()); // unchanged
void fooAsyncResult(params…, std::function<void(logos::AsyncResult<T>)> cb,
Timeout timeout = Timeout()); // new
logos::AsyncResult<T> (new, Qt-free, cpp/logos_async_result.h) is {value, error}
plus ok(); AsyncResult<void> carries only the error so every fooAsyncResult has
the same callback shape. The name is distinct rather than an overload because
std::function<void(AsyncResult<T>)> next to std::function<void(T)> is ambiguous
for a generic lambda.
Applied to both emitters that produce this surface — legacy/generator_lib.cpp
(the module-builder path) and experimental/lidl_gen_client.cpp (`--lidl
--module-only`, from a published contract) — since a consumer can reach either
for the same contract.
The Qt-free (ApiStyle::Lp) surface gets the sync timeout (spelled `int
timeout_ms`; `Timeout` lives behind a Qt header) but NOT fooAsyncResult:
logos-protocol's lp_invoke_async hard-codes `cb(1, …)`, so an AsyncResult there
would report ok() on a failed call. Measured, not assumed. See the note in
makeHeaderLp.
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
|
||
|
Dario Lipicar
|
d972fe207a |
fix(generator): no silent drops — unread STRUCTURE in a record is a build error (#131)
#127 made an unknown C++ TYPE a build error instead of a silent `any`. This is the same hole one level down, in the record scanner: a line inside a `struct` body that the scanner could not read as a field was `continue`d, and the struct was published anyway — MINUS that field. A field list is not a detail of a record, it IS the record: the promise every other language binds to. A contract with two fields is exactly as well-formed as one with three, so nothing downstream could tell. Measured on the generator built from master: struct SplitRecord { -> type SplitRecord { exit 0, no diagnostic std::string id; id: tstr std::vector<std::string> n: uint tags; } uint64_t n; }; `tags` is simply GONE struct Outer { -> type Outer { a: int } exit 0 struct Inner { the record is made of the INNER type's int64_t a; field and has none of its own }; Inner inner; std::string label; }; struct Pair { -> method echoPair(v: Pair) with NO `type Pair` std::string first; std::string second; emitted — a contract naming a }; type it never declares struct Defaulted { -> the brace-initialised field is dropped std::string id{"none"}; (only the `= v` form was recognised) int64_t n = 0; }; struct Allman -> not a record AT ALL, so every mention of it { falls to the `any` fallback — since #127 an std::string id; error whose hint says "declare a struct", }; which is the thing the author declared The scanner now reads a body as DECLARATIONS rather than lines. Physical lines are joined until the declaration is whole — a `;` at the struct's own brace depth, or a `}` there, which is how a member function defined inline ends — exactly as the caller already joins a method signature until its parentheses balance. Where a brace sits, and where a line wraps, cannot decide what a header means. Allman and base-clause openings are recognised for the same reason. Whatever is left after that, and after the constructs that definitively are NOT fields (member functions, `using`/`typedef`/`friend`/`static`, access specifiers), is reported instead of skipped, naming the struct, the declaration and the fix. Same withdrawal discipline as #127: the diagnostic is keyed on the struct and tested against the API-REFERENCED set, so a helper struct the module never publishes may be as unreadable as it likes — every real module carries one (openmetrics' `ModuleSource`, the package manager's `PendingAction`). Comments come off with a literal-aware strip. The old bare `indexOf("//")` truncated `std::string url = "http://x";` inside the literal, which merely lost the field before and would now reject valid code. Blast radius, measured: all 27 universal-interface impl headers in the workspace emit BYTE-IDENTICAL .lidl and stderr, with identical exit codes. All 6 structs those headers declare are K&R with no unparsed body line, so nothing in tree changes. logos-qt-generator, which compiles this same file, builds; and test_fullapi_ext_cpp — the records-heavy module — builds end to end. 14 new tests. 10 of them fail on the unpatched parser; the rest are the "must still work" controls, including one for an inline member-function body that an earlier cut of this change would have broken. |
||
|
Dario LipicarandClaude Opus 5
|
b4c2e5bb2d |
fix(generator): one declaration, one binding — optional record fields on the legacy path (#130)
`? maybe: tstr` and `maybe: ?tstr` are the same declaration. logos-lidl's
docs/spec.md says so and requires them to produce byte-identical code, and
logos-lidl#7 added fieldIsOptional()/fieldValueType() precisely so no backend
re-derives the answer. The cdylib, client-stub and both Rust backends honour
that. The LEGACY consumer path — the one every real C++ module builds through,
`logos-cpp-generator --general-only --dep <name>=<lidl>` from
logos-plugin-qt's buildPlugin.nix — did not:
? maybe: tstr -> QString maybe{}; __m.value("maybe").toString()
maybe: ?tstr -> QVariant maybe{}; __m.value("maybe")
and on `--api-style lp`, `std::string maybe{}` vs `LogosMap maybe{}` — neither
of them std::optional. The flag spelling is the one production contracts use:
logos-chat-module writes all five of its optionals that way, so every one of
them landed on the branch that silently defaults. A `QString` has no empty
inhabitant at all; an absent `nickname` and an empty-string one were the same
value by the time the consumer saw them.
ROOT CAUSE. legacy/main.cpp's moduleRecordsToJson / moduleMethodsToJson /
moduleEventsToJson flatten every TypeExpr into a single Qt TYPE-NAME STRING.
Answering "is this optional" from a name means answering it from the verbatim
spelling, which is the one thing the accessors exist to stop.
WHAT THIS CHANGES. The record-field half of that boundary, and only it. A field
object now carries `optional` alongside `type`, where `type` is the VALUE type
(fieldValueType) and `optional` is true for either spelling (fieldIsOptional).
Both spellings arrive at the emitter as the same object, so they leave as the
same code. Per surface, matching the sibling backend that already serves it:
Qt — QVariant, as lidl_gen_client.cpp already emits. Qt has no optional
template; an invalid QVariant is its single empty inhabitant. Two-state,
untyped.
Lp — std::optional<T>, as lidl_gen_cdylib.cpp already emits, encoded by
logos-protocol's Codec<std::optional<T>>. Keeps the value type.
`?any` / `?{K:V}` / `?[any]` collapse onto the bare LogosMap/LogosList:
nlohmann::json already carries null, so wrapping it would give the slot
two empty spellings — three states, which the two-state rule forbids.
Encode omits the key when empty (a record field is a NAMED slot); decode treats
an absent key and an explicit null as the same state, so neither turns empty
into "" or 0. The round trip is canonicalising, as the spec requires.
The three functions move to legacy/lidl_to_json.{h,cpp}. Not cosmetic: the rule
is a property of frontend -> JSON -> emitter, and while they sat inside a TU
with main() no test could observe it. tests/generator/test_optional_spellings.cpp
now runs that composition end to end.
WHAT THIS DOES NOT CHANGE, and why. POSITIONAL slots — method parameters,
return types, event parameters — are still flattened to QVariant (Qt) /
LogosMap (Lp). They have no name to hang a flag on, so they only ever had the
type-kind spelling and there is no divergence there to fix; what they lose is
the value type. Closing that changes generated method SIGNATURES, i.e. a source
break for every existing call site, for a defect this commit is not about.
`OptionalSpellings.PositionalSlotsAreStillFlattened` pins the current behaviour
so closing it later is deliberate, and the generator still prints a `Note:`
naming every slot it flattens. Nesting, map key types and descriptions still do
not cross the boundary either — the flag is per-field, not a general widening.
VERIFIED BY RUNNING, each check shown to fail when the property does not hold:
- Two contracts identical but for the spelling, generated with `--dep` on both
`--api-style qt` and `--api-style lp`: byte-identical. The SAME comparison on
a generator built from this base without the fix reports a difference on
both styles.
- Harness sensitivity: two contracts differing only in `count: uint` ->
`count: int` are reported as different, so the compare is not vacuous.
- A contract with no optional field generates byte-identically to the
unpatched generator, both styles — and the same compare reports a difference
when given genuinely different output.
- The 5 new assertions FAIL on this base with only the extraction grafted in
(generator_lib.cpp pristine) and pass with the fix; the 3 control assertions
pass on both, which is what makes them controls.
- Generated wrappers compile on both surfaces, for the probe contracts and for
the real logos-chat-module and test_fullapi_ext_rust contracts.
- Emitted lp record codec exercised at runtime: empty omits the key, an
explicit null decodes to nullopt rather than "", uint64 above 2^63 survives,
bstr stays canonically tagged, and `{"maybe": null}` re-encodes omitted.
- `nix build .#tests` green (107 tests), `nix build .#cpp-generator` green.
logos-qt-sdk's qt-generator has the same defect in lidl_gen_qt_consumer.cpp
(record fields read `f.type` directly); it is a separate repo and not touched
here.
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
|
||
|
Dario Lipicar
|
b594d1a38c |
fix(generator): the Qt consumer must not flatten a rejection into an empty value (#129)
* fix(generator): the Qt consumer must not flatten a rejection into an empty value
A provider that REJECTS a call answers the canonical
{"code":"dispatch_failed", "message":..., "origin":...} object as its RESULT,
not as a transport error — the same object from every provider flavour
(logos-qt-sdk dispatchFailedVariant, the generated cdylib dispatch, the Rust
provider's args::dispatch_failed).
The generated Qt consumer wrapper converted it like any other value, which
ERASES it. `_result.toList()` on that map is `[]`, `.toString()` is "",
`.toLongLong()` is 0 — so `echoUintList([1, -1, 3])`, answered `dispatch_failed`
by the provider, reached the caller as `[]`: the whole list, not the bad
element, and indistinguishable from "the provider returned nothing". Losing the
error is a consumer bug whatever the provider did.
Reported through the channel this surface already uses for a failed call: the
`logos::CallError*` out-parameter every generated sync method carries (today
"object_unavailable" when the target cannot be acquired). No signature changes,
no return value changes — a caller that passes `err` now sees
code="dispatch_failed" with the provider's diagnostic and origin; a caller that
does not gets the same default it always got, plus the qWarning the wrapper
already emits for a failed call.
* the detector is emitted once per wrapper, in an anonymous namespace, and
matches EXACTLY (three string fields, that code) for the same reason
logos_rpc_status.h's isUnauthorizedSentinel matches exactly: an `any` or map
return carrying user data must never false-match.
* the result is now captured for `void` returns too — a void method can be
rejected, and the rejection object is the only place that says so.
* the async overload's callback takes the value alone and has no error
parameter; giving it one would change the generated public surface (which
logos-qt-sdk's veneer mirrors 1:1), so an async rejection is logged rather
than delivered. Recorded in cpp-generator/docs/project.md.
Scope: CONSUMER side only. The provider half of registry entry Q1 (a Qt-typed
provider cannot validate the ELEMENT of a typed numeric array, because a C++
signature spells [uint] and [any] alike as QVariantList) is explicitly out of
scope and unchanged. The lp wrapper is untouched — its generated output is
byte-identical before and after.
* fix(generator): guard the emitted detector — the umbrella is one translation unit
`logos_sdk.cpp` textually `#include`s every generated `<dep>_api.cpp`, so a
module with more than one dependency compiles several copies of the detector
into ONE translation unit:
test_fullapi_rust_api.cpp:15:6: error: redefinition of 'logosDispatchRejection'
Internal linkage covers the separate-TU case; only the preprocessor covers this
one. Found by building test_fullapi_qtproxy (3 wrappers: two concrete deps plus
the bound `full_api` interface) against this generator — a single-wrapper
contract cannot reach it.
|
||
|
Dario LipicarandClaude Opus 5
|
44b92b0480 |
feat(generator): no silent admissions — an unknown C++ spelling is a build error (salvage of #112) (#127)
* feat(parser): name nlohmann::json explicitly, ahead of killing the fallback `nlohmann::json` (and the `json` alias) has never had a branch in cppTypeToLidl. It reaches the opaque `any` the same way every unrecognised spelling does: the fallback at the bottom of the function. That is fine while the fallback is silent and wrong the moment it becomes an error, because `any` is the RIGHT answer here. test_fullapi_cpp declares `nlohmann::json echoAny(const nlohmann::json&)`, `bool fireAnyEvent(const nlohmann::json&)` and `logos_events: void anyEvent(const nlohmann::json&)`, and those three are the cross-language conformance chain's `any` cells — they must keep publishing `any`. So this lands first and on its own. It maps to the bare `any` primitive, which is exactly what the fallback already produced, making the change output-neutral: generating over every impl header and every .lidl in the workspace produces 764 byte-identical artifacts. That is what lets the later commit treat everything still reaching the fallback as a silent admission rather than a legitimate `any`. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> * feat(cdylib): a typed map decodes into the author's own container `{tstr: T}` has two C++ spellings — std::map<std::string, T> and std::unordered_map<std::string, T> — and logos_codec.h specializes Codec for both, because they are the same wire shape. The generated dispatch named one: lidlImpl().echoIntMap(logos::fromJson<std::map<std::string, int64_t>>(...)) fromJson returns a std::map, and a std::map does not convert to an unordered_map parameter. So the container the author picked decided whether generated code they never wrote compiles — with the diagnostic pointing at that generated line, not at their declaration. logos::JsonArg (logos-protocol, already on master) exists for exactly this: it instantiates its conversion operator with the parameter's own type, so the author's declaration drives the decode. The return side is the same problem mirrored, and `logos::toJson(result)` deduces instead of asserting. Restricted to Map because every other LIDL type has one C++ spelling here, and because JsonArg documents one target it cannot serve — std::optional<X>, whose converting constructor out-ranks the proxy's conversion operator. The Optional branch returns before reaching this code. For a std::map author nothing changes: JsonArg instantiates the same Codec<std::map<...>>::from at the same path, and toJson deduces the same T. Compile-checked against both spellings, including a bad element still failing with `expected string at arg0.k, got number`. Across the whole workspace the emitted delta is 5 methods, all in test_fullapi_ext. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> * feat(parser): an unknown C++ spelling is a build error, not a silent `any` cppTypeToLidl ended with `// Fallback: treat as opaque` -> `any`, and `any` is ADMITTED by every backend gate. So a spelling nobody had written a branch for was accepted in silence, published as `any`, and dispatched as a bare `lidlImpl().f(args.at(0))` — no logos::fromJson<>, no check. That is the one hole #113-#122 closed for every typed slot and left open for anything that reached `any`. An 11-method hostile probe was admitted wholesale: uint32_t, size_t, float and uint8_t all typed as `any`; std::set, std::vector<std::pair<...>> and a non-string-keyed map as `any`; std::vector<uint32_t> as `[any]`. Now every spelling with no LIDL type is collected with the declaration that carries it, and parseImplHeader turns the list into a parse error naming the offending type and the fix: method 'send_generic_public_transaction': parameter 'instruction' declared `const std::vector<uint32_t>&`, whose element `uint32_t` has no LIDL type. LIDL numbers are 64-bit only. Declare it `uint64_t` (LIDL `uint`). Widening is source-compatible for every caller; a narrow type on the wire is not, which is why LIDL has none. Numbers get that tailored hint (uint8_t its own — it means bytes here, and only as std::vector<uint8_t>); sets, pairs/tuples, non-string map keys, list/deque/array, Qt types and pointers each get theirs; anything else gets the full table of recognised spellings. A hint that does not name a replacement just moves the guesswork, so all of them do. std::unordered_map<std::string, T> joins std::map as a spelling of `{tstr: T}` — the codec has always handled both, and the previous commit made the generated dispatch bind whichever the author declared. Two slots are the exception and say so: a record FIELD and an event PARAMETER, where the generator writes the spelling out into code the author's own declaration has to match and can only pick one name. Three properties keep this from breaking things it should not: * The MAPPING is unchanged. cppTypeToLidl still returns `any` for an unsupported spelling; only a diagnostic is recorded. A diagnostic that is later withdrawn therefore leaves output byte-identical. * Diagnostics are withdrawn for declarations that never reach the contract — a helper struct dropped by keepOnlyReferencedRecords, a reserved lifecycle hook (onContextReady and friends), a struct with no parsed fields. Publishing is what makes a type a promise. * An empty spelling is not a C++ type, it is this line-based parser failing to find one. It keeps the old behaviour rather than reporting `''`. Also fixes a latent ordering bug found while threading the context through: metadata.json's event parameters were typed BEFORE the header was read, i.e. against whatever g_recordNames the previous module's parse had left behind. They are now read there and typed after scanForRecords, in the same position in module.events as before. Verified by generating over every impl header and every .lidl in the workspace: 31 derived contracts byte-identical, 368 consumer-umbrella files byte-identical under both --api-style qt and --api-style lp, 46 generated types headers byte-identical. Exactly two modules now fail, at exactly the four slots a prior scan identified as silent admissions. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> * feat(cdylib): a wrong argument count reports invalid_args The arity gate was `if (args.size() < N) return nullptr;`. The Qt glue turns a NULL reply into an empty QVariant, so "you passed 2 of 4 arguments" was indistinguishable from a method that legitimately returned nothing. logos-rust-sdk already ships the other half. src/args.rs::invalid_args is documented "Same code and message as the C++ generated glue" and pinned by a test named invalid_args_shape_matches_cpp — both of which were false: Rust answered a structured object and C++ answered NULL. Checked against the JSON that crate actually emits rather than against its comment, the two are now byte-identical: {"code":"invalid_args","message":"expected 4 arguments, got 2","origin":"my_module"} `expected` counts REQUIRED parameters in both, so a trailing optional does not change it. The guard is emitted only when the method has at least one required parameter. args.size() is unsigned, so `< 0` never fires: a zero-argument method carried a dead branch. The Rust generator skips it for the same reason, so the two now agree on when a check exists at all. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> * refactor(cdylib): delete the dead emitted base64 codec #117 replaced the codec the generator emitted into every module with logos-protocol's logos_codec.h, but left the base64 pair it had grown around: * lidlB64Idx + lidlBytesFromJson — 55 emitted lines with NO call site at all. Every byte parameter had already moved to logos::bytesFromJsonLenient. Confirmed across the whole workspace: in 48 generated export TUs, all 48 mentions of lidlBytesFromJson are its own definition line. * lidlB64UrlEncode + lidlBytesToJson — 34 emitted lines that are logos::bytesToJson rewritten, in a translation unit that already includes it through "<module>_types.h". Scalar bstr slots now call logos::bytesToJson. Composite ones ([bstr], {tstr: bstr}, records) have gone through logos::Codec since #117, so this removes the last place a module carried its own copy of an encoder — the arrangement that once let the emitted and canonical halves disagree about padded base64, and that #117's own comment set out to end. hasBytesEventParam goes with it: it existed only to keep the emitted copy from sitting unused in the events sidecar of modules whose events carry no binary data, and the `namespace { }` block it gated is gone too. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> * docs(doctest): the bstr event marshal is logos::bytesToJson now Commit (D) deleted the dead emitted base64 codec, so the cdylib events sidecar calls logos::bytesToJson -- the one in logos_codec.h, included in the same translation unit -- instead of emitting its own lidlBytesToJson. The doctest still pinned the old spelling and failed on the new output: expected 'args.push_back(lidlBytesToJson(frame));' not found in output The prose around it is unchanged and still correct: the payload is still the canonical {"_bytes": "<base64url>"} form, which is the property that assertion exists to guard (#99). Only the symbol moved. Verified against real generator output rather than by search-and-replace: built the branch generator, ran --backend cdylib over a sensor_module contract with a bstr event param, and read the emitted line: args.push_back(logos::bytesToJson(frame)); Checked the rest of the specs for other stale symbols (lidlStrdup, b64Url, hasBytesEventParam, the old 'return nullptr' arity gate) -- this was the only one. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5 <noreply@anthropic.com> |
||
|
Dario LipicarandClaude Opus 5
|
5a809c17a6 |
docs(generator): record where a dependency entry is read (#126)
The generator's own description still had `dependencies[]` as a list of names, and the shared-helper list under share/lidl-frontend still had the three files it had before metadata_dependencies.h joined them — the one a reader consults before adding a header that impl_header_parser.cpp includes, which is how the qt-generator's build breaks from a change made here. Documents both entry forms where the umbrella is described, and why the array is read once rather than per emitter: the aggregate is emitted by several passes, and passes that answer that question separately can answer it differently, which is a member whose type was never included. Also notes the umbrella emitters' new home in generator_lib, the tests that cover them, and the fixture declaring both forms. Co-authored-by: Claude Opus 5 <noreply@anthropic.com> |
||
|
Dario LipicarandClaude Opus 5
|
198f0317ca |
feat(optional): ?T is two-state, and the generators finally read it (#125)
* feat(optional): ?T is two-state, and the generators finally read it
No generator in any language read the optional flag — it had never been
implemented. `?T` was a HARD REJECT on the cdylib backend ("module not
cdylib-eligible"), `std::optional<T>` in an impl header fell through to the
opaque `any` with no diagnostic, and a `? name: T` field was emitted as a
required `T`. Three real contracts in the workspace already declare optionals
and were silently getting one of those three answers.
`?T` is TWO-state: a value of T, or empty. Never three — "one LIDL type <-> one
type per language" leaves nowhere for a third state, because every target has
exactly one empty inhabitant.
ONE MEANING, TWO SPELLINGS. `? name: T` (the field flag) and `name: ?T` (the
type kind) are the same declaration. Backends no longer answer that themselves:
logos-lidl's fieldIsOptional/fieldValueType are re-exported from lidl_compat.h
and every site THIS COMMIT TOUCHES reads them, so the two spellings emit
byte-identical code on the cdylib and client backends. That
caught a live drift on the way in — lidlRecordCollidesWithBytesTag read `f.type`
and so refused `? _bytes: tstr` while letting `_bytes: ?tstr` straight through,
one declaration with two answers.
THE WIRE RULE DEPENDS ON THE SLOT. Absent and explicit null are the SAME state
on decode and DIFFERENT on encode:
- decode is liberal, by exactly one inhabitant: in an optional slot absent and
null both mean empty; in a required slot both stay errors. A present value
goes through the decoder a required T would get, so a wrong type still fails
at the same path — optional widens the domain, it does not switch checking
off. `?bstr` therefore keeps the LENIENT bytes decode a bare `bstr` gets,
rather than silently becoming stricter in the optional slot.
- encode has one canonical form: empty OMITS the key where the slot is NAMED
(a record field) and is spelled null where it is POSITIONAL (an argument, a
return, an event parameter — no key to omit, and arity must not change). Key
omission lives in the record emitter because a Codec only ever sees a value,
never the slot it sits in. A round trip therefore canonicalises.
- `?any` collapses onto `any`: nlohmann::json already carries null, so
std::optional<LogosMap> would give the slot two spellings of empty.
The dispatch gate now admits a missing trailing optional argument and
materialises it as null, exactly the way a missing record field already was. A
method with no optional parameter emits the byte-identical gate it always did.
Header-first: `std::optional<T>` <-> `?T`, composing with records and
containers. `std::optional<std::optional<T>>` has NO LIDL type (three C++ states
over a two-state wire), so it maps down to `?T` — which makes the author's own
declaration stop compiling against the generated codec, deliberately — and says
so at derivation time instead of leaving a conversion error in generated code.
The Qt/Lp consumer surface is NOT fixed and does not pretend to be. The wrappers
real modules get come from legacy/main.cpp, where the AST is flattened to a
single Qt type-name string per slot before optionality could be seen; Qt has no
optional metatype, so `?T` lands on QVariant — the right shape (an invalid
QVariant is Qt's empty inhabitant) with no type. The generator now prints a Note
naming every flattened slot so an affected build is never silent, and
docs/project.md records exactly what a Qt consumer will still do with an
optional field.
Verified by output equivalence, not by a green build: the generator was built
before and after and run over every .lidl in the workspace plus the impl-header
fixtures, in cdylib, consumer-qt, consumer-lp, client and header-first modes.
428 of 465 artefacts are byte-identical; all 37 that differ belong to one of the
four contracts that declare an optional (the 38th path is the manifest). The
harness's sensitivity is pinned by a negative control: qt vs lp output differs
in 45 files. The emitted codec was additionally compiled under -Wall -Wextra and
run against the rules above — omission, absent==null, required-still-rejects,
present-but-wrong-still-fails, and canonicalising round trip.
Tests: 199 pass, 0 fail (180 before, 19 new).
Requires logos-lidl's optionality accessors and logos-protocol's
Codec<std::optional<T>>.
NOT FIXED, AND IT IS THE PATH THAT MATTERS MOST. The legacy interface-wrapper
path is untouched, and it is the one every real module builds through
(buildPlugin.nix:145 -> logos-cpp-generator --general-only). There the two
spellings still diverge:
? maybe: tstr -> QString maybe{}; __m.value("maybe").toString()
maybe: ?tstr -> QVariant maybe{}; __m.value("maybe")
and --api-style lp diverges too, neither side being std::optional. So R3 holds
on the backends below and NOT on the Qt consumer a shipping module actually
gets. logos-chat-module -- the contract that prompted this work -- uses the
field-flag spelling, so it lands on the branch that silently defaults.
The cause is upstream of codegen: legacy/main.cpp's moduleRecordsToJson and
moduleMethodsToJson flatten every TypeExpr to a single Qt TYPE-NAME STRING, so
optionality (along with nesting, map key types and descriptions) is gone before
generator_lib.cpp sees it. Widening that interface is a larger change and is
deliberately not attempted here. The only R3 test on a Qt surface covers
lidl_gen_client.cpp, which is on no live build path.
* chore: re-pin logos-lidl to master for the optionality accessors
lidl_compat.h re-exports typeIsOptional / optionalValueType / fieldIsOptional /
fieldValueType / paramIsOptional / paramValueType, which landed in
logos-lidl#7. The pinned lidl predated it, so CI failed to compile.
logos-lidl 8c95d4f -> 35f33d8. Tests: 199 pass, 0 fail.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* chore: re-pin logos-protocol to master for Codec<std::optional<T>>
The generated record codecs emit Codec<std::optional<T>> for an optional
field; that specialisation landed in logos-protocol#37 and the pinned
protocol predated it.
Note this repo's own tests would NOT have caught the omission -- the
generator tests string-assert emitted text rather than compiling it, so a
missing codec specialisation only surfaces when a real module compiles
generated optional code (logos-test-modules' ext provider).
logos-protocol 4359557 -> 72754ab. Tests: 199 pass, 0 fail.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* test(doctests): override logos-lidl alongside every logos-cpp-sdk override
The doc-tests build downstream repos (logoscore-cli, capability_module,
accounts_module) with --override-input logos-cpp-sdk. Nix does not carry the
overridden input's OWN lock, so those builds got this branch's cpp-sdk source
while still resolving logos-lidl from their own, older locks. The shipped
share/lidl-frontend/lidl_compat.h then calls accessors that lidl does not
have:
lidl_compat.h:46: error: 'paramValueType' has not been declared in 'lidl'
lidl_compat.h:92: error: 'fieldValueType' was not declared in this scope
Every --override-input logos-cpp-sdk now has a matching
--override-input <same-path>/logos-cpp-sdk/logos-lidl.
This is specific to the override path. A normal consumer running
'nix flake update logos-cpp-sdk' inherits cpp-sdk's own lock, which pins the
lidl carrying these accessors, and is unaffected.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* test(doctests): move logos-lidl at the qt-sdk nodes, not under logos-cpp-sdk
The doc-tests failed to build logos-qt-generator:
share/lidl-frontend/lidl_compat.h:46: error: 'paramValueType' has not been
declared in 'lidl'
MECHANISM. This SDK installs cpp-generator/experimental/lidl_compat.h into
$out/share/lidl-frontend/, and logos-qt-sdk's logos-qt-generator *compiles*
that installed header against qt-sdk's OWN logos-lidl input. Under
logos-qt-sdk, logos-lidl is a SIBLING of logos-cpp-sdk, not a descendant:
logos-qt-sdk
|-- logos-cpp-sdk <- --override-input moves this to the commit under test
`-- logos-lidl <- stays on qt-sdk's lock (8c95d4f), lacks the accessors
logos-logoscore-cli and logos-module-builder both declare
`logos-qt-sdk.inputs.logos-cpp-sdk.follows = "logos-cpp-sdk"` but no lidl
follows, so overriding the SDK hands qt-sdk a new lidl_compat.h next to its
old lidl. The failing derivation is logos-qt-generator — not anything in
logos-cpp-sdk, which is why the previous attempt aimed at the wrong node.
THE FIX is one `<path-to-logos-qt-sdk>/logos-lidl` override per qt-sdk node
that ends up on the SDK under test. A tree-walk over the resolved lock found
four in logoscore-cli's closure and two per module build; with the overrides
applied the walk reports zero remaining.
WHAT WAS REMOVED, and why it was doing nothing:
* The `.../logos-cpp-sdk/logos-lidl` overrides added in
|
||
|
|
fde0f6fccb |
fix: read dependency entries declared in object form (#123)
* fix: read dependency entries declared in object form The manifest schema lets a dependency entry be an object carrying the name alongside the constraints an installer resolves it by, but every reader took the element as a plain string and skipped what came back empty, so an object entry disappeared: the module it names was left out of the generated LogosModules aggregate, and every call through it failed to compile. The rule lives in one place now, since the copies of it are how the gap spread. It ships in share/lidl-frontend alongside the parser that includes it, which consumers compile from there. * fix: read every dependency entry through one pass over the array The object form reached the umbrella's members and constructor but not its includes: that emitter still read each element as a plain string, so a module declared in object form came out as a member whose type was never included, and the aggregate no longer compiled. It is the Qt-free umbrella, which is what every universal core module and every cdylib module generates, so the form the previous commit set out to support failed there in a new way rather than working. Reading the array element by element is what let one pass disagree with the next, so no reader does that any more: dependencyNames() answers what an array declares, once, and the emitters walk names. That leaves the entry form knowable in exactly one place, and the includes and members of an aggregate can no longer be built from different answers. The umbrella emission moves to generator_lib alongside the per-module wrapper emitters it mirrors, returning the text instead of writing it, so what it generates can be asserted on directly; main.cpp writes what it returns. Output for string-form dependencies is byte-identical in both API styles, with and without interface dependencies. The listing mode (`--metadata` with no `--module-dir`) went the same way — it was the last reader still deciding on its own. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> --------- Co-authored-by: Dario Gabriel Lipicar <dario@status.im> Co-authored-by: Claude Opus 5 <noreply@anthropic.com> |
||
|
Dario LipicarandClaude Opus 5
|
d11fbb2220 |
docs: drop the reference to the deleted lidlGenerateProviderGlue (#124)
logos-qt-sdk#23 deleted that function -- it was a second, parallel top-level pipeline over the same emitters main.cpp already drives, with no callers anywhere in the workspace. This doc bullet was its only surviving reference. The .lidl sidecar the bullet described is emitted elsewhere and is unaffected: main.cpp:92/:184 here, and module-builder's modulePreConfigure.nix:75. Co-authored-by: Claude Opus 5 <noreply@anthropic.com> |
||
|
Dario LipicarandClaude Opus 5
|
99a2bffeb3 |
refactor(generator): drop the dead event-source surface from consumer wrappers (#120)
* fix(generator): one Qt type mapper, and it knows about void; LpClient takes a timeout
Two near-duplicate LIDL->Qt type mappers existed — legacy/main.cpp's
lidlTypeExprToQtTypeName and experimental/lidl_emit_common.cpp's lidlTypeToQt —
and they disagreed. The legacy one had no `void` case, so a `-> void` method
arriving as Primitive("void") from the impl-header parser fell through to
QVariant. (The .lidl parser spells it Named("void"), which survived only by
accident, through mapReturnType's `base == "void"` early-out.)
That was not a Qt-consumer bug: the std/lp tables are DERIVED from this name, so
the same method generated `LogosMap doVoid(...)` on the Qt-free surface too.
Measured, from `void doVoid();` in a .h interface:
QVariant doVoid(...) --api-style qt before
void doVoid(...) --api-style qt after
LogosMap doVoid(...) --api-style lp before
void doVoid(...) --api-style lp after
lidlTypeExprToQtTypeName is now a delegation, so there is one table to disagree
with. This changes the generated signature for any module consuming a `-> void`
method through a .h interface; the two in-tree call sites discard the value and
are unaffected.
logos::LpClient::invoke/invokeAsync gain a timeout_ms parameter, defaulted to
the C ABI's "use the default" (0) so no existing caller changes. The Qt-typed
consumer surface takes a Timeout on every async overload and had nowhere to put
it — a wrapper delegating to the lp path silently dropped it.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* refactor(generator): drop the dead event-source surface from consumer wrappers
Generated consumer wrappers carried setEventSource / eventSource / trigger — an
author-facing way to SOURCE events through a wrapper whose job is to CONSUME
them. Both emitters (legacy and experimental) shipped it.
Nothing used it. Zero call sites across every repo in the workspace including
the vendored SDK copies; the only `trigger(` in the tree is a QML Action's own
method. The generated code did not use it internally either — m_eventSource was
written only by its own setter and read only by trigger, so calling trigger()
without a prior setEventSource() warned and returned.
It was not free. `trigger` routes through m_client->onEventResponse, which has
no lp equivalent — lp_* offers only lp_provider_emit_event, on a handle a
consumer wrapper does not own. That single call was the reason a Qt wrapper had
to keep a LogosAPIClient alongside its lp client, carrying two clients and two
lots of token state per wrapper. Removing an unused surface removes a real
constraint on the veneer.
Worth noting what it would have taken otherwise: either widening the C ABI with
a consumer-side emit (softening a provider/consumer split the ABI currently
enforces), or rerouting through the module's own provider handle. Neither is
needed if nobody is asking.
Pinned by a test rather than left to convention — the emitters are the kind of
code where a convenience accessor grows back.
181/181.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
|
||
|
Dario LipicarandClaude Opus 5
|
5d0a99a182 |
chore(generator): retire ApiStyle::Std (#122)
The Std surface — std-typed signatures over a QVariant + LogosAPIClient body — no longer had a caller. `interface: "universal"` modules moved to `lp` (std types over the Qt-free logos-protocol C ABI), and nothing else ever selected it, so every Std branch was dead weight sitting in front of the two live ones. `--api-style=std` is now rejected with a message naming the retirement rather than aliased to `qt`. A stale caller that still passes it wants std signatures; handing it the Qt surface would fail later, further from the cause. The collapse is deliberate about the branches where Std was tested BEFORE Qt, since a naive "delete the block containing ApiStyle::Std" changes Qt output: - makeHeader's include block tested Std first, so its `else` is the Qt include list — the Qt includes are kept and promoted, not deleted. - recordToWireExpr / recordFromWireExpr returned the Qt map form from a guarded `if` and the Std form from the function's trailing `return`. The guard is dropped and the Qt form promoted to the tail; deleting only the trailing return would have left a path falling off the end. - The private-member `else if (!events.isEmpty())` arm reads as an event test but was Std-only; the Qt arm (m_eventReplica + m_eventSource) survives, so setEventSource/trigger still have their storage. - `if (apiStyle == Qt || !events.isEmpty())` is a disjunction, not an Std branch: it unwraps to an unconditional emit, keeping ensureReplica()'s declaration next to its definition. - `isRec || style == Std` loses only the right disjunct — dropping `isRec ||` would double-wrap record fields in QVariant::fromValue. mapParamTypeStd / mapReturnTypeStd / isStdRefType stay: they are the shared std type table that ApiStyle::Lp reaches through the non-Qt arm of paramTypeFor / returnTypeFor / byRefFor and directly from lpPushExpr / lpFromJsonExpr. Verified by output equivalence rather than by the build succeeding: the generator was run over 13 fixture cases (the full_api contract as both a bound interface and a baked dep, three record-bearing contracts incl. map-of-record fields, the chat module's production contract, and a no-events contract) for both qt and lp, before and after. `diff -r` over the 194 resulting files reports no differences, and the experimental --lidl backends are byte-identical too. Test suite: 180/180, unchanged. Co-authored-by: Claude Opus 5 <noreply@anthropic.com> |
||
|
Dario LipicarandClaude Opus 5
|
8e7ed6e0ec |
fix(generator): the provider dispatch decodes arguments, it does not coerce (#121)
The --provider-header dispatch emitted args.at(0).toULongLong() and friends, so echoUint(-1) reached the author's method as 18446744073709551615 and echoInt(3.7) as 4. It now emits logos::qtArgFromVariant<T> per parameter — the same canonical codec the cdylib dispatch and the Rust provider use — and turns a logos::CodecError into the canonical dispatch_failed object. toQVariantConversion is untouched: it is also used for the Qt CONSUMER wrapper's return conversion, which is a different direction. The new toProviderArgDecode covers only the incoming-argument job, and falls back to the old conversion for a type the codec has no rule for so a module author's own type still compiles. Co-authored-by: Claude Opus 5 <noreply@anthropic.com> |
||
|
Dario LipicarandClaude Opus 5
|
9a6d7b8228 |
fix(generator): one Qt type mapper, and it knows about void; LpClient takes a timeout (#119)
Two near-duplicate LIDL->Qt type mappers existed — legacy/main.cpp's
lidlTypeExprToQtTypeName and experimental/lidl_emit_common.cpp's lidlTypeToQt —
and they disagreed. The legacy one had no `void` case, so a `-> void` method
arriving as Primitive("void") from the impl-header parser fell through to
QVariant. (The .lidl parser spells it Named("void"), which survived only by
accident, through mapReturnType's `base == "void"` early-out.)
That was not a Qt-consumer bug: the std/lp tables are DERIVED from this name, so
the same method generated `LogosMap doVoid(...)` on the Qt-free surface too.
Measured, from `void doVoid();` in a .h interface:
QVariant doVoid(...) --api-style qt before
void doVoid(...) --api-style qt after
LogosMap doVoid(...) --api-style lp before
void doVoid(...) --api-style lp after
lidlTypeExprToQtTypeName is now a delegation, so there is one table to disagree
with. This changes the generated signature for any module consuming a `-> void`
method through a .h interface; the two in-tree call sites discard the value and
are unaffected.
logos::LpClient::invoke/invokeAsync gain a timeout_ms parameter, defaulted to
the C ABI's "use the default" (0) so no existing caller changes. The Qt-typed
consumer surface takes a Timeout on every async overload and had nowhere to put
it — a wrapper delegating to the lp path silently dropped it.
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
|
||
|
Dario LipicarandClaude Opus 5
|
461cfed52d |
refactor: the LIDL codec exists once (#117)
* refactor: the LIDL codec exists once
The cdylib generator emitted its own copy of the codec — ~186 lines of
C++-emitting-C++ mirroring logos-protocol's logos_codec.h by hand. Every codec
fix had to be written twice or it silently only half-applied, which happened
twice in a row recently (routing scalars through the codec + signedness; then
accepting 3.0 while still rejecting 3.7).
It was worse than duplication. The two copies had DRIFTED — the emitted integer
decode gated on is_number() where the canonical one checked is_number_integer()
|| is_number_unsigned() — and logos_json.h's byte helpers were the same mangled
symbols with weak linkage and DIFFERENT bodies as logos_codec.h's, both reaching
one program (module TUs compiled one; liblogos_protocol.a carries TUs that
included the other). Which body won was down to link order.
logos_json.h goes back to its documented charter — "LogosMap/LogosList aliases
for impl classes", per its own CMakeLists — and loses 77 lines. jsonToBytes moves
beside its sibling jsonToStringVec in logos_lp_client.h, rebuilt on the canonical
isTaggedBytes/b64UrlDecode; it keeps its own narrow spelling because every lp
decoder is documented to yield the default-constructed value on a mismatch,
which neither bytesFromJson (throws) nor bytesFromJsonLenient (accepts more) does.
Emptying it rather than making it include logos_codec.h is deliberate: some
thirty alias-only include sites across the module repos get ZERO new includes,
and logos-cpp-sdkConfig's "only dependency is nlohmann_json" stays true.
With the clash gone the generic half is deletable. emitGeneratedCodec becomes
emitRecordCodecs: one logos::detail::Codec<::Rec, void> per declared record, and
nothing else. That residue is irreducible — a LIDL `type` is a per-contract
struct whose fields exist only in that module's header, and C++17 has no field
reflection. Nesting composes for free: Codec<std::vector<Blob>> and deeper come
from the shared half once Codec<::Blob> exists.
One asymmetry dies with it. The scalar bstr decode and the [bstr] element decode
were different functions with different strictness, so echoBytes("hi") succeeded
while echoBytesList(["hi"]) threw — inside one module, for the same type. They
are one function now.
Build wiring: ONE line, in this repo's own test CMake, using a variable
nix/tests.nix already supplies. Nothing in logos-module-builder, logos-qt-sdk, or
any module repo.
verified: cpp-sdk + protocol suites green; test_fullapi_cpp, test_fullapi_ext_cpp
and test_basic_module_cpp build; test-modules 176/176. Conformance delta is
exactly one cell, baselined first in logos-test-modules#31.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* chore: bump logos-protocol to the path-threaded bstr decoder
logos-protocol 4359557 (#33). Required by this branch, not incidental: deleting
the emitted codec swaps its path-carrying bstr decode for the canonical one, and
without #33 the canonical one reported "at value" instead of "[0].payload" —
losing the diagnostic exactly where a malformed bstr is hardest to find.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
|
||
|
Dario LipicarandClaude Opus 5
|
c364133066 |
fix(cdylib): the emitted codec accepts a whole-valued float as an integer (#116)
Mirrors logos-protocol fix/integral-float-accept. The generator emits its OWN codec into <name>_types.h, so the rule has to be applied twice or half the platform disagrees — the same split that made the original scalar fix need two sites. #115 made the emitted integer codecs check signedness, and in doing so they started rejecting 3.0 as well as 3.7. Four test_basic_module_cpp cases pass a whole-valued double where the contract declares an integer, and they are right: JSON does not distinguish 3 from 3.0, and this same codec accepts an integral number for float64 on exactly that reasoning. A float now decodes as an integer when it has no fractional part and fits. 3.7 is still refused. Also adds <cmath> to the emitted include set for std::modf. verified: test-modules 176/176 with the four cases restored, conformance matrix unchanged at 170 pass / 2 xfail. Co-authored-by: Claude Opus 5 <noreply@anthropic.com> |
||
|
Dario LipicarandClaude Opus 5
|
ff8c3003a4 |
fix(cdylib): typed scalars go through the codec, and the emitted codec checks signedness (#115)
* fix(cdylib): typed scalars go through the codec, and the codec checks signedness
The cdylib dispatch decoded composites with the generated codec but scalars with
a bare nlohmann accessor. Two silent conversions lived in that gap:
echoUint(-1) -> 18446744073709551615 (.get<uint64_t>() wraps)
echoInt(3.7) -> 3 (.get<int64_t>() truncates)
The Rust provider rejects both. So a contract both providers share answered
differently depending on which one a consumer resolved to, and one of the two
answers was a sign flip on a nominal value.
The reason this was left in place was circular, and it was written in the source:
the leniency "is pinned by the conformance matrix (`hostile/int/fractional`
expects 3 from 3.7 on this provider)". Those cells exist to DOCUMENT the
divergence — their own `why` text says the strict behaviour is correct. The
expectations moved with this change.
TWO sites, because fixing one relocates the bug rather than closing it:
* jsonArgToStd no longer special-cases int/uint/float64/bool/tstr — everything
typed goes through Codec<T>. `any` still passes through, since it declares
nothing to check against; bstr keeps its tagged-bytes decoder.
* the EMITTED codec (this generator writes its own copy into <name>_types.h,
separate from logos_codec.h) gated integers on `is_number()`, which admits
floats AND negatives. Routing scalars into it without fixing it would have
changed nothing. The integer specializations are now spelled out rather than
driven from the scalar table, because a category check is not enough for them.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* chore: bump logos-protocol to the signedness + sentinel fixes
logos-protocol c0df466 (#31):
* Codec<T> checks integer signedness and range, so a negative can no longer
wrap into an unsigned and a wide value can no longer truncate.
* the pending-call sentinel is matched by shape rather than key presence, so
a user map merely carrying that key no longer hangs the call.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
|
||
|
Dario LipicarandClaude Opus 5
|
3d322bd315 |
fix(codegen): LIDL int/uint are 64-bit in the Qt spelling too (#113)
* fix(codegen): LIDL int/uint are 64-bit in the Qt spelling too lidlTypeToQt mapped BOTH `int` and `uint` to plain `int`. Everywhere else in the stack a LIDL int/uint is 64-bit — int64_t/uint64_t in C++ impls, i64/u64 in the Rust SDK — so the Qt spelling broke the one-type-per-LIDL-type rule and lost data: a Qt consumer reading a `uint` return got a SIGNED 32-bit value, so anything above 2^31 came back wrong and anything above 2^63 was never expressible. int -> qlonglong, uint -> qulonglong, and returnConversion() gains the matching accessors (toLongLong / toULongLong instead of toInt). qlonglong/qulonglong rather than qint64/quint64 so the generated introspection JSON uses the same names Qt's own metaobject normalisation produces — otherwise a cdylib module's generated `signature` and a legacy module's QMetaObject-derived one would disagree for the same LIDL type. Nothing looks these strings up: the only QMetaType::fromName call in the stack is for "LogosResult". This changes two generated surfaces: the Qt consumer wrapper signatures and the introspection JSON. Passing an int argument still converts implicitly, so callers keep compiling; code that assigns a wrapper's return into an `int` narrows and may warn, which is the bug being surfaced rather than a regression. Tests: 168/168, with the type-mapping and client-emitter expectations updated to the 64-bit spelling. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> * feat(records): typed C++ structs for Qt consumers Completes the Qt half of the type mapping this branch started. A `type Foo { … }` in a contract now generates a real struct in the client header, so a consumer writes `Status s = client.makeStatus();` instead of digging fields out of a QVariantMap. One LIDL type, one type per language. lidlTypeToQt - Named -> the record's struct (was QVariant) - [Record] -> QList<Record>, {tstr: Record} -> QMap<QString, Record>. QVariantList CANNOT hold a record without Q_DECLARE_METATYPE, and a typed list is the point. client emitter - struct + inline ToVariant/FromVariant per record, emitted before the class; conversions come after all structs so records may reference each other. Recursive, so a field may itself be [Status] or {tstr: bstr}. - records pass by const&, decode on return, and convert at the call site (sync and async) bstr fields are QByteArray on purpose: logos-protocol's QVariant<->JSON conversion already materialises the canonical {"_bytes": base64url} form as a QByteArray and back, so the record conversions stay pure field mapping and binary survives at any depth with no record-specific bytes handling. Verified by COMPILING and RUNNING the generated code, not just asserting on text — the string tests would not have caught either bug this found: [Record] first mapped to QVariantList (appending a Status to it does not compile) and the decode lambdas shadowed their accumulator. Extracted the emitted record block for a contract with a nested record and a bytes field, compiled it against Qt6Core, and round-tripped Batch -> QVariant -> Batch asserting items[0].port, the QByteArray blob and the label all survive. Exit 0. The LidlTypeToQt.NamedType expectation flips from "QVariant" to the struct name, which is the behaviour change. Tests: 169/169. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> * fix(codegen): teach the lp/std consumer wrappers the 64-bit spellings Caught while checking whether the SDKs are ready for the `any` migration, and it is a regression THIS BRANCH would otherwise have shipped. legacy/generator_lib.cpp generates the lp/std consumer wrappers a universal module uses to call its dependencies. It matches type names against an allow-list and falls back to QVariant for anything else: static const QSet<QString> known = { "void","bool","int","double","float","QString", … }; if (known.contains(base)) return base; return QString("QVariant"); Once lidlTypeToQt reports `qlonglong`/`qulonglong`, every LIDL int/uint method misses that list — so a typed `int` parameter would have silently become an opaque QVariant in those wrappers. Worse than the truncation this branch set out to fix, and invisible until someone read the generated header. Adds the two spellings to both allow-lists, plus the conversions they imply: QVariant->Qt (toLongLong / toULongLong), the std spellings (int64_t / uint64_t), the QVariant->std return path, the Qt-style return, and the default-value case. The existing `int` entries stay for legacy Qt plugins, whose QMetaObject still reports `int` for a 32-bit parameter. Tests: 171/171, with the allow-list pinned in both mapping test files — including that an unknown spelling still falls back to QVariant, so the fallback itself is not what regressed. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> * feat(records): typed structs for C++ dependency wrappers Closes the second record gap. The wrapper every C++ module actually gets for its dependencies comes from the LEGACY generator (`--dep <name>=<lidl>`), not the client-stub backend the previous commit taught about records — and there a contract's `type Status { ... }` reached the consumer as an untyped bag: QVariant on the Qt surface, LogosMap on the std/lp one. Worse than inconvenient for a `bstr` field: the caller received the canonical `{"_bytes": "..."}` envelope and had to know to unwrap it, while Rust and the stub backend handed back real bytes. Now, on all three api styles: struct Status { uint64_t port{}; std::vector<uint8_t> blob{}; }; Status getStatus(logos::CallError* err = nullptr); std::string describeStatus(const Status& s, ...); std::vector<Status> listStatuses(...); The struct is NESTED in the wrapper class (`InfoModule::Status`) because a module consuming two deps that each declare a `Status` includes both wrappers into one translation unit. Conversions are file-local statics in the generated .cpp, so a Qt-free module's own TUs still never see QVariant or nlohmann. Records reach parameters, returns, event callbacks, `[Record]` and `{tstr: Record}` — at any depth, with bytes tagged throughout. Same commit, the legacy path's half of the 64-bit fix: `lidlTypeExprToQtTypeName` mapped BOTH int and uint to `int` ("wire-as-int for now"), so a `uint` method on a dep reached a Qt consumer as a signed 32-bit value and a std/lp consumer as a signed int64_t. Now qlonglong/qulonglong, matching the spelling the other half of this PR gave the stub backend. `lpFromJsonExpr` grew the uint64_t branch it needed — without it a mistyped payload THREW out of nlohmann's implicit conversion instead of defaulting like every other scalar. Verified by generating a contract with a record, a record-of-records, a `uint` above 2^32 and a high-byte `bstr`, then COMPILING the output for qt/std/lp and round-tripping the emitted conversions: - `{"_bytes":"gAH_"}` at every depth, decoding back to the same bytes - 4294967296 intact through both directions - garbage/missing fields default rather than throw That compile is what caught the one real bug here: the container decode lambdas declared `__m`/`__j`, shadowing the record decoder's own locals, so a map-of-records field read from its own uninitialized local — it compiled with nothing but a -Wuninitialized warning. Locals are `__acc`/`__src` now, pinned by a test. Also: 8 generator tests (one asserting an empty record set leaves every byte of the output as it was), 179 total green; logos-test-modules builds and tests green against this generator. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> * fix(records): async record returns decoded a default-constructed struct Two correctness holes in the record work, both silent, found while scoping the cdylib provider side. 1. The async consumer overload emitted `qvariant_cast<Status>(v)` while the sync one emitted `StatusFromVariant(_result)`. The wire delivers a QVariantMap and no Q_DECLARE_METATYPE is emitted for the struct, so the cast does not fail — it returns a DEFAULT-CONSTRUCTED Status and the caller sees empty fields with no diagnostic. The sync path being correct is what makes it bad: the same call is right or wrong depending only on which overload the caller reached for. Async now decodes field by field through the same conversion. (The legacy dependency-wrapper generator already did this correctly — this was the experimental client-stub backend only.) 2. A record whose ONLY field is a tstr named `_bytes` is wire-identical to a canonical tagged byte string: `isTaggedBytes()` is checked before `is_object()` in both logos_codec.h and logos_json_convert.cpp, so such a record decodes as bytes and the struct silently disappears. The ambiguity is inherent to the tagged form — the codec's own comment says not to name a map key `_bytes` — but the generator can refuse to emit the one shape guaranteed to misdecode instead of leaving it to be found at runtime. Both front doors (the .lidl client-stub path and the --dep/--interface path) now reject it with a message naming the type and the fix. A second field disambiguates it (isTaggedBytes requires exactly one key), so that shape still generates — verified, not assumed. 181 tests, +2. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> * feat(cdylib): records, [bstr], typed maps and nested composites on the C++ provider The C++ cdylib backend could express scalars, `[scalar]`, `any` and an untyped map. Everything else it rejected BY NAME — `[bstr]`, `[[int]]`, `{tstr: int}`, and records, which the impl-header parser could not even declare because it skipped `struct` outright. That is why the ext contract had to be Rust-only. Four changes, in the order they matter: typeSupported() recurses instead of whitelisting element names, which admits [bstr], [[int]], [Record] and [{tstr: T}] in one rule; a declared record is admitted; a map now REQUIRES a tstr key (it used to `return true` for any map and then silently flatten {int: tstr} to an untyped LogosMap, losing the key type). lidlTypeToStdCdylib() became total. Its `lidlTypeToStd` fallback answers QVariantList / QVariantMap — Qt names in a Qt-FREE translation unit — and only failed to appear because the gate rejected everything that reached it. Widening the gate made that fallback a live leak, so composites now recurse and never reach it. <name>_types.h new: the generated codec, recursive, with a FULL specialization for std::vector<uint8_t> that wins over the generic vector rule — which is what keeps a bstr tagged at any depth instead of becoming a plain array of numbers. One Codec specialization per declared record, field by field, with the field path in the error. impl_header_parser learned `struct` (two passes, because a record field may name another record and the type mapper only answers Named() for an already-registered name — one pass silently typed `Blob inner;` as `any`), std::map<std::string, T>, and recursion into vector elements so std::vector<Blob> is [Blob] rather than falling through to `any`. Records are only names the contract DECLARES: `void` is not a LIDL builtin, so `-> void` arrives as Named("void"), and treating every Named as a record is the exact trap that made the Rust generator emit `-> Void`. Two things the interface JSON got wrong, both found by running it: - it spelled a record `Blob` and a `[Record]` `QList<Blob>`. Those are the CONSUMER's names, correct in a generated wrapper where the struct exists — but this JSON is the module's getMethods(), read by the host to marshal a QVariant, and there is no metatype called `Blob`. The host SIGSEGV'd on the first call to any record method. A record IS a variant map at that boundary; lidlTypeToQtWire() says so. - the types header emitted the structs. Header-first, the author owns them and the contract was derived from those very declarations, so it was a redefinition. It emits forward declarations and the codec. Also: `jsonReturn` is set by the front end for any map return, which no longer implies the C++ type IS nlohmann::json now that a typed map is std::map<std::string, T> — checking the flag before the spelling emitted `result.dump()` on a std::map. The spelling decides. Scalars keep their nlohmann accessor verbatim rather than routing through the codec: `.get<int64_t>()` TRUNCATES a float instead of throwing, and the conformance matrix pins that leniency (hostile/int/fractional expects 3 from 3.7). Changing it would silently move behaviour something depends on. The pinned-rejection test for [bstr] is INVERTED rather than deleted — the cell it pinned still matters, only its answer changed — plus new tests for the non-tstr map key rejection and for declared-vs-undeclared records. 183 tests. Every existing module still builds; test_fullapi_cpp is unchanged. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> * fix(records): only structs the API mentions become contract types Teaching the impl-header parser to read `struct` (previous commit) published EVERY struct in a header as a contract `type`. Two production modules already carry private helpers: openmetrics-module struct ModuleSource (namespace scope, internal) logos-package-manager struct PendingAction (PRIVATE, inside the class) Both were being published — a module's interface changing as a side effect of an internal refactor, which is not something deriving a contract from a header may do. PendingAction was published WRONG as well: its fields carry trailing `// comments`, the field regex requires a line ending in ';', and the unmatched fields were silently dropped. A record with a partial field list is worse than no record, because it looks like a contract. A struct now earns its place by appearing in a method or event signature — transitively, since a published record's own fields may name others. Verified on the real headers: package-manager and openmetrics publish zero types again, while the ext provider keeps both Blob and Wrapper (Wrapper is reachable only through Blob's use in a signature). Trailing comments are stripped before the field match, so no field is dropped. Two tests over a fixture carrying both an internal namespace-scope struct and a private in-class one; 185 tests. test-modules and openmetrics both rebuild. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> * docs(doctests): the generator round-trip now shows 64-bit ints and typed records The two failing assertions in cpp-sdk-generator-roundtrip were documentation asserting the OLD behaviour, and both changes are the point of this branch: int record(int id, …) -> qlonglong record(qulonglong id, …) QVariant translate(QVariant p) -> Point translate(const Point& p, …) `record`'s `id` is a `uint64_t` in the impl header, so the old signature was handing a caller a SIGNED 32-bit value for a `uint` — the doc showed the bug. The surrounding prose was wrong too, not just the expectations, so both blocks are rewritten rather than patched: * Flow 3 now states the mapping as int->qlonglong / uint->qulonglong and says why (LIDL int/uint are int64_t/uint64_t in every other binding), pointing at `record` as the worked example. * The composite section claimed "records and optionals surface as QVariant". Records now generate a struct, `[Point]` a QList<Point> and `{tstr: Point}` a QMap<QString, Point>; maps of `any`, optionals and bare `any` still cross untyped and stay QVariant/QVariantMap — a record has a declared shape, those do not. The new text draws that line explicitly. Expectations added for `struct Point` and `Point bounds(const QList<Point>&…)` so the record path is pinned in the doc, not just described. Verified the way CI runs it — `--release-for logos-cpp-sdk=feat/qt-64bit-numerics`, which is what makes `{release}` resolve to this branch instead of master: 10 passed, 0 failed. (A plain local run builds master and is not representative — that is why it still showed the old signatures.) outputs/ regenerated; the diff also picks up unrelated pre-existing drift where the committed Markdown had fallen behind the spec. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5 <noreply@anthropic.com> |
||
|
Dario LipicarandClaude Opus 5
|
8a40a98322 |
feat(cdylib): support [bstr] — arrays of byte strings (#111)
A universal module could take or return a single blob (`bstr`) but not a list
of them: `std::vector<std::vector<uint8_t>>` parsed to `[bstr]`, which the
cdylib gate rejected by name. Authors had to flatten to hex or base64 strings
by hand. logos-execution-zone-module hit this on
send_generic_private_transaction(..., program_dependencies), whose dependency
ELFs are exactly a list of blobs.
The gate excluded `[bstr]` because the array path decodes with a blanket
`expr.get<inner>()`, and `[bstr]` elements arrive as the canonical tagged
{"_bytes": base64url} OBJECT — nlohmann refuses that, and a number-array
element would silently bypass the base64 decode. So admitting the type needed
a per-element codec, not just a whitelist entry.
Adds one, on top of the scalar codecs already emitted:
- lidlBytesListFromJson: element-wise lidlBytesFromJson, so each element may
independently be tagged, a plain string or a number array; a non-array arg
yields an empty list instead of throwing, matching the scalar decoder.
- lidlBytesListToJson: element-wise lidlBytesToJson, so a returned or emitted
list carries the tagged form per element instead of nested number arrays
that no consumer decodes as bytes.
Wired into all three places the type can appear — method params, method
returns, event payloads — and both helpers are gated (usesBytesArray /
hasBytesArrayEventParam) so a module that never carries a byte-string array
gains no unused static function, matching how the scalar encoder is gated.
Nothing outside this generator needed changing: lidlTypeToStd already spelled
`[bstr]` as std::vector<std::vector<uint8_t>>, the wire form is protocol's
existing tagged-bytes encoding, and consumers see `[bstr]` as QVariantList
exactly like `[int]` — with nested QByteArray preserved through
qvariantToNlohmann since logos-protocol#23.
Tests: 171/171. New coverage for the param decode, the return encode, the event
encode, and the unused-helper gating; the test that enshrined the rejection is
now an eligibility + tagging assertion.
Verified end to end, not just as generated text — a module with `[bstr]` as
param, return and event payload, driven through logoscore over the real
transport:
param : json:[{"_bytes":"AH-A_w"},{"_bytes":""},{"_bytes":"3q2-7w"}]
-> "3|4:0:255:510|0👎-1:0|4:222:239:824" (byte-exact; the 0x80
and 0xff bytes survive, and the empty element stays an element)
return : [{"_bytes":"AH-A_w"},{"_bytes":""},{"_bytes":"3q2-7w"}]
event : {"arg0":[{"_bytes":"AH-A_w"},{"_bytes":""},{"_bytes":"3q2-7w"}]}
And lez_core's real header now generates, decoding program_elf with the scalar
codec and program_dependencies with the list one.
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
|
||
|
Dario Lipicar
|
e8966bf7a9 |
fix(codegen): correct [T]-array arg packing and any-value return in the Qt client (#105)
* fix(codegen): pack Qt client args as one element each, not a spread list
The generated Qt client wrapper packed a method's arguments with
`QVariantList{a, b, ...}` (sync) and `QVariantList{...}` / `QVariantList() << a`
(async). For a QVariantList-typed argument -- every `[T]` list type (`[any]`,
`[int]`, `[uint]`, `[float64]`, `[bool]`) -- a braced `QVariantList{v}` and
`<< v` both CONCATENATE the list's elements into the args list, so
`echoList([1,2,3])` went out as three positional args instead of one array arg.
The receiver saw an arg-count mismatch and the list round-tripped empty; through
a UI->proxy->provider 2-hop it hung the call outright. This is the long-standing
"typed arrays empty over the Qt path" bug.
Fix both generators that emit the Qt client:
- legacy `generator_lib.cpp` (the production `logos-cpp-generator`): wrap each
arg in `QVariant::fromValue(...)` in the sync and async call sites.
- experimental `lidl_gen_client.cpp`: route both paths through the existing
`packVariantList` helper (which already wraps with `QVariant::fromValue`), the
same helper the event `trigger` path uses.
`QVariant::fromValue` does not double-wrap an already-QVariant (`any`) arg, and
scalars/QString/QVariantMap/QByteArray were never affected (they don't
concatenate). Empirically: `QVariantList{v}` / `<< v` give size 3 for a 3-element
list; the wrapped forms give size 1.
Tests: legacy generator_tests gain ListArgWrappedAsOneElement and update the
param-packing assertions to the wrapped form; experimental gains
ListArgIsPackedAsOneElement. 167/167 green.
* fix(codegen): pass `any` (QVariant) return through raw in the lp wrapper
The Qt-free (lp) client generator maps both `any` (QVariant) and the `{tstr:any}`
map (QVariantMap) to the same `LogosMap` std type, and decoded a `LogosMap`
return as `jv.is_object() ? jv : LogosMap::object()`. For a genuine map that is
a no-op, but for `any` it collapsed every NON-object value (a string, a number,
an array) to an empty object `{}`. A universal proxy forwarding `echoAny("x")`
through this wrapper therefore returned `{}` instead of `"x"` — the concrete
cross-version blocker (the UI's runMethods verified echoAny and got FAIL:echoAny
through the 2-hop).
`lpFromJsonExpr` still receives the original Qt type, so it can tell `any`
(mapReturnType == "QVariant") from the map (QVariantMap): pass `any` through
unchanged, keep the object coercion only for the map.
Test: MakeSourceTest.LpAnyReturnPassesThroughButMapForcesObject.
With this + the arg-spread fix, a UI drives the full method surface (incl.
echoAny and every array type) through a universal proxy 2-hop end to end.
|
||
|
Dario LipicarandClaude Opus 4.8
|
2f34804948 |
fix(codegen): handle bstr + composite-any types in cdylib + Qt wrappers (#103)
Three type-handling gaps surfaced by a module that exercises the full type surface (every method param/return + event param type): 1. cdylib method-param decode used lidlTypeToStd() for array params, which falls back to Qt containers (QVariantList) for [any] — undeclared in the Qt-free cdylib TU. Use the Qt-free lidlTypeToStdCdylib() so [any] decodes as LogosList. (lidl_gen_cdylib.cpp) 2. The Qt sync wrapper had no QByteArray return case, emitting a bare 'return _result;' (QVariant) for a bstr return — no implicit QVariant -> QByteArray conversion. Add toByteArray(). (generator_lib.cpp) 3. The Qt event-callback arg converter had no QByteArray case, so a bstr event param was delivered via .toString() to a std::function<void(QByteArray)>. Add toByteArray(). (generator_lib.cpp) Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com> |
||
|
Dario LipicarandClaude Opus 4.8
|
c7444bc29a |
Follow-ups to #100: lp-consumer bstr decode, Qt-free cdylib event types, binary-event coverage (#102)
* cdylib events: Qt-free types, and drop the unused bytes encoder Three follow-ups to the bstr event fix, all in the cdylib events sidecar -- a Qt-FREE translation unit: - An `any`/map event parameter was emitted as a bare QVariant/QVariantMap, which does not compile there. Spell those as their nlohmann aliases (LogosMap / LogosList) and pull in <logos_json.h> when they appear. - std::vector<std::vector<uint8_t>> fell through the impl-header parser's unknown-type fallback to `any`, so the cdylib gate admitted it and the generator then emitted QVariant. Parse it as `[bstr]` so the gate rejects it with a message naming the offending parameter. - The bytes encoder was emitted into every module's sidecar, leaving an unused static function (-Wunused-function) wherever no event carries binary data. Emit it only when a bstr event parameter exists. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * lp consumer: decode bstr into std::vector<uint8_t> The Qt-free (`lp`) consumer wrappers -- what every universal C++ module gets for its dependencies -- had no QByteArray in their type tables, so a `bstr` event parameter, method argument, or return degraded to QVariant and then to LogosMap. A consumer subscribing to a binary event was handed the raw tagged JSON object {"_bytes": "<base64url>"} instead of the bytes, with no generated decode. Teach the tables about QByteArray (-> std::vector<uint8_t>) and marshal it through the canonical tagged form in both directions: logos::bytesToJson on the way out, logos::jsonToBytes on the way in. Those live in logos_json.h -- Qt-free and protocol-free, so the generated wrappers and module code can share them. The Qt apiStyle already did this via QByteArray::toBase64/fromBase64. Without this, a subscriber written the obvious way -- onBinaryReady([](const std::string&, const std::vector<uint8_t>& payload) {...}) -- compiles (nlohmann::json has an implicit conversion operator) and then throws at runtime on every event, so the callback body silently never runs. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * tests: cover binary event payloads by value, not just by source text The regression test for #99 asserts on generated source text, so it stays green against an encoder that emits the wrong bytes. Add the value-level half: - tests/sdk/test_logos_json_bytes.cpp exercises the canonical tagged-bytes codec against the RFC 4648 vectors, the URL-safe alphabet, every len%3 tail group, embedded NULs and high bytes, a 109,447-byte payload (the size from #99), and the lenient/padded decode paths. - tests/experimental/test_lidl_gen_cdylib.cpp additionally pins the Qt-free spelling of JSON event payloads, the rejection of [bstr], and the omission of the bytes encoder from modules whose events carry no binary data. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * doctests: prove a binary event payload survives the round trip Neither doc-test covered bytes-in-an-event -- the gap #99 fell through. The generator round-trip carried `bstr` only as a method argument and return, and the composition doc-test, which is the one that actually runs two modules under logoscore and subscribes to an event, carried only a string. So a generator that dropped every bstr event argument kept both of them green. - cpp-sdk-module-composition: greeter_module gains a `blobReady(label, payload)` event and an `emitBlob(size)` method; orchestrator_module subscribes and reports the length AND a checksum of what it received. Length alone would not catch a corrupted payload -- a wrong alphabet round-trips to the same size. - cpp-sdk-generator-roundtrip: sensor_module gains a `capture(id, frame: bstr)` event, and a new step shows the generated event body encoding it through lidlBytesToJson rather than pushing it raw. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * logos_json.h: include <cstddef> for size_t The tagged-bytes codec uses size_t but relied on it arriving transitively through the other includes. Include <cstddef> directly so the header is self-contained. (Copilot review, PR #102.) Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com> |
||
|
Santiago Galván
|
b60b230e66 | Fix binary payloads in cdylib events (#100) | ||
|
Dario LipicarandClaude Opus 4.8
|
aea29d3797 |
Per-module concurrent dispatch: C++ module async export (#93)
* feat: emit logos_module_dispatch_async for concurrency:multi C++ modules The cdylib C-ABI exports gain an async dispatch entry (each call run on a worker thread, reply on completion) for universal + cdylib C++ modules. --concurrency multi flag in logos-cpp-generator. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * docs: cpp-sdk concurrent-dispatch doctest (concurrency:"multi" showcase) A concurrency:multi C++ worker + a single driver firing concurrent calls, showing the multi worker overlaps them. The C++ cdylib generator needs NO change — its logos_module_dispatch is already safe to call concurrently; the worker pool lives in the Qt glue and the result is deferred via a sentinel + completion event. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * fix: wire universal-cdylib modules() independent of the context latch A C++ interface:"universal" cdylib that calls another module via modules().<dep>... segfaulted on its FIRST cross-module call: the typed dependency surface (LogosModules) was wired inside lidlTryFireContext, which returns early when no persistence context was stored (g_ctxStored == false). When the daemon never delivers a context (observed: zero set_context calls for a context-less module), maybeSetLogosModules never ran, m_logosModulesPtr stayed null, and LogosModuleContext::modules() dereferenced null. modules() does not need the context — each dependency client bakes its target+origin at codegen time and creates its lp client lazily on first call. So wire it in its own context-independent once-latch (lidlEnsureModulesWired), called at the top of lidlTryFireContext before the context-gated early return, i.e. on the first dispatch / set_context / set_emit_callback. A module with deps but no stored context now has modules() wired before any handler runs. (Bump the concurrent-dispatch doctest's post-daemon-start sleep 3 -> 6 to match the rust spec's cold-start margin.) Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * test: green the universal-cdylib concurrent-dispatch doctest + wire into CI The driver/worker split a declarations-only impl header (so the cpp-generator's --header-to-lidl doesn't choke on inline std calls) from the impl body. That body was never compiled — metadata's nix.cmake.extra_sources is parsed but not consumed by the LogosModule.cmake the build actually uses — so the impl symbols (FanoutDriverModuleImpl::fanOut / ::peak) were UNDEFINED in the dylib and the plugin null-jumped (bl -> 0x0) on the first cross-module call. Pass the impl .cpp via logos_module()'s existing SOURCES argument so it's compiled and linked. With this the cpp universal-cdylib reaches worker peak overlap 4 end-to-end (a single-threaded driver fans out 4 async calls into a concurrency:"multi" worker and all four overlap), matching the Rust half. Wire the spec into doctests.yml so the workspace pipeline runs it. (Auto-wiring metadata.extra_sources — so the split pattern works without listing SOURCES by hand — needs the consumer added to the backend LogosModule.cmake copies in logos-plugin-core / logos-plugin-qt; tracked separately.) Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * chore: bump logos-protocol to merged master (protocol#5) logos-protocol 9de4165 → 4ea32a3 (concurrent-dispatch handshake coalescing, now on master) Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com> |
||
|
haeliusandÁlex
|
f032cf291e |
fix(header-parser): join multi-line method declarations before parsing (#91)
* fix(header-parser): join multi-line method declarations before parsing * strip qualifiers/attributes before return-type matching (#92) --------- Co-authored-by: Álex <alex93cabeza@gmail.com> |
||
|
Dario LipicarandClaude Opus 4.8
|
182d850e72 |
docs: update project.md for the logos-lidl frontend split (#90)
The embedded LIDL frontend (lidl_ast/lexer/parser/serializer/validator) was deleted — it now lives in the logos-lidl repo, linked via find_package and bridged onto the Qt backends by experimental/lidl_compat.h. Update the project doc to match: describe the consumed logos-lidl frontend + the compat shim, list the backends cpp-generator actually keeps (lidl_emit_common, lidl_gen_client, lidl_gen_cdylib, impl_header_parser), drop the deleted frontend files and their (removed) tests, and note the qt provider glue lives in logos-qt-generator and the Rust backend in logos-rust-sdk. Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com> |
||
|
Dario LipicarandClaude Opus 4.8
|
1bc101df1f |
feat: cpp-generator consumes logos-lidl; delete embedded frontend (#89)
* feat: cpp-generator consumes logos-lidl; delete embedded frontend The canonical LIDL frontend now lives in logos-lidl. cpp-generator links it and keeps only the C++/Qt-specific parts (impl-header parsing, the gen_client/ gen_cdylib backends, the Qt type-name mapping). - Delete the embedded lidl_lexer/parser/serializer/validator/ast. - Add experimental/lidl_compat.h: brings logos-lidl's std AST into the global scope the backends use (via `using`), a qs() std::string→QString helper, a QTextStream<<std::string overload, and name-compatible shims (lidlParse/ lidlSerialize/lidlValidate) so the emission code keeps compiling. - Re-point impl_header_parser, lidl_gen_client (+ Doxygen /// docs on the generated client methods), lidl_gen_cdylib, lidl_emit_common, and legacy/ main at lidl::ModuleDecl. - CMake: C++17 + find_package(logos-lidl) + link logos-lidl::logos_lidl. - bin.nix: distribute only the shared C++/Qt backend helpers (compat + impl_header_parser + emit_common) under share/lidl-frontend, not the frontend. - tests: drop the 4 frontend test files (covered by logos-lidl now); the backend tests link logos-lidl. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * fix: pin logos-lidl to the C-ABI commit + lock it The logos-lidl input was declared in flake.nix but missing from flake.lock, so override chains that don't reach the nested input (the doctest harness building a scaffolded module) couldn't resolve it. Pin the branch rev and lock it so the component is self-contained. Re-point at master once logos-lidl lands. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * chore: re-point logos-lidl to merged master (#5) --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com> |
||
|
Dario LipicarandClaude Opus 4.8
|
676154070c |
codegen: Qt-free outbound — ApiStyle::Lp typed wrappers over the lp_* C ABI (#88)
* ci: drop doctest chain pins — the qt-split chain is fully merged logoscore-cli and module-builder masters now contain the chain; the temporary --release-for pins (added so stacked-branch CI could resolve compatible cross-repo revs) default back to latest releases. * codegen: Qt-free outbound — ApiStyle::Lp typed wrappers over the lp_* C ABI Adds a third generator flavor (ApiStyle::Lp, --api-style lp) whose typed dependency wrappers + LogosModules umbrella call the logos-protocol C ABI directly via a new header-only logos::LpClient, instead of LogosAPIClient. This lets a module make outbound typed calls and event subscriptions with NO Qt in its translation units — Qt stays confined to the QRO transport (inside logos-protocol) and the generated plugin glue. - cpp/logos_lp_client.h: header-only logos::LpClient (lazy lp_client_create on a baked origin; invoke / invokeAsync / subscribe; std<->nlohmann JSON; CallError out-param) + RAII logos::LpSubscription (unsubscribes on drop) + json<->std helpers. The C++ analog of rust-sdk PluginProxy. - generator: makeHeaderLp/makeSourceLp emit the Lp wrappers; the Lp umbrella drops the LogosAPI ctor and bakes this module name as the lp_client origin (LogosModules() default-constructible). Qt/Std emission is byte-unchanged (dispatch added at the top of makeHeader/makeSource). Verified: generator builds; generated wrappers + umbrella compile to .o with ONLY cpp-sdk + logos-protocol headers + nlohmann (no Qt); cpp-sdk tests pass. * cdylib: wire the Qt-free typed dependency surface (modules()) into the impl When a cdylib module declares dependencies, the generated exports now include the Lp umbrella (logos_sdk.h) and construct LogosModules() + maybeSetLogosModules on the impl just before onContextReady — so the author can call modules().<dep>... and subscribe to dep events from a Qt-free cdylib. Guarded on module.depends so dependency-less cdylib modules are byte-unchanged. The umbrella + dep wrappers themselves are produced by the --general-only --api-style lp generation; feeding the dep .lidl files into that during the module build is the remaining build-system wiring (module-builder + plugin-qt dep resolution). * cdylib: wire modules() unconditionally (deps come from metadata, not the .lidl) The umbrella wiring was guarded on the .lidl module.depends, but a cdylib module declares its dependencies in metadata.json#dependencies — the .lidl contract.depends is typically empty — so modules() was left null and a typed outbound call segfaulted. Always include the generated logos_sdk.h umbrella and maybeSetLogosModules(impl, new LogosModules()) before onContextReady; the overload is a no-op for context-less impls and the umbrella codegen emits an (empty) logos_sdk.h for every cdylib, so this is safe in all cases. * fix(headers): ship logos_lp_client.h in the include/cpp source-export root A cdylib module's generated dep wrapper includes "logos_lp_client.h" and, transitively, "logos_result.h". The wrapper is compiled with the cpp-sdk source-export include root (include/cpp), so logos_lp_client.h must sit beside logos_result.h there — a quoted include resolves siblings relative to the including file's directory. Previously logos_lp_client.h shipped only at the top-level include/ (the CMake-export layout via cpp/ CMakeLists.txt), so it pulled in include/logos_result.h while the impl's logos_module_context.h pulled include/cpp/logos_result.h. Those are two distinct realpaths under the symlinkJoin, so #pragma once could not dedup them and StdLogosResult was redefined. Install every std header into both roots so a single TU only ever sees one logos_result.h. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * fix(cdylib): route logos_module_accept_token into the protocol TokenManager The generated module-impl export stored accepted tokens in a process-local std::map (g_tokens) that nothing ever read, so a cdylib module's OUTBOUND lp_client (modules().<dep>...) never saw the capability_module bootstrap token the host delivers at load. The automatic requestModule flow then ran unauthenticated: capability_module rejected requestModule, no per-target token was issued, and the cross-module call was rejected (returning a default-constructed result, e.g. 0). Forward the token into lp_token_save, which writes the same TokenManager::instance() singleton the cdylib's lp_client reads. The capability/token handshake now completes and typed Qt-free outbound calls return real results. Drop the dead g_tokens map + mutex. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * fix(lidl): exclude LogosModuleContext hooks from header-derived contracts --header-to-lidl parses an impl class's public methods. An impl commonly overrides onContextReady() (and could redeclare a context accessor) in its own public section, so the derived LIDL would include onContextReady / modules / modulePath / instanceId / instancePersistencePath. Those are framework plumbing, not API methods — and feeding them to the cdylib backend breaks cdylib-eligibility (e.g. the inherited accessors' Qt-free return-type check), which is exactly what header-first universal modules now hit. Skip the reserved LogosModuleContext names in the parser so both the Qt --from-header path and the cdylib --header-to-lidl path emit clean, API-only contracts. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * fix(cdylib): support the full std type set in header-derived contracts Routing core universal modules through the cdylib backend surfaced gaps between the cdylib subset and what the std apiStyle handled — a universal module that built under std must also build as a header-first cdylib. - lidl parser: restore the return-shape flags (resultReturn / jsonReturn) from the parsed return TypeExpr, so a header -> .lidl -> cdylib round-trip (the universal path, needed to feed the Qt glue) preserves the semantics the impl-header parser sets from C++ types (StdLogosResult -> result; LogosMap/LogosList -> json). Without this the cdylib codegen/eligibility mis-handled result / map / list returns. - cdylib eligibility + dispatch: `void` is not a lidlBuiltinType, so the parser yields it as a Named "void" (header path uses empty name) — treat both as void in the eligibility check and the dispatch (was relying on lidlTypeToQt=="void", which didn't match Named "void" -> generated an `auto result = <void call>`). - typeSupported: accept `any` (both directions), `void`/`result` (returns), arrays-of-any, and Map ({k:v}/LogosMap) — the Qt-free-via-nlohmann set. Verified: a probe with void / LogosMap / LogosList / StdLogosResult / const returns is cdylib-eligible and dispatches correctly. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * fix(lidl): carry method/event descriptions across the .lidl round-trip Header-first universal modules go header -> .lidl -> cdylib backend. The impl-header parser captures /// and /** */ doc comments into method/event descriptions, but the .lidl serializer emitted only the signature, so the descriptions were dropped — introspection (lm methods / --json, getMethods) then showed no docs (regressing the wrap-external-lib + tutorial doctests). Serialize each method/event's description as a trailing `description "..."` clause (escaped for the string literal; the lexer already decodes \\ \" \n \t) and parse it back in parseMethodDef/parseEventDef. Module description now escaped too. Verified: /// docs survive header -> .lidl -> getMethods. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * fix(lp): bound-interface wrappers are handles over umbrella-owned state The Lp interface (bind_<iface>) wrapper owned its LpClient + RAII subscriptions BY VALUE, so the idiomatic transient handle — modules().bind_calculator(p).fibonacciAsync(...) modules().bind_calculator(p).onVersionReady(...) — tore the client/subscription down when the temporary died, cancelling the async callback and the event subscription. (Sync calls completed before the temporary's destruction, so they worked; the Qt/std flavor works because its handle is thin over a LogosAPI-owned persistent client.) Make the Lp Bound wrapper a THIN, copyable handle over `State { LpClient client; vector<LpSubscription> subs; }` that the LogosModules umbrella OWNS per provider (std::map<provider, unique_ptr<State>>) for the module's lifetime. bind_<iface>(p) creates/looks up the State and returns a handle to it, so a transient handle's async/event registrations outlive it. Concrete (Static) dep wrappers are unchanged — they're already persistent umbrella members, so by-value ownership is fine there. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * fix(cdylib): lenient bytes-param decode (string / array / tagged) A universal module's bstr (std::vector<uint8_t>) PARAM arrived empty when the caller sent a plain string rather than the tagged {"_bytes": base64url} form — lidlBytesFromJson only accepted the tagged object, so byteArraySize("12345") and byteArraySize(b"\x01..") both saw 0 bytes (the return direction already worked). The std path was lenient (a QString or QByteArray arg both became bytes). Accept all three forms: a plain JSON string (raw UTF-8 bytes), an array of byte values, and the tagged {"_bytes"} form (base64url). Return direction unchanged. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * fix(cdylib): a number arg to a bytes param decodes as its decimal text byteArraySize("12345") arrives as a JSON number (the logoscore CLI's type auto-detection turns the string "12345" into int 12345), and the Qt path gives QVariant(int)->QByteArray "12345" (5 bytes). The cdylib bstr decode returned 0 for a number. Treat a JSON number as its decimal text bytes (j.dump()), matching the Qt behaviour, so a bare-number arg to a bytes param round-trips identically. Verified: byteArraySize 12345 -> 5. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com> |
||
|
Dario LipicarandClaude Opus 4.8
|
87abcd8044 |
Cdylib authoring: --backend cdylib emits the common C ABI wrapper + uniform Qt glue (#84)
* Extract the protocol layer into logos-protocol; consume it as a flake input
The transport/token/IPC layer (transports incl. QRO + plain TCP/TLS,
consumer core LogosAPIClient/LogosAPIConsumer with the capability
auto-requestModule flow, ModuleProxy, token manager, QVariant<->JSON
conversion, the abstract LogosProviderObject interface) now lives in the
logos-protocol repo behind the versioned lp_* C ABI.
This SDK keeps the typed C++ developer layer (LogosAPI, provider base
classes + Qt provider glue, module context, code generator) and still
compiles the protocol sources INTO liblogos_sdk.a from the flake input,
so the installed artifact (archive symbols, include/ + include/cpp
layouts, cmake config) stays byte-compatible: existing consumers need
no changes. Public headers are unchanged; logos_provider_object.h keeps
its name and now re-exports the abstract interface from
logos_provider_interface.h.
Transport/protocol component tests moved to logos-protocol with the
code; the remaining sdk/generator/experimental suites are unchanged
(432/432 green against the local protocol checkout).
* lock: add logos-protocol input
* Make the base SDK Qt-free: move the Qt developer layer to logos-qt-sdk
LogosAPI, LogosAPIProvider, LogosProviderBase/LOGOS_PROVIDER macros, the
QObject provider glue (QtProviderObject) and the legacy PluginInterface
(core/interface.h) move to the new logos-qt-sdk repo. The protocol
sources are no longer compiled into a monolithic archive — consumers
link logos-qt-sdk (which layers on logos-protocol) instead.
What remains here is header-only std C++: logos_module_context.h,
logos_result.h (StdLogosResult), logos_json.h — exported as the CMake
INTERFACE target logos-cpp-sdk::logos_headers — plus the code generator
(a build-time tool; its introspection mode now includes
logos_provider_interface.h from logos-protocol, where
LogosProviderPlugin moved).
Mechanically verified Qt-free: the logos-cpp-lib / logos-cpp-include
closures contain only nlohmann_json. Tests: 245/245 (module-context std
suite + generator + experimental).
* Cdylib authoring backend: --backend cdylib emits the common C ABI wrapper + uniform Qt glue
From a module's LIDL contract the generator now emits:
- <name>_module_impl.cpp — the Qt-FREE logos_module_impl.h export
wrapper (dispatch/get_methods/set_context/set_emit_callback/
accept_token/get_protocol_version/string_free) around the universal
impl class; compiled into the module's cdylib. Tagged {"_bytes"}
bytes, StdLogosResult -> {success,value,error}, context via the
existing _logos_codegen_::maybeSet* SFINAE helpers.
- <name>_events_cdylib.cpp — typed logos_events: bodies marshalling
into nlohmann::json (the cdylib flavor of the events sidecar).
- <name>_cdylib_glue.{h,cpp} — the UNIFORM Qt-plugin glue forwarding
LogosProviderObject to the C symbols; identical regardless of the
module's source language (the Rust SDK emits the same exports).
- the .lidl sidecar.
Qt-container types are rejected at generation time (a cdylib impl is
Qt-free by definition). Verified end-to-end by dlopen smoke: typed
dispatch, typed events through the emit callback, result returns,
introspection, and the protocol-version handshake — and the SAME C
harness passes against a Rust cdylib generated by logos-lidl-gen
--provider.
* fix: accept the installed source-export layout in the protocol-root check
The fail-fast only tested <root>/cpp/logos_protocol.h, but the LP_SRC
selection right below (and the error message itself) support the
installed export layout <root>/include/cpp as well. Pointing
LOGOS_PROTOCOL_ROOT at an installed export tripped the FATAL_ERROR
before that fallback could apply.
Caught by Copilot review on #82.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* lock: pin logos-protocol to the qt-free-split branch head
The Qt-free SDK (and the cdylib backend stacked on it) reference
LogosProviderPlugin from protocol's logos_provider_interface.h, which
lands on feat/qt-free-split — the P1-branch pin no longer compiles
standalone. Temporary — drop when the chain PRs merge.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* lock: pin logos-protocol to the qt-free-split branch head
The Qt-free SDK (and the cdylib backend stacked on it) reference
LogosProviderPlugin from protocol's logos_provider_interface.h, which
lands on feat/qt-free-split — the P1-branch pin no longer compiles
standalone. Temporary — drop when the chain PRs merge.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* doctest: pin the logoscore runtime via its {release} placeholder
The spec built logoscore-cli at bare master with only the cpp-sdk inputs
overridden — master's stack cannot compile against the qt-free SDK, so
the suite failed on the chain branches. With the placeholder, CI's
--release-for pins expand it to the workspace's logoscore commit (and
local runs without a pin still fall back to master, unchanged).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* doctest: override the nested module builders to {release} too
capability_module (via logoscore's lock) and the cloned accounts module
resolve module-builder from their own locks — pre-split revs whose
LogosModule.cmake still detects the SDK by logos_api.h, which the
qt-free SDK no longer ships ('logos-cpp-sdk not found'). Overriding the
builder itself to the workspace-pinned chain rev (keeping the nested
cpp-sdk override) builds both modules with the split-aware builder.
Verified end-to-end locally with the exact doctest command.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* doctest: apply the {release} + nested-builder overrides to all three specs
The runtime spec got the treatment in 210eea1; the composition and
worker-thread specs have the same logoscore/module build commands and
failed identically (pre-split builders from the modules' own locks).
All executed run: blocks now pin logoscore-cli{release} and override
the nested module builders to logos-module-builder{release}; the
displayed code_block: variants stay in their generic master form.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* generator: glue-only cdylib mode for non-C++ impls
--lidl x.lidl --backend cdylib (no --impl-class) emits just the uniform
Qt-plugin glue; the C exports come from the module's own language
backend (e.g. the Rust SDK's lidl-gen --provider). This is the entry
point logos-module-builder's cdylib interface uses for Rust modules.
* cdylib glue: root plugin implements PluginInterface
logos_host's module_initializer hard-requires PluginInterface on the
root plugin before any provider detection — same bases as the qt glue
(QObject, PluginInterface, LogosProviderPlugin). Caught by the first
host-loaded run of a cdylib-authored module; the dlopen smoke harness
exercised only the C seam.
* cdylib glue: seed the cdylib's protocol stack with the host auth token
The glue's init() reads the authToken property module_initializer now
surfaces on the LogosAPI object and forwards it across
logos_module_accept_token under the initializer's own keys
(core / capability_module). Without it the cdylib's TokenManager (a
separate static copy of the singleton) is empty and every outbound
call — including the capability requestModule bootstrap — is rejected
as unauthorized.
* cdylib glue: provider derives LogosProviderBase; tokens land in BOTH stacks
informModuleToken = base save (host-stack TokenManager — what ModuleProxy
validates inbound calls against, incl. the grant the daemon pushes after
a capability requestModule) + C-ABI forward (the cdylib's own stack, for
outbound auth). init folds into onInit; the authToken property seeding
stays.
* lock: protocol integer-fidelity fix
* codegen: typed wrappers throw on call failure; dispatch catches escapes
Generated sync client wrappers call the new err-out invokeRemoteMethod
overload and throw logos::LogosCallError when the call fails (e.g. the
bound module is missing) — previously the empty QVariant silently
degraded to the return type's default and a caller could not tell
failure from a legitimate 0 / "". Both generators (legacy + LIDL),
both API styles. Async paths unchanged.
Generated provider dispatch (universal qt glue + LOGOS_PROVIDER) wraps
the method body in a catch-all that logs and returns an invalid QVariant
— an escaped exception becomes an ordinary METHOD_FAILED instead of
unwinding through Qt event dispatch and killing the module process.
* fix: restore a clean flake.lock after the merge conflict (protocol 176fbc8)
* codegen: CallError out-param instead of throwing wrappers
Per review, the generated sync wrappers expose the error channel as an
optional trailing parameter — add(a, b, &err) — rather than throwing:
explicit, stateless, works on temporaries, and existing call sites
compile unchanged (they keep default-on-failure, now with a qWarning so
failures are visible in the module log). The dispatch catch-all from the
previous commit stays: it contains author exceptions, it doesn't
introduce any.
* generator: contract-first C++ cdylib modules from --lidl
--lidl x.lidl --backend cdylib with --impl-class/--impl-header now emits
the FULL set (C-ABI export wrapper + events + uniform glue) around the
named hand-written Qt-free impl class — the C++ mirror of declaring the
contract in .lidl and implementing the Rust trait. Without --impl-class
the glue-only mode is unchanged.
* glue: fire onContextReady AFTER modules()/event wiring
The generated onInit set the context (which fires the impl's
onContextReady hook) before constructing the LogosModules aggregate and
wiring typed event emission — so an impl doing its documented one-time
setup there (typed dependency calls, event subscriptions) dereferenced
a null aggregate and crashed the module process (signal 11). Found by
the first module to subscribe to a dependency's typed event from
onContextReady. Context now goes last.
* cdylib: fire the context-ready hook at module LOAD, not first dispatch
The impl-exports wrapper used to fire maybeSetContext (stamp + hook)
immediately inside logos_module_set_context — which the glue calls during
onInit, BEFORE the auth token is seeded and BEFORE ModuleProxy wires the
emit callback. A hook that made outbound calls or emitted events ran
half-wired; and the Rust scaffold deferred its hook to first dispatch
entirely, so a Rust module couldn't subscribe/emit/act until someone
called it — a capability gap vs the universal C++ path.
Two changes, applied uniformly:
- The glue's onInit now seeds the auth token FIRST and forwards the
context LAST (the same context-last convention as the universal onInit
ordering fix).
- The impl-exports gained a ready-latch (lidlTryFireContext): the context
is stored on set_context and the hook fires ONCE as soon as both the
context AND the emit callback have been delivered — during module
registration, before the module is published for inbound calls. Hosts
that never wire an emit callback still get the hook before the first
dispatch (requireEmit=false fallback in logos_module_dispatch).
The Rust scaffold (logos-rust-sdk lidl-gen) implements the same latch, so
on_context_ready now matches C++ onContextReady semantics: subscriptions,
authenticated outbound calls and typed emission all work from the hook at
load time.
* ci: run workflows on stacked PRs + workflow_dispatch
Both workflows filtered pull_request to master-based PRs, so stacked PRs
(feat/qt-free-sdk -> feat/extract-logos-protocol, feat/cdylib-authoring
-> feat/qt-free-sdk) ran NO checks at all. Drop the base-branch filter
for pull_request and add workflow_dispatch for manual runs. Same fix as
logos-module-builder 232b8a2.
* generator: --backend ui — universal authoring for UI plugin backends
New emitters (lidl_gen_ui.{h,cpp}) for type=ui_qml + interface=universal
modules: from the author's single clean impl class (optionally deriving
LogosModuleContext), generate
<name>.rep — the view contract, one SLOT per public method
(framework hooks like onContextReady excluded)
<name>_ui_interface.h — PluginInterface subclass + IID
<name>_ui_glue.{h,cpp}— plugin deriving <Cls>SimpleSource +
<Cls>Interface + <Cls>ViewPluginBase; slots
forward to the impl (std<->Qt at the boundary);
Q_INVOKABLE initLogos(LogosAPI*) builds
LogosModules, wires modules(), stamps context
(fires onContextReady — same order as the
universal core glue), then setBackend(this).
Typed dependency callers, typed event subscriptions and bind_<interface>
binders come from the existing umbrella pass (logos_sdk.h), which already
runs for UI modules. ui-host's reflection-based initLogos call site is
unchanged; legacy LogosAPI*-based UI plugins are untouched. v1 view API
types: void/int/uint/float64/bool/string; logos_events: rejected for ui
backends (views talk to QML via .rep, not module events).
* generator: distribute the LIDL frontend for external generators
First step of moving ALL Qt glue emission out of this repo into
logos-qt-sdk's logos-qt-generator (cpp-sdk's generator keeps only the
Qt-free outputs: std typed wrappers, logos_sdk umbrella, cdylib
impl-exports, LIDL derivation).
- Shared emit helpers (lidlToPascalCase, lidlTypeToQt, lidlTypeToStd,
lidlIsStdConvertible) move to a new lidl_emit_common.{h,cpp} unit, used
by both generators.
- The frontend set (AST, lexer, parser, serializer, validator,
impl-header parser, emit-common) is installed under
share/lidl-frontend/ — the qt generator compiles these sources in
directly, so the two tools share one frontend without a binary ABI.
* generator: Qt glue emission removed — logos-qt-generator owns it
The deletion half of the generator split (counterpart: logos-qt-sdk
3b37474, builder d7a2272). This tool now emits ONLY Qt-free outputs:
kept std typed wrappers + logos_sdk umbrella (--general-only),
cdylib C-ABI impl-exports + typed event emitters
(--lidl/--from-header --backend cdylib --impl-class),
LIDL derivation/serialization (--header-to-lidl), client stubs
removed universal Qt glue (--backend qt), the uniform cdylib Qt glue,
the ui backend emitters — all relocated verbatim (byte-identical
output verified) to logos-qt-generator; invocations here now
fail with a pointer to the right tool
The provider-glue golden tests travel with the emitters (to be re-homed
in logos-qt-sdk's test suite); test_lidl_type_mapping stays — it covers
lidl_emit_common, which both generators compile.
* lock: protocol at the typed-requestModule port (3de5398)
* lock: protocol at the typed-requestModule port (3de5398)
* ci: chain pins for the doc-tests (drop at merge)
In repo CI only cpp-sdk's {release} is the commit under test —
logoscore-cli and module-builder expanded to master, which doesn't link
against the chain SDK the specs override in ('Build the CLI with the SDK
override' failed on every run since the stacked-PR triggers were
enabled). Pin both to the extraction-chain heads; the workspace pipeline
is unaffected (it pins every repo itself).
* generator: distribute the LIDL frontend for external generators
First step of moving ALL Qt glue emission out of this repo into
logos-qt-sdk's logos-qt-generator (cpp-sdk's generator keeps only the
Qt-free outputs: std typed wrappers, logos_sdk umbrella, cdylib
impl-exports, LIDL derivation).
- Shared emit helpers (lidlToPascalCase, lidlTypeToQt, lidlTypeToStd,
lidlIsStdConvertible) move to a new lidl_emit_common.{h,cpp} unit, used
by both generators.
- The frontend set (AST, lexer, parser, serializer, validator,
impl-header parser, emit-common) is installed under
share/lidl-frontend/ — the qt generator compiles these sources in
directly, so the two tools share one frontend without a binary ABI.
* docs: note the generator split (Qt glue emission lives in logos-qt-sdk)
* lock: protocol#3 merged — pin advances to protocol master
---------
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
||
|
Dario LipicarandClaude Opus 4.8
|
f0fe8cbfeb |
Make the base SDK Qt-free: Qt developer layer moves to logos-qt-sdk (#83)
* Extract the protocol layer into logos-protocol; consume it as a flake input The transport/token/IPC layer (transports incl. QRO + plain TCP/TLS, consumer core LogosAPIClient/LogosAPIConsumer with the capability auto-requestModule flow, ModuleProxy, token manager, QVariant<->JSON conversion, the abstract LogosProviderObject interface) now lives in the logos-protocol repo behind the versioned lp_* C ABI. This SDK keeps the typed C++ developer layer (LogosAPI, provider base classes + Qt provider glue, module context, code generator) and still compiles the protocol sources INTO liblogos_sdk.a from the flake input, so the installed artifact (archive symbols, include/ + include/cpp layouts, cmake config) stays byte-compatible: existing consumers need no changes. Public headers are unchanged; logos_provider_object.h keeps its name and now re-exports the abstract interface from logos_provider_interface.h. Transport/protocol component tests moved to logos-protocol with the code; the remaining sdk/generator/experimental suites are unchanged (432/432 green against the local protocol checkout). * lock: add logos-protocol input * Make the base SDK Qt-free: move the Qt developer layer to logos-qt-sdk LogosAPI, LogosAPIProvider, LogosProviderBase/LOGOS_PROVIDER macros, the QObject provider glue (QtProviderObject) and the legacy PluginInterface (core/interface.h) move to the new logos-qt-sdk repo. The protocol sources are no longer compiled into a monolithic archive — consumers link logos-qt-sdk (which layers on logos-protocol) instead. What remains here is header-only std C++: logos_module_context.h, logos_result.h (StdLogosResult), logos_json.h — exported as the CMake INTERFACE target logos-cpp-sdk::logos_headers — plus the code generator (a build-time tool; its introspection mode now includes logos_provider_interface.h from logos-protocol, where LogosProviderPlugin moved). Mechanically verified Qt-free: the logos-cpp-lib / logos-cpp-include closures contain only nlohmann_json. Tests: 245/245 (module-context std suite + generator + experimental). * fix: accept the installed source-export layout in the protocol-root check The fail-fast only tested <root>/cpp/logos_protocol.h, but the LP_SRC selection right below (and the error message itself) support the installed export layout <root>/include/cpp as well. Pointing LOGOS_PROTOCOL_ROOT at an installed export tripped the FATAL_ERROR before that fallback could apply. Caught by Copilot review on #82. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * lock: pin logos-protocol to the qt-free-split branch head The Qt-free SDK (and the cdylib backend stacked on it) reference LogosProviderPlugin from protocol's logos_provider_interface.h, which lands on feat/qt-free-split — the P1-branch pin no longer compiles standalone. Temporary — drop when the chain PRs merge. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * doctest: pin the logoscore runtime via its {release} placeholder The spec built logoscore-cli at bare master with only the cpp-sdk inputs overridden — master's stack cannot compile against the qt-free SDK, so the suite failed on the chain branches. With the placeholder, CI's --release-for pins expand it to the workspace's logoscore commit (and local runs without a pin still fall back to master, unchanged). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * doctest: override the nested module builders to {release} too capability_module (via logoscore's lock) and the cloned accounts module resolve module-builder from their own locks — pre-split revs whose LogosModule.cmake still detects the SDK by logos_api.h, which the qt-free SDK no longer ships ('logos-cpp-sdk not found'). Overriding the builder itself to the workspace-pinned chain rev (keeping the nested cpp-sdk override) builds both modules with the split-aware builder. Verified end-to-end locally with the exact doctest command. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * doctest: apply the {release} + nested-builder overrides to all three specs The runtime spec got the treatment in 210eea1; the composition and worker-thread specs have the same logoscore/module build commands and failed identically (pre-split builders from the modules' own locks). All executed run: blocks now pin logoscore-cli{release} and override the nested module builders to logos-module-builder{release}; the displayed code_block: variants stay in their generic master form. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * codegen: typed wrappers throw on call failure; dispatch catches escapes Generated sync client wrappers call the new err-out invokeRemoteMethod overload and throw logos::LogosCallError when the call fails (e.g. the bound module is missing) — previously the empty QVariant silently degraded to the return type's default and a caller could not tell failure from a legitimate 0 / "". Both generators (legacy + LIDL), both API styles. Async paths unchanged. Generated provider dispatch (universal qt glue + LOGOS_PROVIDER) wraps the method body in a catch-all that logs and returns an invalid QVariant — an escaped exception becomes an ordinary METHOD_FAILED instead of unwinding through Qt event dispatch and killing the module process. * codegen: CallError out-param instead of throwing wrappers Per review, the generated sync wrappers expose the error channel as an optional trailing parameter — add(a, b, &err) — rather than throwing: explicit, stateless, works on temporaries, and existing call sites compile unchanged (they keep default-on-failure, now with a qWarning so failures are visible in the module log). The dispatch catch-all from the previous commit stays: it contains author exceptions, it doesn't introduce any. * glue: fire onContextReady AFTER modules()/event wiring The generated onInit set the context (which fires the impl's onContextReady hook) before constructing the LogosModules aggregate and wiring typed event emission — so an impl doing its documented one-time setup there (typed dependency calls, event subscriptions) dereferenced a null aggregate and crashed the module process (signal 11). Found by the first module to subscribe to a dependency's typed event from onContextReady. Context now goes last. * ci: run workflows on stacked PRs + workflow_dispatch Both workflows filtered pull_request to master-based PRs, so stacked PRs (feat/qt-free-sdk -> feat/extract-logos-protocol, feat/cdylib-authoring -> feat/qt-free-sdk) ran NO checks at all. Drop the base-branch filter for pull_request and add workflow_dispatch for manual runs. Same fix as logos-module-builder 232b8a2. * lock: protocol at the typed-requestModule port (3de5398) * ci: chain pins for the doc-tests (drop at merge) In repo CI only cpp-sdk's {release} is the commit under test — logoscore-cli and module-builder expanded to master, which doesn't link against the chain SDK the specs override in ('Build the CLI with the SDK override' failed on every run since the stacked-PR triggers were enabled). Pin both to the extraction-chain heads; the workspace pipeline is unaffected (it pins every repo itself). * generator: distribute the LIDL frontend for external generators First step of moving ALL Qt glue emission out of this repo into logos-qt-sdk's logos-qt-generator (cpp-sdk's generator keeps only the Qt-free outputs: std typed wrappers, logos_sdk umbrella, cdylib impl-exports, LIDL derivation). - Shared emit helpers (lidlToPascalCase, lidlTypeToQt, lidlTypeToStd, lidlIsStdConvertible) move to a new lidl_emit_common.{h,cpp} unit, used by both generators. - The frontend set (AST, lexer, parser, serializer, validator, impl-header parser, emit-common) is installed under share/lidl-frontend/ — the qt generator compiles these sources in directly, so the two tools share one frontend without a binary ABI. * lock: protocol#3 merged — pin advances to protocol master --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com> |
||
|
Dario LipicarandClaude Opus 4.8
|
38bc77e127 |
Consume logos-protocol: the transport/token/IPC layer moves behind the lp_* C ABI (#82)
* Extract the protocol layer into logos-protocol; consume it as a flake input The transport/token/IPC layer (transports incl. QRO + plain TCP/TLS, consumer core LogosAPIClient/LogosAPIConsumer with the capability auto-requestModule flow, ModuleProxy, token manager, QVariant<->JSON conversion, the abstract LogosProviderObject interface) now lives in the logos-protocol repo behind the versioned lp_* C ABI. This SDK keeps the typed C++ developer layer (LogosAPI, provider base classes + Qt provider glue, module context, code generator) and still compiles the protocol sources INTO liblogos_sdk.a from the flake input, so the installed artifact (archive symbols, include/ + include/cpp layouts, cmake config) stays byte-compatible: existing consumers need no changes. Public headers are unchanged; logos_provider_object.h keeps its name and now re-exports the abstract interface from logos_provider_interface.h. Transport/protocol component tests moved to logos-protocol with the code; the remaining sdk/generator/experimental suites are unchanged (432/432 green against the local protocol checkout). * lock: add logos-protocol input * fix: accept the installed source-export layout in the protocol-root check The fail-fast only tested <root>/cpp/logos_protocol.h, but the LP_SRC selection right below (and the error message itself) support the installed export layout <root>/include/cpp as well. Pointing LOGOS_PROTOCOL_ROOT at an installed export tripped the FATAL_ERROR before that fallback could apply. Caught by Copilot review on #82. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * lock: protocol at the typed-requestModule P1 port (1e4bc72) * lock: protocol at master (protocol#2 merged) The extraction is on protocol master now (29afbac); the temporary branch pin is dropped. --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com> |
||
|
Dario LipicarandClaude Opus 4.8
|
bb6d87b6ec |
fix: parse declarations on same line as logos_events: / access specifier (#76) (#81)
* fix: parse declarations on same line as logos_events/access specifier (#76) The impl header parser updated its section state and immediately broke out of line processing when it matched `logos_events:` (or `public:`/`private:`), discarding any declaration on the same physical line. This meant clang-format / prettier output like logos_events : void versionReady(const std::string &version); silently dropped the event, while the newline-separated form parsed fine — the same valid C++ was handled differently based on formatting. Strip any leading section specifiers in a loop, updating the section state, then let the remainder of the line fall through to the declaration parser. Brace counting still happens once per physical line and blank-line doc-comment reset is preserved. Adds a regression test (SameLineSectionSpecifiers) with a fixture covering the exact prettier form from the issue, a follow-on same-line event, the newline form alongside it, and the symmetric inline `public:` method case. Fixes #76 Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * fix: attach doc comments to same-line logos_events/access-specifier decls Address review feedback: the first pass cleared pendingDoc on every specifier match, so a `///` comment above a collapsed `logos_events : void foo();` did not attach to the event. In the collapsed form there is nowhere else to put the doc comment, so this left documentation formatting-dependent — the same bug class as #76, one level up. Only clear pendingDoc for a *bare* specifier (a section boundary, matching Qt `signals:` semantics); when a declaration shares the line, keep the pending doc so the declaration parser attaches it. Extend the fixture with a `///`-documented same-line event and assert the description is captured. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com> |
||
|
Dario LipicarandClaude Opus 4.7
|
42a8b9ed5c |
feat: LIDL interface IR — --header-to-lidl frontend + --dep backend (#77)
* feat: LIDL as the interface IR — --header-to-lidl frontend + --dep backend Make the generator pivot around LIDL so every binding flows source -> LIDL -> C++ (and a future Rust module plugs into the same backend via Rust -> LIDL): - --header-to-lidl <impl.h> --impl-class X --metadata m.json -o out.lidl: the standalone C++ frontend. Runs parseImplHeader -> lidlSerialize and emits ONLY the <name>.lidl contract (no Qt glue/dispatch), so a module can publish a cheap `lidl` artifact without compiling its plugin. - --dep <name>=<lidl>: the LIDL backend for concrete dependencies. Reuses the interface-wrapper path with BindMode::Static, emitting the name-baked modules().<dep> wrapper from the dep's published LIDL. Deduped vs each other and vs --interface names. - generateInterfaceWrappers gains a BindMode param (default Bound); --interface stays Bound, --dep is Static. parseInterfaceFlags generalized to parseSpecFlags(args, flag) for both --interface and --dep. The umbrella already emits a `<dep>` member per metadata dependency, so no umbrella change is needed. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * address review: strip leading '@' in --header-to-lidl paths; fix doc comment - --header-to-lidl now strips a leading '@' from the header/metadata/output path args (matches legacy_main; some build drivers pass @/abs/path). - Remove the stale "bound wrapper" doc comment above generateInterfaceWrappers (it now generates Static dep wrappers too via BindMode). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
Dario LipicarandClaude Opus 4.7
|
eb71a1aa90 |
feat: dependency interfaces — SDK code generator (bound wrappers) (#74)
* feat: dependency interfaces — runtime-bound typed wrappers
Turn a declared "interface" (a .lidl file or a pure-C++ header with a
logos_events: block) into a BOUND client wrapper: the target module name
is a constructor argument instead of a baked-in literal, so one interface
can be bound to any satisfying module at runtime.
- generator_lib: new BindMode { Static, Bound }. In Bound mode the ctor
takes (LogosAPI*, const QString& moduleName) and every invokeRemoteMethod*
/ ensureReplica routes through m_moduleName. Default Static leaves existing
name-baked output byte-for-byte unchanged.
- legacy/main.cpp: repeatable --interface <name>=<path>[=<impl_class>] flag,
consumed in --general-only. Parses .lidl via lidlParse and .h via
parseImplHeader, emits the bound <name>_api.{h,cpp}, and adds
bind_<name>(moduleName) factories (QString + std::string) to the
LogosModules umbrella. Also self-resolves local interface_dependencies
from metadata.json for non-nix builds.
- experimental/lidl_gen_client: same BindMode parity for the --lidl path.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* address review: dedup and validate --interface specs
A repeated --interface <name>=... would emit duplicate #include and
bind_<name>(...) into logos_sdk.h and fail to compile; empty name/path were
silently accepted. Dedup the flag-derived specs by name and drop malformed
ones with an explanatory message on stderr. (Copilot review, PR #74.)
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
|
||
|
Dario LipicarandClaude Opus 4.8
|
3bdd8858f5 |
handle reserved words in names/parameters correctly (#72)
* handle reserved words in names/parameters correctly * parser: accept reserved words as dependency names too Addresses review feedback on the depends list: parseMetadata() still hard-required LidlToken::Ident for each entry, so a dependency named after a keyword (e.g. `version`) would fail to parse even though lidlSerialize() emits it unquoted. Use atName() there too, consistent with the contextual-keyword rule applied to the other name positions. Adds a KeywordAsDependencyName regression test. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com> |
||
|
Dario LipicarandClaude Opus 4.8
|
7b62ac2017 |
Per-event documentation + getPluginEvents introspection (#71)
* Add per-event documentation + getPluginEvents introspection Mirror the per-method documentation pipeline for events. Events (declared in a universal module's logos_events: section) now carry a description parsed from their /// doc comments, and are introspectable at runtime via a new getPluginEvents framework call. - lidl_ast: EventDecl gains a description field. - impl_header_parser: capture the event's doc comment (previously discarded) and an optional metadata.json events[].description. - lidl_gen_provider: generated universal provider emits getEvents() override, mirroring getMethods() (name/signature/ parameters/description; no returnType/isInvokable — events are void). - logos_provider_object: default-empty virtual getEvents() so the legacy provider path and QtProviderObject inherit empty. - module_proxy / qt_provider_object: intercept getPluginEvents next to the getPluginMethods special-case. - docs: spec + README event-documentation notes. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * Add unit tests for event documentation + getEvents generation Address review feedback (#71): cover the event-introspection paths that previously only had method-side tests. - impl_header_parser test: assert metadata.json events[].description is parsed; new documented_events fixture asserts `///` doc-comment capture on a logos_events: block (multi-line joined with \n, adjacent-only, plain // ignored). - lidl_gen_provider test: assert the generated dispatch contains getEvents() emitting each event's name/signature/parameters and an escaped description, and that events carry no returnType/isInvokable. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * Fold event introspection into getMethods() to keep the provider ABI stable The previous approach added a getEvents() virtual to LogosProviderObject, which inserted a new vtable slot and shifted every later slot — an ABI break that would misdispatch virtual calls whenever an old and new host/module were mixed across the in-process plugin boundary. Instead, report events INSIDE the existing getMethods() call: it now returns the module's whole interface, with each entry tagged type "method" or "event" (events omit returnType/isInvokable). The provider vtable is therefore byte-for-byte unchanged, so old/new hosts and modules stay binary-compatible — a new host reading an old module sees no event entries (zero events), and an old host reading a new module just ignores the "type" field (cosmetic). An entry with no "type" is treated as a method. - logos_provider_object.h: remove the getEvents() virtual; document that getMethods() carries both, and why. - generator (lidl_gen_provider): emit events as type "event" entries inside getMethods(); tag methods type "method"; no getEvents() output. - module_proxy / qt_provider_object: getPluginMethods()/getPluginEvents() are now type-filtered views of getMethods(), plus a new getPluginInterface() returning the whole list. (These are name- dispatched Q_INVOKABLEs, not vtable surface — adding them is safe.) - tests: generator asserts events fold into getMethods() tagged "event"; ModuleProxy asserts the three filtered views; parser tests unchanged. - docs: spec/project/docs/README updated, incl. an ABI rationale note. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com> |
||
|
Dario LipicarandClaude Opus 4.8
|
760916e97b |
Parse method doc comments into per-method description (#70)
* parse adjacent method comments to populate description * preserve line breaks in method descriptions Join doc-comment lines with newlines instead of spaces (markers stripped, leading/trailing blank lines dropped, interior blanks kept), and escape \n when emitting the description into the generated getMethods(). Both codegen paths updated; docs corrected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * don't count braces inside comment lines (impl-header parser) A brace in a doc/line comment (e.g. `/// returns { ... }`) no longer affects class-scope tracking, which previously could make the parser think the class ended early and drop later declarations. Addresses review feedback on #70. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com> |
||
|
Dario Lipicar
|
c71089abe9 | remove legacy emitEvent (#69) | ||
|
Dario Lipicar
|
8bdbd13848 |
Extend universal modules with module context (#61)
* extend universal modules with module context * implement module calls and events for universal modules * pr comments |
||
|
Iuri Matias
|
f7c855b110 | add logos result type (#55) |