Noteworthy changes in version 1.4.5 (2009-12-11)
------------------------------------------------
* Fixed minor memory leak in DSA key generation.
* No more switching to FIPS mode if /proc/version is not readable.
* Fixed a sigill during Padlock detection on old CPUs.
* Fixed a hang on some W2000 machines.
* Boosted SHA-512 performance by 30% on ia32 boxes and gcc 4.3;
SHA-256 went up by 25%.
This changes the buildlink3.mk files to use an include guard for the
recursive include. The use of BUILDLINK_DEPTH, BUILDLINK_DEPENDS,
BUILDLINK_PACKAGES and BUILDLINK_ORDER is handled by a single new
variable BUILDLINK_TREE. Each buildlink3.mk file adds a pair of
enter/exit marker, which can be used to reconstruct the tree and
to determine first level includes. Avoiding := for large variables
(BUILDLINK_ORDER) speeds up parse time as += has linear complexity.
The include guard reduces system time by avoiding reading files over and
over again. For complex packages this reduces both %user and %sys time to
half of the former time.
* Publish GCRY_MODULE_ID_USER and GCRY_MODULE_ID_USER_LAST constants.
This functionality has been in Libgcrypt since 1.3.0.
* MD5 may now be used in non-enforced fips mode.
* Fixed HMAC for SHA-384 and SHA-512 with keys longer than 64 bytes.
* In fips mode, RSA keys are now generated using the X9.31 algorithm
and DSA keys using the FIPS 186-2 algorithm.
* The transient-key flag is now also supported for DSA key
generation. DSA domain parameters may be given as well.
is incorrect because:
1) _gcry_rngfips_deinit_external_test() is void function
2) the calling function, random, is declared void
The unpatched code will not compile with Sun compiler.
* Try to auto-initialize Libgcrypt to minimize the effect of
applications not doing that correctly. This is not a perfect
solution but given that many applicationion would totally fail
without such a hack, we try to help at least with the most common
cases. Folks, please read the manual to learn how to properly
initialize Libgcrypt!
* Auto-initialize the secure memory to 32k instead of aborting the
process.
* Log fatal errors via syslog.
* Changed the name and the semantics of the fips mode config file.
* Add convenience macro gcry_fips_mode_active.
* More self-tests.
* Documentation cleanups.
Noteworthy changes in version 1.4.1 (2008-04-25)
------------------------------------------------
* Fixed a bug introduced by 1.3.1 which led to the comsumption of far
too much entropy for the intial seeding.
* Improved AES performance for CFB and CBC modes.
* Removed build problems for the Padlock support.
VIA PadLock Security Engine:
- The new option is turned on by default.
- The new option is only available on i386 systems except Mac OS X (which
doesn't work on VIA CPU).
- The new option isn't available on system which uses GCC 3.x because it
causes build failures. This fixes PR pkg/38197.
Approved by Dieter Baron.
Noteworthy changes in version 1.4.0 (2007-12-10)
------------------------------------------------
* New configure option --disable-padlock-support which is mostly
useful in case of build problems.
Noteworthy changes in version 1.3.2 (2007-12-03)
------------------------------------------------
* The visibility attribute is now used if supported by the toolchain.
* The ACE engine of VIA processors is now used for AES-128.
* The ASN.1 DER template for SHA-224 has been fixed.
Noteworthy changes in version 1.3.1 (2007-10-26)
------------------------------------------------
* The entire library is now under the LGPL. The helper programs and
the manual are under the GPL. Kudos to Peter Gutmann for giving
permissions to relicense the rndw32 and rndunix modules.
* The Camellia cipher is now under the LGPL and included by default.
* Fixed a bug in the detection of symbol prefixes which inhibited the
build of optimzied assembler code on certain systems.
* Updated the entropy gatherer for W32.
Noteworthy changes in version 1.3.0 (2007-05-04)
------------------------------------------------
* Changed the way the RNG gets initialized. This allows to keep it
uninitialized as long as no random numbers are used. To override
this, the new macro gcry_fast_random_poll may be used. It is in
general a good idea to spread this macro into the application code
to make sure that these polls happen often enough.
* Made the RNG immune against fork without exec.
* Reading and writing the random seed file is now protected by a
fcntl style file lock on systems that provide this function.
* Support for SHA-224 and HMAC using SHA-384 and SHA-512.
* Support for the SEED cipher.
* Support for the Camellia cipher. Note that Camellia is disabled by
default, and that enabling it changes the license of libgcrypt from
LGPL to GPL.
* Support for OFB encryption mode.
* gcry_mpi_rshift does not anymore truncate the shift count.
* Reserved algorithm ranges for use by applications.
* Support for DSA2.
* The new function gcry_md_debug should be used instead of the
gcry_md_start_debug and gcry_md_stop_debug macros.
* New configure option --enable-random-daemon to support a system
wide random daemon. The daemon code is experimental and not yet
very well working. It will eventually allow to keep a global
random pool for the sake of short living processes.
* Non executable stack support is now used by default on systems
supporting it.
* Support for Microsoft Windows.
* Assembler support for the AMD64 architecture.
* New configure option --enable-mpi-path for optimized builds.
* Experimental support for ECDSA; should only be used for testing.
* New control code GCRYCTL_PRINT_CONFIG to print the build
configuration.
* Minor changes to some function declarations. Buffer arguments are
now typed as void pointer. This should not affect any compilation.
Fixed two bugs in return values and clarified documentation.
* Interface changes relative to the 1.2.0 release:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
gcry_fast_random_poll NEW
gcry_md_debug NEW
gcry_sexp_nth_string NEW
GCRY_MD_SHA224 NEW
GCRY_PK_USAGE_CERT NEW
GCRY_PK_USAGE_AUTH NEW
GCRY_PK_USAGE_UNKN NEW
GCRY_PK_ECDSA NEW
GCRY_CIPHER_SEED NEW
GCRY_CIPHER_CAMELLIA128 NEW
GCRY_CIPHER_CAMELLIA192 NEW
GCRY_CIPHER_CAMELLIA256 NEW
GCRYCTL_FAKED_RANDOM_P NEW
GCRYCTL_PRINT_CONFIG NEW
GCRYCTL_SET_RNDEGD_SOCKET NEW.
gcry_mpi_scan CHANGED: Argument BUFFER is now void*.
gcry_pk_algo_name CHANGED: Returns "?" instead of NULL.
gcry_cipher_algo_name CHANGED: Returns "?" instead of "".
gcry_pk_spec_t CHANGED: Element ALIASES is now const ptr.
gcry_md_write_t CHANGED: Argument BUF is now a const void*.
gcry_md_ctl CHANGED: Argument BUFFER is now void*.
gcry_cipher_encrypt CHANGED: Arguments IN and OUT are now void*.
gcry_cipher_decrypt CHANGED: Arguments IN and OUT are now void*.
gcry_sexp_sprint CHANGED: Argument BUFFER is now void*.
gcry_create_nonce CHANGED: Argument BUFFER is now void*.
gcry_randomize CHANGED: Argument BUFFER is now void*.
gcry_cipher_register CHANGED: Argument ALGORITHM_ID is now int*.
Noteworthy changes in version 1.2.4 (2007-02-01)
------------------------------------------------
* Fixed a bug in the memory allocator which could have been the
reason for some of non-duplicable bugs.
* Other minor bug fixes.
Noteworthy changes in version 1.2.3 (2006-08-28)
------------------------------------------------
* Rewrote gcry_mpi_rshift to allow arbitrary shift counts.
* Minor bug fixes.
and add a new helper target and script, "show-buildlink3", that outputs
a listing of the buildlink3.mk files included as well as the depth at
which they are included.
For example, "make show-buildlink3" in fonts/Xft2 displays:
zlib
fontconfig
iconv
zlib
freetype2
expat
freetype2
Xrender
renderproto
RECOMMENDED is removed. It becomes ABI_DEPENDS.
BUILDLINK_RECOMMENDED.foo becomes BUILDLINK_ABI_DEPENDS.foo.
BUILDLINK_DEPENDS.foo becomes BUILDLINK_API_DEPENDS.foo.
BUILDLINK_DEPENDS does not change.
IGNORE_RECOMMENDED (which defaulted to "no") becomes USE_ABI_DEPENDS
which defaults to "yes".
Added to obsolete.mk checking for IGNORE_RECOMMENDED.
I did not manually go through and fix any aesthetic tab/spacing issues.
I have tested the above patch on DragonFly building and packaging
subversion and pkglint and their many dependencies.
I have also tested USE_ABI_DEPENDS=no on my NetBSD workstation (where I
have used IGNORE_RECOMMENDED for a long time). I have been an active user
of IGNORE_RECOMMENDED since it was available.
As suggested, I removed the documentation sentences suggesting bumping for
"security" issues.
As discussed on tech-pkg.
I will commit to revbump, pkglint, pkg_install, createbuildlink separately.
Note that if you use wip, it will fail! I will commit to pkgsrc-wip
later (within day).
developer is officially maintaining the package.
The rationale for changing this from "tech-pkg" to "pkgsrc-users" is
that it implies that any user can try to maintain the package (by
submitting patches to the mailing list). Since the folks most likely
to care about the package are the folks that want to use it or are
already using it, this would leverage the energy of users who aren't
developers.
Noteworthy changes in version 1.2.2 (2005-10-05}
------------------------------------------------
* Made the RNG immune against fork without exec.
* Minor changes to some function declarations. Buffer arguments are
now typed as void pointer. This should not affect any compilation.
* A bug in the definition of gcry_cipher_register has been fixed.
* Interface changes relative to the 1.2.1 release:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
gcry_cipher_encrypt CHANGED: Arguments IN and OUT are now void*.
gcry_cipher_decrypt CHANGED: Arguments IN and OUT are now void*.
gcry_create_nonce CHANGED: Argument BUFFER is now void*.
gcry_md_ctl CHANGED: Argument BUFFER is now void*.
gcry_sexp_sprint CHANGED: Argument BUFFER is now void*.
gcry_mpi_scan CHANGED: Argument BUFFER is now void*.
gcry_cipher_register CHANGED: Argument ALGORITHM_ID is now int*.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
when other source files depend on fd_set being defined in a local header.
(Required on Interix, which does not expose <sys/select.h>/<sys/time.h>
automagically via other system headers as some OS's do by default.)
* Portability fixes, memory allocation fixes and other minor things.
* Support to build as a W32 static library.
* Changed the way the RNG gets initialized. This allows to keep it
uninitialized as long as no random numbers are used. To override
this, the new macro gcry_fast_random_poll may be used. It is in
general a good idea to spread this macro into the application code
to make sure that these polls happen often enough.
buildlink did not bring in PTH. For example, this made opencdk fail to
build. Fix this by adding mk/pthread.buildlink3.mk to
security/gcrypt/buildlink3.mk
If __sparcv8 is defined, use the SPARC v8 version of udiv_qrnnd. Without
this change, udiv_qrnnd will not be defined using the SPARC v8 udiv
instruction, and the default SPARC v7 case of defining it to be the native
__gmpn_udiv_qrnnd will be used, but this does not exist on SPARC v8, so
at link time, the user will get an error about _gmpn_udiv_qrnnd being
undefined. gcc defines __sparcv8.
From PR pkg/26815.
in the process. (More information on tech-pkg.)
Bump PKGREVISION and BUILDLINK_DEPENDS of all packages using libtool and
installing .la files.
Bump PKGREVISION (only) of all packages depending directly on the above
via a buildlink3 include.
All library names listed by *.la files no longer need to be listed
in the PLIST, e.g., instead of:
lib/libfoo.a
lib/libfoo.la
lib/libfoo.so
lib/libfoo.so.0
lib/libfoo.so.0.1
one simply needs:
lib/libfoo.la
and bsd.pkg.mk will automatically ensure that the additional library
names are listed in the installed package +CONTENTS file.
Also make LIBTOOLIZE_PLIST default to "yes".
file descriptors (0, 1, 2) are open. This avoids multiple warnings issued
under NetBSD about running set[ug]id programs with those descriptors closed.
Fixes PR pkg/26079; although it talks about gaim, the problem is here, in
libgcrypt. Bump PKGREVISION to 1.
