As camlp5 is required with ocaml 3.10, bring it as a dependency anyway,
instead of requiring ocaml 3.10
Changes include:
* Installation
- Support for compilation with ocaml 3.10 and (transitional) camlp5.
- Many bugs have been fixed (cf coq-bugs web page)
- All known failures of ROmega have been fixed. It should now be a
faithful and quicker replacement for Omega (except when nat parts
are involved). ROmega and Omega now handle <->.
- Better computational behavior of some constants (eq_nat_dec and
le_lt_dec more efficient, Z_lt_le_dec and Positive_as_OT.compare
transparent) [exceptionally source of incompatibilities].
- Loading FSets/FMap used to open unwanted scopes of integer datatypes
(see bug #1347). These scopes may need to be manually opened now.
Camlp5 is a preprocessor and pretty-printer for OCaml programs. It also
provides parsing and printing tools.
As a preprocessor, it allows to:
- extend the syntax of OCaml,
- redefine the whole syntax of the language.
As a pretty printer, it allows to:
- display OCaml programs in an elegant way,
- convert from one syntax to another,
- check the results of syntax extensions.
Camlp5 also provides some parsing and pretty printing tools:
- extensible grammars
- extensible printers
- stream parsers and lexers
- pretty print module
It works as a shell command and can also be used in the OCaml toplevel.
Change in GNU Prolog version 1.3.0:
* official release (Jan 4 2007)
* change error messages emitted by the compiler to follow GNU standards
* modify doc (mainly rename manual.xxx to gprolog.xxx)
* add DESTDIR variable support in main Makefile for staged installs
* fix a bug with Prolog floats in x86_64/linux (bad stack alignment)
* port for ix86/darwin (Mac OS X)
* add check target to main Makefile
* improve Win32 ports (cygwin, MinGW, MSVC 6.0 and 2005 Express Edition)
(MSVC port uses MinGW as.exe instead of nasm.exe - named mingw-as.exe
provided in the setup.exe)
* rename call/2 to call_det/2
* implement call/2-11 as will be defined in the next standard
* fix various problems when compiling with gcc 4 (gcc 4.1.1)
* emit .note.GNU-stack to mark the stack as no executable
in x86_any.c, x86_64_any.c and powerpc_any.c
* change the way objects are found (obj_chain.c) using gcc ctors
* use Doug Lea malloc for OpenBSD (problem with malloc using mmap)
* fix problems in various ports:
alpha/linux, powerpc/darwin (Mac OS X), sparc/solaris, ix86/OpenBSD
Change in GNU Prolog version 1.2.19:
* fix 2 bugs in global variables
Change in GNU Prolog version 1.2.18:
* fix problem when compiling with gcc 3.4.0
* fix bug in term comparison involving negative integers
* add consult, ... and fix minor bugs in the Win32 GUI console menu
* fix the stack overflow detection under cygwin
* port to ix86/MinGW - many thanks to:
Cesar Rabak <csrabak@ig.com.br>
* fix a problem in the port to x86/OpenBSD
* port to sparc/NetBSD and powerpc/NetBSD - many thanks to:
Jason Beegan <jtb@netbsd.org>
* fix a bug in =../2 involving FD variables
* fix a bug in arithmetics (in float_{integer/fractional}_part)
* fix a bug in FD solver (wrong union with a singleton)
Change in GNU Prolog version 1.2.17:
* change configure.in: by default ebp is not used
* fix a but with CTRL+C handler not reinstalled
* fix a bug with _XXX (re)displayed under the top-level
* port to x86_64/linux - many thanks to:
Gwenole Beauchesne <gbeauchesne@mandrakesoft.com>
Changes since 5.2.1:
5.2.2 - added SHA-224
- put SHA-256, SHA-384, SHA-512, RSASSA-PSS into DLL
5.2.3 - fixed issues with FIPS algorithm test vectors
- put RSASSA-ISO into DLL
5.3 - ported to MSVC 2005 with support for x86-64
- added defense against AES timing attacks, and more AES test vectors
- changed StaticAlgorithmName() of Rijndael to "AES", CTR to "CTR"
5.4 - added Salsa20
- updated Whirlpool to version 3.0
- ported to GCC 4.1, Sun C++ 5.8, and Borland C++Builder 2006
5.5 - added VMAC and Sosemanuk (with x86-64 and SSE2 assembly)
- improved speed of integer arithmetic, AES, SHA-512, Tiger, Salsa20,
Whirlpool, and PANAMA cipher using assembly (x86-64, MMX, SSE2)
- optimized Camellia and added defense against timing attacks
- updated benchmarks code to show cycles per byte and to time key/IV setup
- started using OpenMP for increased multi-core speed
- enabled GCC optimization flags by default in GNUmakefile
- added blinding and computational error checking for RW signing
- changed RandomPool, X917RNG, GetNextIV, DSA/NR/ECDSA/ECNR to reduce
the risk of reusing random numbers and IVs after virtual machine state
rollback
- changed default FIPS mode RNG from AutoSeededX917RNG<DES_EDE3> to
AutoSeededX917RNG<AES>
- fixed PANAMA cipher interface to accept 256-bit key and 256-bit IV
- moved MD2, MD4, MD5, PanamaHash, ARC4, WAKE_CFB into the namespace "Weak"
- removed HAVAL, MD5-MAC, XMAC
5.5.1 - fixed VMAC validation failure on 32-bit big-endian machines
5.5.2 - ported x64 assembly language code for AES, Salsa20, Sosemanuk, and Panama
to MSVC 2005 (using MASM since MSVC doesn't support inline assembly on x64)
- fixed Salsa20 initialization crash on non-SSE2 machines
- fixed Whirlpool crash on Pentium 2 machines
- fixed possible branch prediction analysis (BPA) vulnerability in
MontgomeryReduce(), which may affect security of RSA, RW, LUC
- fixed link error with MSVC 2003 when using "debug DLL" form of runtime library
- fixed crash in SSE2_Add on P4 machines when compiled with
MSVC 6.0 SP5 with Processor Pack
- added support for newly released compilers: MSVC 2008, GCC 4.2, Sun CC 5.9,
Intel C++ Compiler 10.0, and Borland C++Builder 2007
2.5
Keith Packard (4):
Document several function return values (Bug 13145).
Document that Match calls FcFontRenderPrepare (bug 13162).
Document that FcConfigGetFonts returns the internal fontset (bug 13197)
Revert "Remove fcprivate.h, move the remaining macros to fcint.h."
Tor Lillqvist (1):
Workaround for stat() brokenness in Microsoft's C library (bug 8526)
2.4.92 (2.5 RC2)
Behdad Esfahbod (14):
Make fc-match --sort call FcFontRenderPrepare.
Port fonts-persian.conf to new alias syntax with binding="same"
Fix trivial bugs in edit-sgml.c
Add FcGetLangs() and FcLangGetCharSet().
Add/update config files from Fedora.
Split 40-generic.conf into 40-nonlatin.conf and 45-latin.conf
Use binding="same" in 30-urw-aliases.conf and remove duplicate entries.
Remove redundant/obsolete comments from conf files.
Remove 20-lohit-gujarati.conf. It's covered by 25-unhint-nonlatin.conf now.
Oops, fix Makefile.am.
Remove 25-unhint-nonlatin.conf from default configuration by not linking it.
Fix documented conf-file naming format in README
Remove list of available conf files from README.
Simplify/improve 30-metric-aliases.conf
Keith Packard (25):
Also check configDirs mtimes in FcConfigUptoDate
Respect "binding" attribute in <alias> entries.
Correct documentation for FcAtomicLock (Bug 12947).
Remove fcprivate.h, move the remaining macros to fcint.h.
Correct documentation for FcConfigUptoDate (bug 12948).
Document skipping of fonts from FcFileScan/FcDirScan.
Make file_stat argument to FcDirCacheLoadFile optional.
Clean up exported names in fontconfig.h.
Track line numbers in sgml edit tool input.
Typo error in function name: Inverval -> interval
Don't check cache file time stamps when cleaning cache dir.
Use FcLangDifferentTerritory instead of FcLangDifferentCountry.
Verify documentation covers exposed symbols.
Document previously undocumented functions. (bug 12963)
Update documentation for FcStrCopyFilename (bug 12964).
Update documentation for stale FcConfigGetConfig function.
Have FcConfigSetCurrent accept the current configuration and simply return
Remove references to FcConfigParse and FcConfigLoad.
Replace incorrect documentation uses of 'char' with 'FcChar8' (bug 13002).
Fix formatting syntax in doc/fccache.fncs
Generate fccache.sgml, fcdircache.sgml and fclangset.sgml.
Formatting syntax mistake in doc/fclangset.fncs.
Link new function documentation into the fontconfig-devel.sgml
Ignore new generated documentation
Export FcConfig{G,S}etRescanInverval from .so, mark as deprecated.
2.4.91 (2.5 RC1)
Behdad Esfahbod (1):
Update CaseFolding.txt to Unicode 5.1.0
Dwayne Bailey (1):
Add/fix *.orth files for South African languages
Hideki Yamane (1):
Handle Japanese fonts better. (debian bug #435971)
Keith Packard (32):
rehash increment could be zero, causing rehash infinite loop.
Work around FreeType bug when glyph name buffer is too small.
Free temporary string in FcDirCacheUnlink (Bug #11758)
Fix ChangeLog generation to avoid circular make dependency
Store font directory mtime in cache file.
Comment about mmaping cache files was misleading.
Make FC_FULLNAME include all fullname entries, elide nothing. [bug 12827]
Remove unneeded call to access(2) in fc-cache.
Improve verbose messages from fc-cache.
Verbose message about cleaning directories was imprecise
Don't use X_OK bit when checking for writable directories (bug 12438)
Have fc-cache remove invalid cache files from cache directories.
FcConfigParseAndLoad doc was missing the last param.
Place language name in constant array instead of pointer.
Must not insert cache into hash table before completely validating.
Eliminate relocations for glyph name table.
Eliminate relocations from FcCodePageRange structure (bug 10982).
Leave generated headers out of distribution (bug 12734).
Move <cachedir> elements to the end of fonts.conf.
Add BRAILLE PATTERN BLANK to list of blank glyphs.
Replace makealias pattern with something supported by POSIX grep (bug 11083)
FcInit should return FcFalse when FcInitLoadConfigAndFonts fails. (bug 10976)
There is no U+1257 (bug 10899).
Spelling errors in documentation. (bug 10879).
Oops. Left debugging printf in previous commit.
Handle UltraBlack weight.
Fix parallel build in fontconfig/docs (bug 10481).
Distribute man source files for command line programs (bug 9678).
Ensure weight/slant values present even when style is supplied (bug 9313).
fontconfig needs configure option to use gnu iconv (bug 4083).
Match 'ultra' on word boundaries to detect ultra bold fonts. (bug 2511)
Build fix for Solaris 10 with GCC.
Mike FABIAN (1):
Avoid crashes if config files contain junk.
Stephan Kulow (1):
Make FcPatternDuplicate copy the binding instead of always using Strong.
Tilman Sauerbeck (2):
Store FcNoticeFoundries in read-only memory.
Store FcVendorFoundries in read-only memory.
pkg-config 0.22
===
- Make Requires.private a whole lot more useful by traversing the
whole tree, not just the top-level, for Cflags.
- Add support for using the system glib.
- Update URL to pkg-config website
- Fix some win32 problems.
- Other minor fixes.
1.6.1
New versions of OpenEXR and CTL are now available. This release
fixes a buffer overrun in OpenEXR and a Windows build problem in
CTL, and it removes a few unnecessary files from the .tar.gz
packages.
1.6.0
* Reduced generational loss in B44- and B44A-compressed images.
* Added B44A compression. This is a variation of B44, but with
a better compression ratio for images with large uniform areas,
such as in an alpha channel.
* Bug fixes.
1.5.0
* OpenEXR supports a new image compression method, called B44.
It has a fixed compression rate of 2.28:1, or 4.57:1 if used
in combination with luminance/chroma encoding. B44-compressed
images can be uncompressed fast enough to support real-time
playback of image sequences.
* The new playexr program plays back moving image sequences.
Playexr is multi-threaded and utilizes the threading capabilities
of the IlmImf library that were introduced in OpenEXR 1.3.0.
The program plays back B44-compressed images with fairly
high-resolution in real time on commodity hardware.
* The playexr program and a new version of the existing exrdisplay
image viewer both support color rendering via color transforms
written in the new Color Transformation Language or CTL. CTL
is not part of OpenEXR; it will be released separately. CTL
support in playexr and exrdisplay is optional; the programs
can be built and will run without CTL.
* In preparation for the release of CTL, OpenEXR has been split
into three separate packages:
o IlmBase 0.9.0 includes the Half, Iex, Imath and IlmThread
libraries
o OpenEXR 1.5.0 includes the IlmImf library, programming
examples and utility programs such as exrheader or
exrenvmap
o OpenEXRViewers 0.9.0 includes the playexr and exrdisplay
programs
* The "Technical Introduction to OpenEXR" document now includes
a recommendation for storing CIE XYZ pixel data in OpenEXR
files.
* A new "OpenEXR Image Viewing Software" document describes
the playexr and exrdisplay programs. It briefly explains
real-time playback and color rendering, and includes recommendations
for testing if other image viewing software displays OpenEXR
images correctly.
* The OpenEXR sample image set now includes B44-compressed
files and files with CIE XYZ pixel data.
Half is a class that encapsulates our 16-bit floating-point format.
IlmThread is a thread abstraction library for use with OpenEXR
and other software packages. It currently supports pthreads and
Windows threads.
Imath implements 2D and 3D vectors, 3x3 and 4x4 matrices, quaternions
and other useful 2D and 3D math functions.
Iex is an exception-handling library.
- fix man pages and improve consistency
- improved handling for malformed IXFRs
- report source and zone for denied AXFR requests
- improved handling of malformed nsec3 records
- fix ignored return value in region-allocator.
- portability fixes
- run format auction only once, simpler and more reliable
- speed up handling of uncompressed archives
- fix skipping of archive entries of exactly 4GB size on 32bit platforms
- fix reading files > 8 GB from tar archives
- many improvements for bsdcpio
