[Pkg-bitcoin-commits] [libsecp256k1] annotated tag upstream/0.1_20150202 created (now fa3453c)

Jonas Smedegaard dr at jones.dk
Sun Apr 3 18:25:48 UTC 2016


This is an automated email from the git hooks/post-receive script.

js pushed a change to annotated tag upstream/0.1_20150202
in repository libsecp256k1.

        at  fa3453c   (tag)
   tagging  c73f1bec516e3391cbcf3947e56ac491a16e04c3 (commit)
 tagged by  Jonas Smedegaard
        on  Mon Feb 2 15:36:53 2015 +0100

- Log -----------------------------------------------------------------
Upstream version 0.1~20150202
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
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=
=WXJF
-----END PGP SIGNATURE-----

Amir Taaki (1):
      add pkg-config support

CodeShark (1):
      Merge pull request #3 from CodeShark/no_short_circuit_eval

Cory Fields (20):
      autotools: autotools'ify libsecp256k1
      add missing include
      autotools: changes suggested by Pieter Wuille
      autotools: No need to link against libssl
      packaging: remove the --with-pkgconfigdir option
      packaging: fixup pkg-config
      packaging: add pkg-config file to gitignore
      build: add autogen. How was this missing?
      build: add travis support
      build: grab full paths to host tools
      travis: minimize the dependencies available for each build config
      warnings: enable quiet builds
      build: osx: attempt to work with homebrew keg-only packages
      travis: add clang to the test matrix
      build: fix __builtin_expect detection for clang
      build: fix openssl detection for cross builds
      x86 builder
      build: disable benchmark by default
      build: use subdir-objects for automake
      build: nuke bashisms

Diederik Huys (7):
      x86_64 specific assembly version of field code
      Small fix for ExSetSquare
      Comments
      Port to more generally used YASM assembler
      Solved register save/restore bug
      1st assembly version of field 5x64 code
      1st assembly version of field 5x64 code - bugfix

Eric Lombrozo (5):
      Added constant time Normalize operation to FieldElem class.
      First cast the conditional to an int64 so it works on 32-bit platforms.
      Added constant time methods to FieldElem class.
      Added ULL suffix to integer constant.
      Removed short-circuit evaluation potential.

Gregory Maxwell (31):
      Try to not leave secret data on the stack or heap.
      Misc. Warning and cosmetic error cleanups.
      Remove some dead variables in the tests.
      Add non-null and unused-result warnings for the external API.
      Reorder static to comply with C99 and switch to the inline macro.
      Correct function prototypes and avoid unused parameter warnings.
      Remove a VERIFY_CHECK for >=0ness on an unsigned type.
      Avoid a shadowed variable.
      Fix varrious signed/unsigned comparisons.
      Avoid unsigned comparison in scalar arith.
      Signed/unsigned comparisons in tests.
      Avoid forward static decl of undefined functions, also fix a paren warning in the tests.
      field_gmp's negate doesn't need to use the magnitude argument.
      Enable warnings.
      Switch to C89 comments in prep for making the whole codebase C89 compatible.
      Fix a memory leak and add a number of small tests.
      Remove redundant secp256k1_fe_normalize from secp256k1_gej_add_ge_var.
      Add some additional tests.
      Add a few more additional tests.
      Avoid constructing an invalid signature with probability 1:2^256.
      Use secp256k1_fe_equal_var in secp256k1_fe_sqrt_var.
      Add magnitude limits to secp256k1_fe_verify to ensure that it's own tests function correctly.
      Add tests for handling of the nonce function in signing.
      C89 nits and dead code removal.
      Convert field code to strict C89 (+ long long, +__int128)
      Covert several more files to C89.
      Switch scalar code to C89.
      Centralize the definition of uint128_t and use it uniformly.
      Convert tests to C89. (also fixes a use of bare "inline" in field)
      Convert the rest of the codebase to C89.
      Some readme updates, e.g. removal of the GMP field.

HaltingState (1):
      updated documentation

John Driscoll (1):
      Documented autotools build process in readme

Jonas Smedegaard (1):
      Imported Upstream version 0.1~20150202

Lucas Betschart (1):
      Update README.md : Travis status

Luke Dashjr (2):
      Fix build on x32
      Explicitly access %0..%2 as 64-bit so we use the right registers for x32 ABI

Matt Corallo (3):
      Compile with -fPIC so that we can build shared libraries.
      Add JNI wrapper for org.bitcoin.NativeSecp256k1 (bitcoinj)
      Fix JNI for C instead of C++

Pavel Janík (2):
      Include time.h header for time().
      Make git ignore bench_recover when configured with benchmark enabled

Peter Dettman (19):
      Re-write secp256k1_fe_sqrt and secp256k1_fe_inv to use a sliding window over blocks of 1s
      Fix addition chain entry in comments
      Add MacPorts default include/lib paths for OSX
      secp256k1_fe_sqrt checks for success
      Use endomorphism in precomputations
      Support 64bit_asm field on OSX
      A few improvements to the sqrt tests
      Rewrite 5x52 normalize method to be faster
      Improve normalization performance for 32bit
      Use batch inversion in G precomputation
      Test demonstrating discrepancy in sqr output
      Add verification to 32bit field
      Fix build for 64bit field under OSX
      Address 'constant-time' TODOs in field impls
      Rearrange _gej_add_ge to save an _fe_negate
      Rewrite mul/sqr for 32bit/64bit
      Avoid division when decomposing scalars
      Add _fe_normalizes_to_zero method
      Add _normalizes_to_zero_var variant

Phillip Mienk (1):
      Correct .gitignore to correctly account for build-aux.

Pieter Wuille (352):
      First commit
      square root + point decompression
      all kinds of things
      split off GroupElemJac: doesn't compile
      compiles at least
      working jac+aff addition
      fix inversion, simplify, remove templates
      OpenSSL BN_mod_inverse performance is bad
      split up
      begin ecmult
      compiles...
      working ecmult
      GLV optimization
      begin ECDSA
      maybe ecdsa
      small changes
      num_gmp.h and begin tests
      fix tests
      some comments and more loops in test
      more tests
      bugfix FieldElem::SetHex
      minimal
      bugfixes and num-based Field::Inverse
      maybe testable
      add header
      validation bugfixes: validates main bitcoin chain
      num_openssl.h bugfixes
      simple makefile
      remove Context interface... it doesn't help much
      Split field def/impl
      Split headers/code
      add signing
      Make Field::Normalize explicit
      Merge remote-tracking branch 'codeshark/master'
      Merge pull request #2 from CodeShark/master
      added TODO
      make assembly version optional
      Merge pull request #4 from PSYCPU/asm
      Merge pull request #5 from PSYCPU/asm
      First step in converting to C: num
      Merge pull request #8 from sipa/cify
      Merge pull request #6 from PSYCPU/asm
      Move 5x52 specific code to field_5x52
      Second step in converting to C: field
      Merge pull request #9 from sipa/cify
      Switch num/field to extern C; small fixes
      Use variable-time fe_inv in group
      Split 5x52 inner implementations
      Source tree reorganization
      Bugfix: secp256k1_fe_inv_var correct output
      Remove num_builtin.h; it wasn't complete
      Begin group C interface + start/stop
      Init/free field constant
      Merge pull request #7 from TheBlueMatt/master
      Third step in converting to C: group
      Fourth step in converting to C: ecmult
      Fifth step in converting to C: ecdsa
      Final step in converting to C
      Small fix
      move constants in field/group
      Makefile tweaks
      10x26 field implementation
      Some group.h comments
      Add field_gmp; only 64-bit platforms for now
      Reorganize source tree: no .c for non-objects
      Support for 32-bit limbs in field_gmp
      Revamp makefile
      Add obj/ directory
      Use GMP's low-level routines for num (mpn_)
      Some comments
      public interface docs
      Builtin random
      Change num_mod semantics
      Test improvements
      More num unit tests
      Add coverage check
      Better compilation flags for coverage
      Make constants constant
      Remove traces of old mpz-based num
      Add signing to public interface
      More public implementations
      Make pubkeys, set_xo, ... ge instead of gej
      Compact signatures/pubkey recovery
      Privkey import/export
      OpenSSL/EC tests
      Inline verify is slow, this is what unit tests are for
      updates
      MIT License
      Bugfix: correct privkey serialization
      Use privkey serialization to construct OpenSSL keys
      Public inclue improvements
      Merge pull request #10 from TheBlueMatt/master
      Allow private keys with less-than-32-byte secrets
      Private/public key tweaking
      Field 5x64
      Inline asm for 5x64
      Merge pull request #11 from PSYCPU/master
      Merge pull request #12 from PSYCPU/master
      Invert buggy logic in secp256k1_ecdsa_seckey_verify
      Add support for multiplicative tweaking
      Tweak fixes & variable name improvement
      Break malleability by producing S <= order/2
      Merge pull request #15 from haltingstate/master
      Check for overflows when order is added at recovery
      Make endomorphism optimization optional
      Remove 5x64 implementation - no actual benefits
      Merge pull request #14
      Merge pull request #19
      Slice bytes of G multiples to avoid cache timings
      Merge pull request #1 from sipa/slice
      Move implementations from impl/*.h to *_impl.h
      Merge pull request #2 from sipa/impl
      Create README.md
      More details in README.md
      Merge pull request #5 from sipa/morereadme
      Merge pull request #7 from luke-jr/x32
      Merge pull request #9
      Merge pull request #8
      Merge pull request #23
      Merge pull request #22
      Merge pull request #20
      Merge pull request #19
      Merge pull request #17
      Merge pull request #18
      Merge pull request #24
      Merge pull request #25
      Do not free endomorphism constants when disabled
      Introduce CHECK() for tests that works with NDEBUG.
      Merge pull request #28
      Merge pull request #31
      Merge pull request #32
      No releases yet anyway
      Merge pull request #34
      Merge pull request #40
      Merge pull request #42 from lclc/patch-1
      Merge pull request #16
      Correct secp256k1_fe_verify and use it everywhere
      Merge pull request #44
      Merge pull request #46
      Merge pull request #47
      Add high-level secp256k1.c tests
      Merge pull request #49
      Add secp256k1_num_eq and use it in tests
      Merge pull request #50
      Make tests and bench just use asm directly instead of library
      Merge pull request #53
      Merge pull request #55
      Add VERIFY_CHECK/DEBUG_CHECK and use CHECK macros more
      Merge pull request #54
      Merge pull request #51
      Fix a signedness mistake in secp256k1_num_set_hex
      Merge pull request #56
      Nothing-up-my-sleeving blinding for a*G
      Only use the libcrypto part of OpenSSL
      Merge pull request #58
      Merge pull request #57
      Make secp256k1_ge_set_gej work with the point at infinity
      Merge pull request #62
      Split up signing and verification initialization
      Set precomputation table late and unset early.
      Merge pull request #60
      Make bench use external interface
      Merge pull request #67
      Better randomization for tests
      Merge pull request #68
      Split up ecmult and ecmult_gen entirely
      Merge pull request #71
      Verify num=openssl initialization and check repeatability
      Avoid uninitialized access in secp256k1_gej_double
      Tests take too long by default
      Avoid valgrind complaints in OpenSSL initialization check
      [API CHANGE] Use secp256k1_ec_ prefix for non-ECDSA key operations
      Move non-ECDSA operations from ecdsa to eckey
      Use internal secp256k1_eckey_ prefix for functions in eckey
      Abstract out tweak logic to secp256k1_eckey_* functions
      Merge pull request #72
      Introduce secp256k1_scalar_t for future constant-time mod order operations
      Merge pull request #75
      Switch scalar to use get/set 32-byte arrays
      Add unit tests for scalars.
      Merge pull request #76
      Add bench_inv tool
      Remove OpenSSL bignum implementation
      Merge pull request #78
      Merge pull request #79
      Make bench deterministic
      Rename bench to bench_verify
      Merge pull request #80
      Add bench_sign tool
      Better .gitignore for bench binaries
      Merge pull request #83
      Merge pull request #82
      Fix interaction between magnitudes and negation
      Merge pull request #70
      Get rid of {num,scalar,ecdsa_sig}_{init,free}
      Merge pull request #89
      Merge pull request #85
      Merge pull request #93
      Merge pull request #92
      Fix typo
      Label variable-time functions correctly and don't use those in sign
      Merge pull request #94
      Implementations for scalar without data-dependent branches.
      Merge pull request #77
      Merge pull request #95
      Branch-free point addition
      Merge pull request #98
      Document some preconditions
      Merge pull request #99
      Merge pull request #96
      Merge pull request #101
      Merge pull request #102
      Add equalities relating input and output variables
      Add overflow analysis to field_5x52_int128_impl.h
      Add overflow analysis to field_10x26_impl.h
      Merge pull request #103
      Merge pull request #106
      Merge pull request #105
      Update README.md
      Merge pull request #108
      Add DETERMINISTIC to avoid line number/source dependent binaries
      Merge pull request #107
      Make secp256k1_eckey_pubkey_serialize fail for infinity
      Test whether recovered public keys are not infinity
      Add test that recovering infinity fails
      Merge pull request #110
      Merge pull request #115
      Don't split the g factor when not using endomorphism
      Merge pull request #116
      Add bounds checking to field element setters
      Fix secp256k1_num_set_bin handling of 0
      Merge pull request #114
      Tweak precomputed table size for G
      Correct typo in comment
      Merge pull request #121
      Optimize doubling: secp256k1 has no y=0 point
      Explain why no y=0 check is necessary for doubling
      Do signature recovery/verification with 4 possible recid case
      Merge pull request #122
      Add secp256k1_scalar_add_bit
      Generalize secp256k1_scalar_get_bits
      Switch wnaf splitting from num-based to scalar-based
      Make test_point_times_order test meaningful again
      Add secp256k1_scalar_inverse_var which delegates to GMP
      Add scalar splitting functions
      Switch all EC/ECDSA logic from num to scalar
      Move lambda-splitting code to scalar.
      Remove unused num functions
      Merge pull request #117
      Merge pull request #120
      Make constant initializers independent from num
      Make num optional
      Require that r and b are different for field multiplication.
      Make secp256k1_fe_mul_inner use the r != property
      Add a test for r >= order signature handling
      Add a test case for ECDSA recomputing infinity
      Merge pull request #126
      Merge pull request #124
      Merge pull request #119
      Add secp256k1_scalar_mul_shift_var
      Convert lambda splitter to pure scalar code.
      Bugfix: b is restricted, not r
      Merge pull request #129
      Fix typo
      Merge pull request #131
      Use constant-time conditional moves instead of byte slicing
      Avoid undefined shift behaviour
      Make scalar_add_bit test's overflow detection exact
      Merge pull request #133
      Merge pull request #132
      Merge pull request #127
      Merge pull request #134
      Rewrite field assembly to match the C version
      Convert YASM code into inline assembly
      Make {mul,sqr}_inner use the same argument order as {mul,sqr}
      Merge pull request #135
      Allocate precomputation arrays on the heap
      Remove unused secp256k1_fe_inv_all
      Merge pull request #139
      Another redundant secp256k1_fe_normalize
      Merge pull request #140
      Variable time normalize
      Check return value of malloc
      Merge pull request #138
      Merge pull request #128
      Merge pull request #137
      Merge pull request #118
      Merge pull request #143
      Rename bench_verify to bench_recovery
      Make the benchmarks print out stats
      Compile with -O3 by default
      Merge pull request #141
      Merge pull request #144
      Merge pull request #136
      Merge pull request #145
      Merge pull request #146
      Really compile with -O3 by default
      Fix ECDSA message hashes to 32 bytes
      Avoid the stack in assembly and use explicit registers
      Merge pull request #155
      Enable tests in x86 travis builds
      Merge pull request #142
      Merge pull request #149
      Merge pull request #151
      Merge pull request #150
      Merge pull request #158
      Merge pull request #156
      Merge pull request #157
      Configure options reorganization
      Merge pull request #165
      Merge pull request #160
      Optimize verification: avoid field inverse
      Add explanation about how inversion can be avoided
      Merge pull request #123
      Remove GMP field implementation
      Merge pull request #168
      Make signing fail if a too small buffer is passed.
      Merge pull request #169
      [API BREAK] Use a nonce-generation function instead of a nonce
      Implement SHA256 / HMAC-SHA256 / RFC6979.
      Use rfc6979 as default nonce generation function
      weak normalization
      Weak normalization for secp256k1_fe_equal
      Merge pull request #154
      Merge pull request #163
      Rework group tests
      Add a reference consistency test to ge_tests.
      Merge pull request #175
      Merge pull request #176
      Merge pull request #177
      Merge pull request #178
      Convert the scalar constant initialization to static consts
      Remove unused secp256k1_fe_inner_{start, stop} functions
      Convert the field/group/ecdsa constant initialization to static consts
      Merge pull request #191
      Merge pull request #194
      Merge pull request #193
      Merge pull request #195
      Get rid of variable-length hex string conversions
      Merge pull request #196
      Field storage type
      Add group element storage type
      Use group element storage type in EC multiplications
      Switch ecmult_gen to use storage types
      Remove the non-storage cmov
      Merge pull request #197
      Merge pull request #199
      Merge pull request #200
      Merge pull request #201
      Merge pull request #202
      Add group operation counts
      Merge pull request #203

William Swanson (1):
      Check signature nonces for validity

caktux (1):
      fix unsigned warning in num_gmp_impl.h

evoskuil (1):
      Fix x64 gmp init on platforms that define UL as 32 bits.

kiwigb (3):
      Add autoreconf warnings. Replace obsolete AC_TRY_COMPILE.
      Remove INCLUDES. Obsolete, appears unused anyway.
      Use same build template as bitcoin. Add bitcoin_secp.m4.

mb300sd (1):
      fix ifdef/ifndef

-----------------------------------------------------------------------

No new revisions were added by this update.

-- 
Alioth's /usr/local/bin/git-commit-notice on /srv/git.debian.org/git/pkg-bitcoin/libsecp256k1.git



More information about the Pkg-bitcoin-commits mailing list