| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
| |
Avoids some cut and paste, also removes the need for special logic in
configure.py for handling mp module specially.
Merge SIMD classes into a single type SIMD_4x32
|
| |
|
|\ |
|
| | |
|
|\ \ |
|
| | |
| | |
| | |
| | | |
The assembler shipping with current OpenBSD (5.9) does not support avx2
|
| | |
| | |
| | |
| | | |
The binary name for a sufficiently new GCC is 'eg++' on OpenBSD
|
|\ \ \
| | |/
| |/| |
|
| | | |
|
| |/
|/|
| |
| | |
FreeBSD 10 comes with clang installed by default, and no gcc
|
|/
|
|
|
|
| |
command. This broke Sun Studio among others. Reported by Jeffrey Walton.
[ci skip]
|
| |
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
A module policy is a file specifying three types of modules: ones which
are required, ones which are prohibited, and ones which should be used
if otherwise available (this is mostly for platform specific modules).
Finally there are whatever modules which exist in the library of which
the policy makes no mention. These will be included if an explicit
dependency of some other module pulls them in (so there is no reason
to mention base, utils, ... in the file) but skipped otherwise.
For example policy 'sane' does not mention 'utils' or 'twofish' either
way. Since utils is a dependency of other modules which are included,
but Twofish does not. However unlike an explicitly prohibited module,
not mentioned can still be requested as part of the build (here with
--enable-module=twofish)
Also fixes some test bugs noticed by compiling in different build
configs. DLIES test didn't check that the KDF and MAC existed. Adds a
typedef for MessageAuthenticationCode because typing it twice in a
single line in the DLIES test made me think it's way too long. :) Also
fix some fuzzer build problems. Due to a copy and paste bug the PKCS
certificate (it was not).
Inspired by GH #439
|
|/
|
|
|
|
| |
--with-debug-info and --no-optimizations remain to toggle each independently.
Closes GH #433
|
| |
|
| |
|
| |
|
|
|
|
| |
valgrind and throws off benchmarks.
|
|
|
|
| |
Based on GH #362 by emilymaier
|
|\
| |
| | |
make shutil.rmtree and os.makedirs more robust to AV interference
|
| |
| |
| |
| |
| |
| |
| | |
* fix Python 3 compatibility
* add comment explaining purpose of the workarounds
* raise exception if directory exists in robust_makedirs
* pylint fixes
|
| | |
|
| | |
|
| | |
|
|/ |
|
| |
|
| |
|
|
|
|
|
|
| |
sanitizers and coverage as all 4 are usefully set independently.
Use GCC 4.9 on CircleCI as it seems to be supported. Add ASan + UBSan test.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Previously a build had optimizations disabled completely when debug info was
emitted. But there are many reasons to use optimized builds with debug symbols
(running under valgrind, against afl, with Asan/Ubsan, in prod, etc).
And personally I find even debugging at -O2 or -O3 is fine most of the time and
worth it for the speed. Use the (now documented!) --no-optimizations flag if no
optimization is desired while debugging.
This also removes the distinction between library and application compile flags;
there is a single optimization level that is probably good enough for everything.
On Win32 it removes definding _CONSOLE for the application. This seems to be
some mythical value that may have been required at some point, but is not
documented anywhere I can find. Who knows what VC thinks, hoping I don't have to
add this back. It also drops defining `EBUG` which is what happens when you
tell cl.exe to '/DEBUG'. LOL.
|
|
|
|
|
| |
but the meaning of the option is probably easier to understand with
this name.
|
|
|
|
|
|
|
| |
On Darwin platforms, there are frameworks that can bundle libraries and
header files in a standardized directory structure. We need to support
linking to them because most of the OS X or iOS-specific APIs are
provided as frameworks.
|
| |
|
|
|
|
|
|
|
|
| |
With this change the tests pass when linked against a static library
built in the normal (non-amalgamation) fashion.
Remove the restriction in configure.py, and have circleci build the
clang static build as a non-amalg.
|
|
|
|
|
|
|
|
| |
* Rename variable to avoid confusion of build_shared and
options.build_shared_lib
* Don't automatically change shared -> static. Force user to do that
Fixes #211
|
| |
|
|
|
|
|
|
| |
Allow multiple ABI link flags with the same name, e.g.,
-arch armv7 -arch armv7s when cross-compiling
for iOS. This is the first fix necessary for fixing #188.
|
|
|
|
| |
Fixes #241
|
| |
|
|
|
|
|
|
|
|
| |
* Increase minimal python version to 2.7
* Remove monotone revision command
* Set choices for --link-method argument
Closes #217
|
| |
|
|
|
|
| |
Thanks to @vlajos https://github.com/vlajos/misspell_fixer
|
|
|
|
|
|
|
| |
Before a typo in a module name was a warning. It now is an error because
it is potentially harmful.
[ci skip]
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
| |
This commit aims to preserve all settings. In some cases it will drop
optimization flags and fall back to compiler's defaults in the new debug
mode.
|