This was necessary because we used to cross-compile the library but
compile the precomputation programs for the build host. Now it's no
longer necessary and we can cleanly link even the external ASM
(which was the intent of #935).
On the way, remove an obsolete "-I" parameter.
- Improve Makefile.am for both prebuilt tables files
- On the way, tidy EXTRA_DIST: Move the header files to noinst_HEADERS,
where they conceptually belong, and add missing SECURITY.md to EXTRA_DIST
099bad945e Comment and check a parameter for inf in secp256k1_ecmult_const. (Russell O'Connor)
6c0be857f8 Verify that secp256k1_ge_set_gej_zinv does not operate on infinity. a->x and a->y should not be used if the infinity flag is set. (Russell O'Connor)
Pull request description:
a->x and a->y should not be used if the infinity flag is set.
ACKs for top commit:
robot-dreams:
ACK 099bad945e
real-or-random:
ACK 099bad945e I inspected all call sites, they all ensure that a is not infinity
Tree-SHA512: 495fcfe4ec4cacb3fc64bd5d04ecc67ab34f6b63666c6169d473abfd63c2041bc501a9a60d817566517435b986406ea2b7db3f5806043cecf30e214eba9892e9
5eb519e1f6 ci: reduce TEST_ITERS in memcheck run (Pieter Wuille)
e2cf77328a Test ecmult functions for all i*2^j for j=0..255 and odd i=1..255. (Pieter Wuille)
Pull request description:
Instead of just testing properties of the points xG for x=-36..36:
* also compute all xG where x=j*2^i for i=0..255 and odd j=1..255.
* test them against known exact results (SHA256 all of them, and compared against an independently created result)
* test all 4 ecmult functions (and for secp256k1_ecmult and secp256k1_ecmult_multi_var, both as G, and through the generic point input)
ACKs for top commit:
real-or-random:
ACK 5eb519e1f6
jonasnick:
ACK 5eb519e1f6
Tree-SHA512: 5d3fcbff754e859ba27d4f4581fa91fafb450fa3f7880364667dba51287e7f02f489af19b9de6a6e0f52faa183c0c7ae46db6add05180c3d4f45a6557b00c0ed
c0cd7de6d4 build: add -no-undefined to libtool LDFLAGS (fanquake)
fe32a79d35 build: pass win32-dll to LT_INIT (fanquake)
Pull request description:
This takes care of two of the outstanding issues in #923. One being initializing libtool with `win32-dll` and the other being the addition of `-no-undefined` to the libtool LDFLAGS. See each commit for more details.
Builders cross-compiling for Windows (including Core) will no-longer see:
```bash
libtool: warning: undefined symbols not allowed in x86_64-w64-mingw32 shared libraries; building static only
```
I'm planning on making some related changes downstream.
ACKs for top commit:
sipa:
utACK c0cd7de6d4. We indeed have done the work to propertly mark exported symbols, and AFAIK have no imported symbols apart from standard library ones.
real-or-random:
ACK c0cd7de6d4
hebasto:
ACK c0cd7de6d4
Tree-SHA512: 6756bc88ac439a27117a1341d82a801cef70354a9e7a563592ab3ac7298fbefdaa0a2c410ea3fba8953d53f254c449dc491069f30468db12791027a65dd02f80
592661c22f ci: move test environment variable declaration to .cirrus.yml (siv2r)
dcbe84b841 bench: add --help option to bench. (siv2r)
Pull request description:
Fixes#1005
**Changes:**
- added `--help` option to `bench.c`
- `help()` function prints the help to command-line
- `have_invalid_args()` checks if the user has entered an invalid argument
- moved `secp256k1_bench_iters` and `secp256k1_test_iters` environment variables declaration to `.cirrus.yml`
ACKs for top commit:
sipa:
utACK 592661c22f
real-or-random:
ACK 592661c22f
Tree-SHA512: ebc6a2e6e47b529212efa1c9b75cc79649fca7f42aa75ce46502db24ac94f46b6cef59c828d13265d1fa69187a81c140d1951e7daeb7c8e008a6c1ad75259741
The following functions were created:
1. bench.c: help()
- prints the help to the command-line
2. bench.h: have_invalid_args()
- takes a list of arguments that the user is allowed to enter on the command-line
- returns 1 if the user entered an invalid argument
- returns 0 if all the user entered arguments are valid
2b7c7497ef build: replace backtick command substitution with $() (fanquake)
Pull request description:
This is only needed for the very oldest of non-POSIX-compatible shells.
Note that this code will also only be executed on macOS, where it'd be
very unlikely to run into such a shell anyways.
Followup to https://github.com/bitcoin-core/secp256k1/pull/1019#pullrequestreview-815300521. I had thought there were more usages of this
syntax, but seems like it's just the one.
See:
https://github.com/koalaman/shellcheck/wiki/SC2006
Co-authored-by: Hennadii Stepanov <32963518+hebasto@users.noreply.github.com>
ACKs for top commit:
real-or-random:
ACK 2b7c7497ef
hebasto:
ACK 2b7c7497ef, verified that this is the only case.
Tree-SHA512: 6192f5efe437ff428ce7843ac595049a1aa7969a9e696f649cfd4820b28fc96ad0fabd6eec0ec1ca404763f02e64af6a99e57666a00d8749c6212a0646211991
This is a backwards-compatible API change: Before this commit, a context
initialized for signing was required to call functions that rely on
ecmult_gen. After this commit, this is no longer necessary because the
static ecmult_gen table is always present. In practice this means that
the corresponding functions will just work instead of calling the
illegal callback when given a context which is not (officially)
initialized for signing.
This is in line with 6815761, which made the analogous change with
respect to ecmult and contexts initialized for signing. But as opposed
to 681571, which removed the ecmult context entirely, we cannot remove
the ecmult_gen context entirely because it is still used for random
blinding. Moreover, since the secp256k1_context_no_precomp context is
const and cannot meaningfully support random blinding, we refrain (for
now) from changing its API, i.e., the illegal callback will still be
called when trying to use ecmult_gen operations with the static
secp256k1_context_no_precomp context.
This is only needed for the very oldest of non-POSIX-compatible shells.
Note that this code will also only be executed on macOS, where it'd be
very unlikely to run into such a shell.
Followup to #1019.
See:
https://github.com/koalaman/shellcheck/wiki/SC2006
Co-authored-by: Hennadii Stepanov <32963518+hebasto@users.noreply.github.com>
60bf8890df ecmult: fix definition of STRAUSS_SCRATCH_OBJECTS (Jonas Nick)
Pull request description:
This bug was introduced in 7506e064d7 by adding
an allocation but not updating the constant.
ACKs for top commit:
robot-dreams:
ACK 60bf8890df
real-or-random:
ACK 60bf8890df
Tree-SHA512: d7782fe9bf09fea8cf22304ab13679223a48f4d8b09081e662ea162a68c4e35f6b5820fbe4c6030fabad02a48dfdd02eb9eef22262c1dbbf02955bb92b75aef8
214042a170 build: don't append valgrind CPPFLAGS if not installed (fanquake)
Pull request description:
Valgrinds CPPFLAGS, i.e `-I/usr/local/opt/valgrind/include`, are currently added to CPPFLAGS, regardless of whether valgrind is installed. This changes configure so that they are only added if valgrind is available. i.e the output of `brew list --versions valgrind` is non-null.
ACKs for top commit:
real-or-random:
ACK 214042a170
hebasto:
ACK 214042a170, tested on macOS Big Sur 11.6.1 (20G224, Intel).
Tree-SHA512: 5101636a0a12f1941b01967ca8eab7aa20f44db0d1ef4571a5ad6026bb89494b983465d34d93c8b17a260b695116792991da53d135bc19a3c9e974f5266a90af
812ff5c747 doc: remove use of 0xa0 "no break space" (fanquake)
Pull request description:
This is miscellaneous, but I don't think these were being used on purpose?
ACKs for top commit:
siv2r:
ACK 812ff5c. The non-breaking space character is replaced with whitespace. Tested with [NBSP highlighter extension](https://marketplace.visualstudio.com/items?itemName=viktorzetterstrom.non-breaking-space-highlighter) on vscode.
real-or-random:
ACK 812ff5c747
Tree-SHA512: ccfcc64798f5a5eb0c669eb00f4408ab713e6710d67fd15ee2a4dca0d052e27636d7f0ad312aa94be0cd068c7e7874441aa2e114c4118322d0c764398a4ff695
72de1359e9 ci: Enable -g if we set CFLAGS manually (Tim Ruffing)
Pull request description:
ACKs for top commit:
jonasnick:
ACK 72de1359e9
Tree-SHA512: 0456db0ba53410640653e5d11ee4f328be0657e1e9077aa982ed4fd3eb6e326cfc022ec7ab71fc5c62d7942a20bbc7a5e8000cf5b62201fa1c183853d899ea77
16d132215c refactor: Use (int)&(int) in boolean context to avoid compiler warning (MarcoFalke)
Pull request description:
This one should *really* be only a refactor with the goal to silence static analysis warnings. clang-14 (trunk) recently added one in commit f62d18ff14 and I expect other tools will offer similar warnings.
Follow up to #1006, which was not a refactor.
ACKs for top commit:
real-or-random:
ACK 16d132215c
jonasnick:
ACK 16d132215c
Tree-SHA512: c465522ea4ddb58b5974c95bc36423c453e6fcf5948cb32114172113b5244209ceaa9418ec86ebe210390ae5509c2f24a42c41a7353de4cfb8fd063b0d5c0e79
This fixes a compiler warning:
./src/ecdsa_impl.h:312:12: warning: use of bitwise '&' with boolean operands [-Wbitwise-instead-of-logical]
return !secp256k1_scalar_is_zero(sigr) & !secp256k1_scalar_is_zero(sigs);
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
3b157c48ed doc: Suggest keys.openpgp.org as keyserver in SECURITY.md (Tim Ruffing)
73a7472cd0 doc: Replace apoelstra's GPG key by jonasnick's GPG key (Tim Ruffing)
Pull request description:
I have verified the new key via other secure channels.
This closes#1003 .
We can skip the second commit but I expect https://github.com/bitcoin/bitcoin/pull/23466/ to be merged. If it won't be merged, we could still revert.
ACKs for top commit:
sipa:
ACK 3b157c48ed. I've also verified the key out of band.
jonasnick:
ACK 3b157c48ed
Tree-SHA512: 496f98121f14031bc693aa83bf208b253f79b700b4bca0b629deadc8852f76ef6d69ad90109baa771d7b9f6e4b983e4ed8dca404cf5aceffe9d520d3362b533a
Tim Ruffing
Pieter Wuille
siv2r
Russell O'Connor
Hennadii Stepanov
fanquake
Jonas Nick
Elliott Jin
Dimitris Apostolou
MarcoFalke