aboutsummaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Drop two test-generated files. Add one input file needed for ECDSA signaturelloyd2008-10-123-8/+1
| | | | test.
* Cleanup ECDSA tests. Enable test_ec_sign testlloyd2008-10-121-23/+12
|
* Move InSiTo's ECDSA tests into the main test suitelloyd2008-10-125-626/+580
|
* Move ECDSA_Signature into CVC module. It is not used by ECDSA directly now.lloyd2008-10-1111-187/+204
| | | | Change several ECC functions to return const references instead of const values.
* Remove unnecessary includeslloyd2008-10-111-10/+2
|
* Enable throwing of Illegal_Point in check_invariantslloyd2008-10-111-2/+2
|
* Minor additions to the Windows build instructionslloyd2008-10-111-5/+7
|
* Update the algorithms list in doc/info.txtlloyd2008-10-111-21/+41
|
* Generate an ECDSA cert if ECDSA is enabled during testslloyd2008-10-111-0/+9
|
* Default to using SHA-1 with ECDSA in X.509 signatureslloyd2008-10-111-0/+5
|
* Fix ECDSA/SHA-160 OID entrylloyd2008-10-111-1/+1
|
* MSVC does not recognize and as equiv to && in a preprocessor statementlloyd2008-10-111-1/+1
|
* Error was in both x86 and x86-64 asmlloyd2008-10-111-1/+1
|
* Drop files generated each test runlloyd2008-10-114-41/+0
|
* Clarify that HP-UX, QNX, Tru64, and like are tested not a few times a year ↵lloyd2008-10-111-3/+3
| | | | but maybe every few years (at least by me)
* Remove ECDSA/ECDH as todolloyd2008-10-111-8/+0
|
* Disable the implementation of word_add in mp_ia32 for the same reasonlloyd2008-10-111-0/+8
| | | | | as the version in mp_amd64. Presumably they both need the same constraint added for them to work correclty.
* Remove -Werror from example compile flagslloyd2008-10-111-1/+1
|
* Add set of ECC/ECDSA test data, including X.509 and CVC certificates,lloyd2008-10-1147-0/+218
| | | | from InSiTo
* Correct the expected signature name in test_decode_ecdsa_X509 - itlloyd2008-10-111-396/+393
| | | | | | | | | | | wanted ECC_TESTDATA/EMSA1_BSI(SHA-224), while Botan is providing ECDSA/EMSA1_BSI(SHA-224) which seems more useful anyway. Also reindent the code to match more or less the convention in the rest of the codebase. At this point all ECDSA tests are pasing. (That probably just means there aren't enough ECDSA tests)
* Move all ECDSA test data into ecc_testdatalloyd2008-10-111-20/+20
|
* Add more ECC domain parameter sets from InSiTolloyd2008-10-111-10/+295
|
* Don't abort if a curve is not found in test_curve_registrylloyd2008-10-111-10/+18
|
* Catch and print exceptions in the CHECK macroslloyd2008-10-111-4/+4
|
* Generate and check ECDSA signature in examplelloyd2008-10-111-2/+29
|
* Update release noteslloyd2008-10-111-0/+2
|
* Reenable /proc/cpuinfo CPU detectionlloyd2008-10-111-1/+1
|
* Print generated public key in ECDSA examplelloyd2008-10-111-0/+2
|
* Missing tr1 inclusion in freestore.hlloyd2008-10-111-0/+8
|
* Default to using Boost's TR1. There seems to be a bug in the GCC 4.3.2lloyd2008-10-111-2/+2
| | | | | implementation of shared_ptr on x86 that causes memory corruption; default to Boost to avoid this.
* Comment out or remove some of the louder/less useful trace messageslloyd2008-10-111-6/+3
|
* Add comment attributing source for ECDSA OIDlloyd2008-10-111-2/+2
|
* Add OIDs for ECDSA with SHA-2 from draft-ietf-pkix-sha2-dsa-ecdsa-04.txtlloyd2008-10-111-6/+17
|
* Add OIDs for ECDSA and ECDSA with EMSA1_BSI padding from InSiTolloyd2008-10-111-0/+7
|
* Correct OID for ECDSA paramlloyd2008-10-111-1/+1
|
* Remove last use of freestore.h from gfpmath, move it to CVC code which stilllloyd2008-10-116-15/+14
| | | | uses it. Include tr1 header directly in gfp_element.h
* Remove some unused code from freestore.h, including the code that broughlloyd2008-10-111-166/+4
| | | | | in the Boost dependency. So Botan is now back to just requiring TR1 shared_ptr (I think).
* Only run uname if Perl's -f operator returns true forlloyd2008-10-111-34/+52
| | | | | | | | | | | | | | | | '/bin/uname', '/usr/bin/uname', or '/bin/sh' - the existence of these files gives us some hope that exec'ing uname will work, and avoids an error on Windows builds. Only run uname once (uname -a), collecting it into $$config{'uname'} (previously configure.pl might run it as many as three times, uname -s, uname -p, and uname -m). Only submodels were being searched longest to shortest. Modify guess_cpu_from_this to also search arch names and aliases in this manner. This allows Config{'archname'} of 'x86_64-linux' to be correctly detected as x86-64 (was being detected as x86, since before the search was ordered by the Perl hash order).
* s/BOOST_CHECK/CHECK/ in macros since this isn't really using the Boost ↵lloyd2008-10-111-113/+113
| | | | testsuite at all
* Quieter GF(p) testslloyd2008-10-111-2/+6
|
* Document other changes since 1.7.16. Spell check.lloyd2008-10-111-5/+8
|
* Change GFpElement's scalar multiplication operators to take a u64bit insteadlloyd2008-10-112-28/+25
| | | | of u32bit. Some minor format cleanup.
* Disable the x86-64 implementation of word_add. I think there is a buglloyd2008-10-111-0/+8
| | | | | | | | | | | in the constraints. It turns out that the GF(p) tests all pass in 64-bit mode if this function is disabled. I suspect the problem is that innert_montg_mult_sos calls this function in ways that are unusual in terms of how it is used in the rest of the library (in particular calling it with constant zero arguments). I think a constraint error is causing GCC to generate bad code in certain instances with this function. Will need to investigate this further.
* Add a version for x86-64 in mp_asm64, though it is disabled by defaultlloyd2008-10-112-0/+8
| | | | | | because the dependency handling/module picker gets confused otherwise. This allows building a mostly C++ version with 64-bit words with only a very small amount of inline asm (which is good for debugging the asm).
* Modify botan-config so --libs does not use -L if the prefix is /usr orlloyd2008-10-111-1/+6
| | | | /usr/local (same as -I)
* If any are unequal print all values in test_gfp_shared_valslloyd2008-10-111-8/+10
|
* Get GF(p) test failures reported up to the higher level.lloyd2008-10-111-65/+159
|
* Rename SHA-{384,512} base class SHA_64_BASE to SHA_384_512_BASE. Rename ↵lloyd2008-10-104-22/+23
| | | | SHA_224256_BASE SHA_224_256_BASE
* Some small changes to GFpElement, including accepting u64bit insteadlloyd2008-10-092-30/+28
| | | | | | | of u32bit to the operator* and operator*= Instead of passing the n_dash array to inner_montg_mult_sos, which only uses the first word, just pass the word directly.
* Reorder GF(p) tests, add more diagnostic output on test failurelloyd2008-10-091-14/+14
|