| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
Provides an easier way for an application to configure a list of
entropy sources they'd like to use, or add a custom entropy source to
their seeding.
Exposes some toggles for the global/default entropy sources to build.h
Adds basic entropy tests which runs the polls and does sanity checking
on the results, including compression tests if available. These are
less useful for the CSPRNG outputs but a good check for the ones
producing plain ASCII like the /proc reader.
|
| |
|
|
| |
[ci skip]
|
| |
|
|
| |
Should resolve the random CI breaks
|
| |\
| |
| | |
Fix sanitizer build in travis ci
|
| |/
|
| |
The parameter is actually --with-sanitizers.
|
| |
|
|
| |
Also fix a few cast and zero-as-nullptr warnings in the AltiVec header
|
| |\
| |
| | |
Enhance tests main readability
|
| |/ |
|
| |
|
|
| |
[ci skip]
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The only reason mul128.h was included in mp_types.h was for the
definition of dword. But dword is only needed by the generic version
of mp_madd, which is an internal header. So move both the inclusion
of the header and the dword definition to there.
Previously mul128.h was very public (mp_types.h to bigint.h to rsa.h,
for example) and use of __int128 causes problems in pedantic mode. So
additionally, prefer using the TI attribute version since GCC does not
complain about that. Clang's -Wpedantic does not seem to care about it
either way.
GH #330
|
| | |
|
| |\
| |
| | |
Add myself (tiwoc) to credits.rst and license.txt
|
| | | |
|
| |/
|
|
| |
[ci skip]
|
| | |
|
| |
|
|
|
|
| |
Re-encoding the server key exchange meant that any leading zeros
in the values for DHE (or SRP) would be stripped out. This would
cause the signature check to fail.
|
| |
|
|
| |
DB::spin now returns the number of rows affected
|
| | |
|
| |
|
|
| |
since the primality tests are expensive in CPU time.
|
| | |
|
| |
|
|
|
| |
the first block satisfy the request, since it was possible another
record header appeared in some appended corrupted data.
|
| |
|
|
|
|
|
|
|
|
|
| |
Fix a bug which rejected any short server key exchanges. These can
occur with a plain PSK with short or empty identity hints.
Disable SHA-224 by default.
Remove some vestigal RC4 cruft.
Push more on the TLS corruption tests.
|
| |
|
|
|
|
|
|
|
| |
The Boost directory iterator throws an exception in that case.
Interestingly the current VC version does not seem to. In the
interests of portability just avoid this operation rather than trying
to hack around it in the Boost fs version.
GH #328
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
| |
Increase the iterations of the BigInt::random_integer test. Since things
get slow quickly, leave the larger range tests to higher soak levels.
In TLS, if the corrupted data causes an exception immediately that's ok
because it's corrupted data.
|
| |
|
|
| |
[ci skip]
|
| |
|
|
|
|
|
|
|
|
|
| |
The tests previously had used 4 to 6 different schemes internally (the vec file
reader framework, Catch, the old InSiTo Boost.Test tests, the PK/BigInt tests
which escaped the rewrite in 1.11.7, plus a number of one-offs). Converge on a
design that works everywhere, and update all the things.
Fix also a few bugs found by the test changes: SHA-512-256 name incorrect,
OpenSSL RC4 name incorrect, signature of FFI function botan_pubkey_destroy
was wrong.
|
| | |
|
| |
|
|
| |
[ci skip]
|
| | |
|
| | |
|
| |\
| |
| | |
Fix broken Gentoo package hyperlink
|
| |/ |
|
| |
|
|
| |
[ci skip]
|
| |\
| |
| | |
Split up build options
|
| | |
| |
| |
| |
| |
| | |
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.
|
| | | |
|
| | | |
|
| | | |
|
| | |
| |
| |
| | |
[ci skip]
|
| |\ \
| | |
| | | |
Fix McEliece key gen endian dependency.
|
| |/ /
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The tests which generate McEliece keys using a deterministic RNG and
fixed seed failed on PowerPC (or other big endian systems) because the
vectors assumed we were creating elements little endian, which is
what happend with rng.randomize(&u16, 2) on x86
Fix it to always be little endian. No particular reason to prefer one vs the
other here (we're just trying for compatability with ourselves) and choosing
little endian avoids having to regen the vectors.
|
| | |
| |
| |
| | |
[ci skip]
|
| | |
| |
| |
| | |
[ci skip]
|
| | | |
|
| | | |
|
| |\ \
| |/
|/| |
TLS improvements
|
| | | |
|
