aboutsummaryrefslogtreecommitdiffstats
path: root/src
Commit message (Collapse)AuthorAgeFilesLines
* Default to not setting a timestamp in the gzip headerJack Lloyd2017-12-142-5/+6
| | | | But allow it to be set via a constructor arg.
* Remove some unused OS feature flagsJack Lloyd2017-12-1317-20/+1
| | | | | We are unlikely to ever use gettimeofday or memset_s going forward. Also MinGW had the wrong flag for VirtualLock
* Drop IRIX supportJack Lloyd2017-12-135-13/+0
| | | | | | It turns out that GCC dropped support for IRIX in 4.8, so likely there was never even a C++11 compiler that could have been used to compile 2.x in any case.
* Simplify how make flags are setJack Lloyd2017-12-121-6/+5
|
* Try running rst2man in CIJack Lloyd2017-12-122-2/+2
|
* Fix pragma message to avoid warningschregger2017-12-121-1/+1
| | | Missing parentheses around pragma message caused warnings in Visual Studio.
* Add ability to generate man page with rst2man fixing #1349Jack Lloyd2017-12-112-0/+14
|
* Fix make cleanJack Lloyd2017-12-111-1/+8
|
* Merge GH #1351 Cleanup macro generation in build.hJack Lloyd2017-12-1110-22/+61
|\
| * Rename SSE4.x names to avoid underscoresJack Lloyd2017-12-118-14/+14
| | | | | | | | | | | | This breaks how we determine the ISA flags for amalgamation files. The code for doing that is kind of a hack but I don't want to mess with it right now, easier to just rename the ISA internally.
| * Finish with the CPU feature macrosJack Lloyd2017-12-111-2/+2
| |
| * Rename the SSE4 ISA extensionsJack Lloyd2017-12-118-14/+14
| | | | | | | | Simplifies macro generation
| * Further cleanup of macro generation in build.hJack Lloyd2017-12-111-5/+28
| |
| * More simplificationsJack Lloyd2017-12-111-3/+3
| |
| * Clean up macro generationJack Lloyd2017-12-112-8/+24
| |
* | Avoid a shadow warning under old Clang [ci skip]Jack Lloyd2017-12-111-2/+2
|/ | | | Bogus warning but whatever.
* Fix missing comma in test runner [ci skip]Jack Lloyd2017-12-111-1/+1
| | | | | This ended up trying to run the "aeadkdf" suite first, and then the actual aead and kdf suites didn't end up running until later.
* Add compiler language flags in build.h header [ci skip]Jack Lloyd2017-12-111-1/+1
| | | | Fixes #1353
* Fix make clean targetJack Lloyd2017-12-111-7/+3
| | | | | | | | | | It removed the doc output dir, so this sequence: $ ./configure.py $ make clean $ make docs would fail
* Improve debug output of the build_docs scriptJack Lloyd2017-12-111-2/+5
| | | | | Only log if there was output, and convert the binary string to Unicode string.
* Avoid a warning under AIX makeJack Lloyd2017-12-111-4/+2
| | | | It doesn't like variables that don't expand to anything.
* Merge GH #1346 Add for loops to templates, add cmake+bakefile templatesJack Lloyd2017-12-113-10/+148
|\
| * Tweak the bakefile output a bitJack Lloyd2017-12-101-8/+19
| | | | | | | | Multiple files can be listed in the sources {} block, so do that.
| * Fix bakefile arch setting for x86-32Jack Lloyd2017-12-101-1/+1
| | | | | | | | Restrict bakefile to x86 since thats all that is supported.
| * Escape the shared build flags for CMakeJack Lloyd2017-12-101-1/+1
| | | | | | | | It needs this for Windows with its declspec syntax
| * Use for loops in cmake and bakefile buildsJack Lloyd2017-12-102-13/+36
| |
| * Use template file to generate bakefileJack Lloyd2017-12-101-0/+28
| |
| * Generate the CMake using a template fileJack Lloyd2017-12-101-0/+62
| | | | | | | | Less code in configure.py == betters
| * Add basic iteration constructs to template filesJack Lloyd2017-12-101-10/+24
| | | | | | | | Eliminates the last bits of makefile-specific syntax from configure.py
* | Rename CRC24 tablesJack Lloyd2017-12-101-9/+9
| | | | | | | | These conflict with name of temp variables and MSVC gets noisy.
* | Enable using NEON on ClangJack Lloyd2017-12-104-20/+24
|/ | | | | Clang doesn't like the way SIMD shifts were implemented, I guess it fails to inline the constant. Make it a template parameter instead.
* Move -Wstrict-overflow to maintainer mode [ci skip]Jack Lloyd2017-12-091-2/+2
| | | | It tends to have false positives, and is deprecated starting in GCC 8
* Avoid saving a resumed session multiple timesJack Lloyd2017-12-071-1/+3
|
* Handle #1303 on the server sideJack Lloyd2017-12-071-1/+13
|
* On resuming a client session, save the certificates that were used.Jack Lloyd2017-12-073-3/+17
| | | | GH #1303
* Fix formatting in TLS server code [ci skip]Jack Lloyd2017-12-071-193/+179
|
* Remove the doc.stamp file in make clean target [ci skip]Jack Lloyd2017-12-071-0/+2
| | | | | Otherwise make docs followed by make clean leaves docs unbuilt and make won't rebuild them. GH #1337
* Fix make clean target - didn't remove libbotan-2.so symlink files [ci skip]Jack Lloyd2017-12-061-1/+1
|
* Fix seeding of test RNGJack Lloyd2017-12-061-4/+3
| | | | | If no explicit seed was provided, it ended up ignoring the timestamp derived seed.
* Add a new version function returning just the version without extrasJack Lloyd2017-12-063-19/+41
|
* Add some basic tests of CPUID classJack Lloyd2017-12-061-0/+44
|
* Fix uninitialized warning closes #927 [ci skip]Jack Lloyd2017-12-061-1/+1
|
* Add an option to generate PDF version of the documentationJack Lloyd2017-12-061-11/+36
| | | | Fixes #1337
* Merge GH #1339 Add ability to query supported named groupsJack Lloyd2017-12-063-31/+56
|\
| * Add known_named_groups to EC_Group to list known curves. Fixes GX #1338.J08nY2017-12-053-31/+56
| |
* | Fix signed vs unsigned comparison warningJack Lloyd2017-12-051-1/+1
| |
* | Cryptobox doesn't require filters anymoreJack Lloyd2017-12-051-6/+3
| |
* | Merge GH #1267 Multithreaded signing in XMSSJack Lloyd2017-12-0529-170/+649
|\ \
| * | Fixes lambda capture error in MSVCMatthias Gierlings2017-11-231-1/+3
| | | | | | | | | | | | | | | MSVC seems to not allow defining aliases inside lambda capture expressions. Defining the aliases beforehand, outside the lambda fixes the issue.
| * | Adds runtime benchmark to guess phys. core count.Matthias Gierlings2017-11-233-3/+122
| | | | | | | | | | | | | | | | | | | | | Adds a small runtime benchmark to prevent performance degradation by overprovisioning SMT CPUs with too many threads. This is a temporary workaround until a hardware and OS independent detection of the physical core count through Botan::CPUID is in place.