* Replace g1 points with g1 lagrange points
* Swap out validate_kzg_g1 with blst funcs
* Update minimal preset too
* Fix java & nodejs bindings
* Put variables in smaller scope
* Update some comments
* Remove FFTSettings (#2)
* Fix issue when expanding roots of unity
* Fix formatting
* Revert back to using intermediate array
* Add missing c_kzg_free
* Fix some nits
* Replace free_kzg_settings with free_trusted_setup
* Add parens to NUM_ELEMENTS
Co-authored-by: George Kadianakis <desnacked@riseup.net>
* Move memcpy to the end
* Revert "Move memcpy to the end"
This reverts commit 5331c7feadc92e4b5dd5d4e7512e4be563d7f386.
* Add comment about free_trusted_setup
* Move check before alloc
* Add remark
* Delete unnecessary blank line
* Fix asn's nits
* Update comment for roots_of_unity in header
Co-authored-by: Gottfried Herold <GottfriedHerold@users.noreply.github.com>
* Fix formatting
* Update comment about max_scale
Co-authored-by: Gottfried Herold <GottfriedHerold@users.noreply.github.com>
---------
Co-authored-by: Suphanat Chunhapanya <haxx.pop@gmail.com>
Co-authored-by: George Kadianakis <desnacked@riseup.net>
Co-authored-by: Gottfried Herold <GottfriedHerold@users.noreply.github.com>
Since `DirSep` matches the default separator of the target OS, it's
not suitable for performing path operations during compile-time on
the host OS.
The particular fix here solves the issue for cross-compiling a Windows
build from a Linux host, but Nim seems to need a more general solution
for detecting the host OS during compilation, so a host OS specific
separator can be used in paths derived from `currentSourcePath`.
* add eslint and run for TS files
* add eslint and run for JS/TS
* remove eslint file comments
* add endOfLine: "auto" for windows
* fix yarn command to remove excess output
* update eslint versions
* Run C tests on all platforms
* Don't install llvm on mac & add windows define
* Update blst settings for windows
* Install msvc on windows
* Don't specify blst as a library
* Try including -lmingw32
* Clean up makefile & test setting target
* Try using gcc for windows
* Don't treat warnings as errors on windows
* Test upping stack size
* Ignore some tasks on windows
* Clean up some things
* Fix indentations
* Delete extra blank line
* More organization
* Remove windows check around sanitize
* Move XCRUN initialization back to top
* Move extra sanitize cflags out like the others
* Intentionally fail to compile with a FIELD_ELEMENTS_PER_BLOB value that is not a power of 2.
We do not want to support or maintain this case, so let us be explicit about it.
* Update src/c_kzg_4844.h
Co-authored-by: Justin Traglia <95511699+jtraglia@users.noreply.github.com>
* Update src/c_kzg_4844.h
Co-authored-by: Justin Traglia <95511699+jtraglia@users.noreply.github.com>
---------
Co-authored-by: George Kadianakis <desnacked@riseup.net>
Co-authored-by: Justin Traglia <95511699+jtraglia@users.noreply.github.com>
* modify directory structure to isolate the generated bindings
* add a lib.rs file
* move deref impls to the extension file
* remove unused types from generated bindings
* cleanup new lines to reduce diff noise
* reorder definitions to reduce diff noise
* move sync and send impls to the extension file
* generate bindings
* blacklist used bindings; create snapshots dir to ensure freshness in the future
* fix typo. Ty @pawanjay176
* run cargo build after merge
* custom impl for KZGCommitment and KZGProof
* final touches