Lasse Collin [Sun, 31 Jan 2010 21:41:29 +0000 (23:41 +0200)]
Updated TODO.
Lasse Collin [Sun, 31 Jan 2010 21:35:04 +0000 (23:35 +0200)]
Mention TODO in README.
Lasse Collin [Sun, 31 Jan 2010 21:31:14 +0000 (23:31 +0200)]
Updated INSTALL.
Lasse Collin [Sun, 31 Jan 2010 21:28:51 +0000 (23:28 +0200)]
Revise the Windows build files.
The old Makefile + config.h was deleted, because it
becomes outdated too easily and building with the
Autotools based build system works fine even on Windows.
windows/build.sh hasn't got much testing, but it should
work to build 32-bit x86 and x86-64 versions of XZ Utils
using MSYS, MinGW or MinGW-w32, and MinGW-w64.
windows/INSTALL-Windows.txt describes what packages are
needed and how to install them.
windows/README-Windows.txt is a readme file for the binary
package that build.sh hopefully builds.
There are no instructions about using Autotools for now,
so those using a git snapshot may want to run
"autoreconf -fi && ./configure && make mydist" on a UN*X
box and then copy the resulting .tar.gz to a Windows.
Lasse Collin [Sun, 31 Jan 2010 17:52:38 +0000 (19:52 +0200)]
Select the default integrity check type at runtime.
Previously it was set statically to CRC64 or CRC32
depending on options passed to the configure script.
Lasse Collin [Sun, 31 Jan 2010 16:17:50 +0000 (18:17 +0200)]
Improve displaying of the memory usage limit.
Lasse Collin [Sun, 31 Jan 2010 10:53:56 +0000 (12:53 +0200)]
Don't use uninitialized sigset_t.
If signal handlers haven't been established, then it's
useless to try to block them, especially since the sigset_t
used for blocking hasn't been initialized yet.
Lasse Collin [Sun, 31 Jan 2010 10:01:54 +0000 (12:01 +0200)]
Delay opening the destionation file and other fixes.
The opening of the destination file is now delayed a little.
The coder is initialized, and if decompressing, the memory
usage of the first Block compared against the memory
usage limit before the destination file is opened. This
means that if --force was used, the old "target" file won't
be deleted so easily when something goes wrong very early.
Thanks to Mark K for the bug report.
The above fix required some changes to progress message
handling. Now there is a separate function for setting and
printing the filename. It is used also in list.c.
list_file() now handles stdin correctly (gives an error).
A useless check for user_abort was removed from file_io.c.
Lasse Collin [Fri, 29 Jan 2010 20:48:04 +0000 (22:48 +0200)]
Add list.h to src/xz/Makefile.am.
This should have been already in
0bc9eab243dee3be764b3530433a7fcdc3f7c6a1.
Lasse Collin [Fri, 29 Jan 2010 11:24:27 +0000 (13:24 +0200)]
Add lzmainfo.1 to manfiles list to convert to .txt and .pdf.
Lasse Collin [Wed, 27 Jan 2010 14:42:11 +0000 (16:42 +0200)]
Silence two compiler warnings on DOS-like systems.
Lasse Collin [Wed, 27 Jan 2010 11:31:03 +0000 (13:31 +0200)]
Use PACKAGE_URL instead of custom PACKAGE_HOMEPAGE.
Lasse Collin [Tue, 26 Jan 2010 21:37:46 +0000 (23:37 +0200)]
Add a missing space to an error message.
Thanks to Robert Readman.
Lasse Collin [Tue, 26 Jan 2010 20:53:37 +0000 (22:53 +0200)]
Use past tense in error message in io_unlink().
Added a note to translators too.
Thanks to Robert Readman.
Lasse Collin [Tue, 26 Jan 2010 13:42:24 +0000 (15:42 +0200)]
Fix too small static buffer in util.c.
This was introduced in
0dd6d007669b946543ca939a44243833c79e08f4 two days ago.
Lasse Collin [Tue, 26 Jan 2010 12:46:43 +0000 (14:46 +0200)]
Minor comment fix.
Lasse Collin [Sun, 24 Jan 2010 21:50:54 +0000 (23:50 +0200)]
Add initial version of xz --list.
This is a bit rough but should be useful for basic things.
Ideas (with detailed examples) about the output format are
welcome.
The output of --robot --list is not necessarily stable yet,
although I don't currently have any plans about changing it.
The man page hasn't been updated yet.
Lasse Collin [Sun, 24 Jan 2010 20:46:11 +0000 (22:46 +0200)]
Add io_pread().
It will be used by --list.
Lasse Collin [Sun, 24 Jan 2010 20:45:14 +0000 (22:45 +0200)]
Set LC_NUMERIC=C when --robot is used.
It is to ensure that floating point numbers
will always have a dot as the decimal separator.
Lasse Collin [Sun, 24 Jan 2010 14:57:40 +0000 (16:57 +0200)]
Some improvements to printing sizes in xz.
Lasse Collin [Wed, 20 Jan 2010 20:02:35 +0000 (22:02 +0200)]
Fix a typo in README.
Thanks to R. Bijker.
Lasse Collin [Sun, 17 Jan 2010 09:59:54 +0000 (11:59 +0200)]
Updated windows/Makefile.
Thanks to Dan Shechter for the patch.
It is likely that windows/Makefile will be removed
completely, because Autotols based build nowadays
works well with both 32-bit and 64-bit MinGW (I
just need to update the docs).
Lasse Collin [Fri, 15 Jan 2010 09:05:11 +0000 (11:05 +0200)]
Update the xz man page to match the previous two commits.
Lasse Collin [Wed, 13 Jan 2010 17:10:25 +0000 (19:10 +0200)]
Don't read compressed data from a terminal or write it
to a terminal even if --force is specified.
It just seems more reasonable this way.
The new behavior matches bzip2. The old one matched gzip.
Lasse Collin [Wed, 13 Jan 2010 16:12:40 +0000 (18:12 +0200)]
Don't compress or decompress special files unless writing
to stdout even if --force is used.
--force will still enable compression of symlinks, but only
in case they point to a regular file.
The new way simply seems more reasonable. It matches gzip's
behavior while the old one matched bzip2's behavior.
Lasse Collin [Tue, 12 Jan 2010 14:30:33 +0000 (16:30 +0200)]
Updated THANKS.
Lasse Collin [Tue, 12 Jan 2010 14:18:14 +0000 (16:18 +0200)]
Add IRIX-specific code to tuklib_physmem and tuklib_cpucores.
This is untested but it will get tested soon and, if needed,
fixed before 5.0.0.
Thanks to Stuart Shelton.
Lasse Collin [Thu, 31 Dec 2009 22:29:10 +0000 (00:29 +0200)]
Fix _memconfig() functions.
This affects lzma_memusage() and lzma_memlimit_get().
Lasse Collin [Thu, 31 Dec 2009 20:45:53 +0000 (22:45 +0200)]
Revised the Index handling code.
This breaks API and ABI but most apps are not affected
since most apps don't use this part of the API. You will
get a compile error if you are using anything that got
broken.
Summary of changes:
- Ability to store Stream Flags, which are needed
for random-access reading in multi-Stream files.
- Separate function to set size of Stream Padding.
- Iterator structure makes it possible to read the same
lzma_index from multiple threads at the same time.
- A lot faster code to locate Blocks.
- Removed lzma_index_equal() without adding anything
to replace it. I don't know what it should do exactly
with the new features and what actually needs this
function in the first place other than test_index.c,
which now has its own code to compare lzma_indexes.
Lasse Collin [Thu, 31 Dec 2009 19:13:25 +0000 (21:13 +0200)]
Remove c-format tag in cs.po.
It was fixed in the C code earlier.
Lasse Collin [Thu, 31 Dec 2009 19:11:05 +0000 (21:11 +0200)]
Add missing lzma_nothrow in filter.h.
Lasse Collin [Wed, 9 Dec 2009 16:13:44 +0000 (18:13 +0200)]
Remove redefinition of _(msgid) macro from lzmainfo.c.
Jonathan Nieder [Wed, 9 Dec 2009 01:41:57 +0000 (19:41 -0600)]
update po/.gitignore
Since the *.gmo files are deleted by the maintainer-clean target,
I assume they are not meant to be tracked.
Also add the other files listed in the Makefile’s clean targets
(stamp-poT, xz.po, xz.[12].po, *.new.po, xz.mo) to make sure they
are not accidentally tracked. Most of these are intermediate
files that would not appear unless a build is interrupted or
fails.
Split the list of untracked files by origin to make it easier to
tell if files are missing in the future.
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Lasse Collin [Tue, 8 Dec 2009 22:38:55 +0000 (00:38 +0200)]
Always rely on GCC's auto-import on Windows.
I understood that this is nicer, because then people
don't need to worry about the LZMA_API_STATIC macro.
Thanks to Charles Wilson and Keith Marshall.
Lasse Collin [Mon, 7 Dec 2009 19:46:53 +0000 (21:46 +0200)]
Fix file_io.c on DOS-like systems.
The problem was introduced when adding sparse file
support in
465d1b0d6518c5d980f2db4c2d769f9905bdd902.
Thanks to Charles Wilson.
Lasse Collin [Mon, 7 Dec 2009 18:54:21 +0000 (20:54 +0200)]
Add Czech translation.
Thanks to Marek Černocký.
Other people planning to translate xz: Note that the
messages are a little bit in flux still. Translations
are still welcome, just be prepared to some extra work
in case there are changes.
Lasse Collin [Mon, 7 Dec 2009 18:32:08 +0000 (20:32 +0200)]
Add a note for translators to add a bug reporting address
for translation bugs.
Lasse Collin [Mon, 7 Dec 2009 18:07:02 +0000 (20:07 +0200)]
Prevent xgettext from taking one regular string as a C format string.
Thanks to Marek Černocký.
Lasse Collin [Sat, 28 Nov 2009 15:45:22 +0000 (17:45 +0200)]
Remove duplicate code in io_open_dest().
Fix a missing _() in the error message too.
Lasse Collin [Thu, 26 Nov 2009 08:11:23 +0000 (10:11 +0200)]
Typo fix to sysdefs.h.
Thanks to Jonathan Nieder.
Lasse Collin [Thu, 26 Nov 2009 08:10:36 +0000 (10:10 +0200)]
Fix a memory leak in test_index.c.
This was introduced in
bd13b04e202b6f495a68eb0766f97085b7c50a06.
Thanks to Jim Meyering for noticing it.
Lasse Collin [Wed, 25 Nov 2009 12:22:19 +0000 (14:22 +0200)]
Add missing error check to coder.c.
With bad luck this could cause a segfault due to
reading (but not writing) past the end of the buffer.
Lasse Collin [Wed, 25 Nov 2009 11:04:10 +0000 (13:04 +0200)]
Fix bugs in lzma_index_read() and lzma_index_cat().
lzma_index_read() didn't skip over Stream Padding
if it was the first record in the Index.
lzma_index_cat() didn't combine small Indexes correctly.
The test suite was updated to check for these bugs.
These bugs didn't affect the xz command line tool or
most users of liblzma in any way.
Lasse Collin [Wed, 25 Nov 2009 10:52:56 +0000 (12:52 +0200)]
Index decoder fixes.
The Index decoder code didn't perfectly match the API docs,
which said that *i will be set to point to the decoded Index
only after decoding has succeeded. The docs were a bit unclear
too.
Now the decoder will initially set *i to NULL. *i will be set
to point to the decoded Index once decoding has succeeded.
This simplifies applications too, since it avoids dangling
pointers.
Lasse Collin [Wed, 25 Nov 2009 09:19:20 +0000 (11:19 +0200)]
Create sparse files by default when decompressing into
a regular file.
Sparse file creation can be disabled with --no-sparse.
I don't promise yet that the name of this option won't
change before 5.0.0. It's possible that the code, that
checks when it is safe to use sparse output on stdout,
is not good enough, and a more flexible command line
option is needed to configure sparse file handling.
Lasse Collin [Sun, 22 Nov 2009 10:43:06 +0000 (12:43 +0200)]
Updated THANKS.
Lasse Collin [Sun, 22 Nov 2009 10:05:33 +0000 (12:05 +0200)]
Add missing consts to pointer casts.
Lasse Collin [Sun, 22 Nov 2009 10:00:30 +0000 (12:00 +0200)]
Enable assembler code only if it is known to work
on that operating system.
I'm too lazy to think how to make a good Autoconf test
for this and it's not that important anyway.
No longer define HAVE_ASM_X86 or HAVE_ASM_X86_64.
Inline assembler (if any) is used if a macro like
__i386__ or __x86_64__ is defined.
Lasse Collin [Sun, 22 Nov 2009 09:55:03 +0000 (11:55 +0200)]
Make fastpos.h use tuklib_integer.h instead of bsr.h
when --enable-small has been specified.
Lasse Collin [Sun, 22 Nov 2009 09:52:30 +0000 (11:52 +0200)]
Update tuklib_integer.h with bit scan functions.
Thanks to Joachim Henke for the original patch.
Lasse Collin [Fri, 20 Nov 2009 10:51:19 +0000 (12:51 +0200)]
Update tuklib_cpucores.m4 and tuklib_physmem.m4 from tuklib,
which now use AC_CACHE_CHECK. Using the cache variable,
configure now warns if there is no method to detect the amount
of RAM and recommends using --enable-assume-ram.
Lasse Collin [Mon, 16 Nov 2009 16:16:45 +0000 (18:16 +0200)]
Add support for --info-memory and --robot to xz.
Currently --robot works only with --info-memory and
--version. --help and --long-help work too, but --robot
has no effect on them.
Thanks to Jonathan Nieder for the original patches.
Lasse Collin [Sun, 15 Nov 2009 10:54:45 +0000 (12:54 +0200)]
Fix wrong indentation caused by incorrect settings
in the text editor.
Lasse Collin [Sun, 15 Nov 2009 10:40:17 +0000 (12:40 +0200)]
Add lzma_physmem().
I had hoped to keep liblzma as purely a compression
library as possible (e.g. file I/O will go into
a different library), but it seems that applications
linking agaisnt liblzma need some way to determine
the memory usage limit, and knowing the amount of RAM
is one reasonable way to help making such decisions.
Thanks to Jonathan Nieder for the original patch.
Lasse Collin [Sat, 14 Nov 2009 18:21:19 +0000 (20:21 +0200)]
Updated THANKS.
Lasse Collin [Sat, 14 Nov 2009 18:20:03 +0000 (20:20 +0200)]
Some updates to xz man page.
Lasse Collin [Sat, 14 Nov 2009 17:51:03 +0000 (19:51 +0200)]
Fix description of --memory in --long-help.
Lasse Collin [Sat, 14 Nov 2009 17:45:39 +0000 (19:45 +0200)]
Update the debug programs so that they compile again.
Lasse Collin [Sat, 14 Nov 2009 16:59:19 +0000 (18:59 +0200)]
Fix a design error in liblzma API.
Originally the idea was that using LZMA_FULL_FLUSH
with Stream encoder would read the filter chain
from the same array that was used to intialize the
Stream encoder. Since most apps wouldn't use
LZMA_FULL_FLUSH, most apps wouldn't need to keep
the filter chain available after initializing the
Stream encoder. However, due to my mistake, it
actually required keeping the array always available.
Since setting the new filter chain via the array
used at initialization time is not a nice way to do
it for a couple of reasons, this commit ditches it
and introduces lzma_filters_update(). This new function
replaces also the "persistent" flag used by LZMA2
(and to-be-designed Subblock filter), which was also
an ugly thing to do.
Thanks to Alexey Tourbin for reminding me about the problem
that Stream encoder used to require keeping the filter
chain allocated.
Lasse Collin [Sat, 17 Oct 2009 08:11:58 +0000 (11:11 +0300)]
Fix wrong function name in the previous commit.
It was meant to be lzma_filters_copy(), not lzma_filters_dup().
Lasse Collin [Fri, 16 Oct 2009 22:47:07 +0000 (01:47 +0300)]
Add lzma_filters_copy().
This will be needed internally by liblzma once I fix
a design mistake in the encoder API. This function may
be useful to applications too so it's good to export it.
Jonathan Nieder [Fri, 16 Oct 2009 01:44:13 +0000 (20:44 -0500)]
Escape dashes in xzmore.1
A minus sign is larger, easier to see in a printout, and more
likely to use the same glyph as ASCII hyphen-minus in a terminal
than a hyphen. Since broken manual pagers do not find hyphens
when the user searches for a hyphen-minus, minus signs are also
easier to search for. So use minus signs instead of hyphens to
render sample terminal output.
Lasse Collin [Fri, 16 Oct 2009 17:35:39 +0000 (20:35 +0300)]
Mention --check=none in --long-help. It was already in
the man page though.
Thanks to Jim Meyering for noticing this.
Lasse Collin [Sun, 4 Oct 2009 19:57:12 +0000 (22:57 +0300)]
Use a tuklib module for integer handling.
This replaces bswap.h and integer.h.
The tuklib module uses <byteswap.h> on GNU,
<sys/endian.h> on *BSDs and <sys/byteorder.h>
on Solaris, which may contain optimized code
like inline assembly.
Lasse Collin [Fri, 2 Oct 2009 11:35:56 +0000 (14:35 +0300)]
Add support for --enable-assume-ram=SIZE.
Lasse Collin [Fri, 2 Oct 2009 08:28:17 +0000 (11:28 +0300)]
Use unaligned access (if possible) on both endiannesses
in lz_encoder_hash.h.
Lasse Collin [Fri, 2 Oct 2009 08:03:26 +0000 (11:03 +0300)]
Make liblzma produce the same output on both endiannesses.
Seems that it is a problem in some cases if the same
version of XZ Utils produces different output on different
endiannesses, so this commit fixes that problem. The output
will still vary between different XZ Utils versions, but I
cannot avoid that for now.
This commit bloatens the code on big endian systems by 1 KiB,
which should be OK since liblzma is bloated already. ;-)
Mike Frysinger [Sat, 26 Sep 2009 16:51:50 +0000 (12:51 -0400)]
add lzmainfo to gitignore
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Lasse Collin [Sun, 27 Sep 2009 08:53:36 +0000 (11:53 +0300)]
Updated THANKS.
Lasse Collin [Sun, 27 Sep 2009 08:48:54 +0000 (11:48 +0300)]
Work around a bug in Interix header files.
Thanks to Markus Duft for the patch.
Lasse Collin [Thu, 24 Sep 2009 14:50:17 +0000 (17:50 +0300)]
Fix an error in OpenVMS-specific code.
Thanks to Jouk Jansen.
Lasse Collin [Tue, 22 Sep 2009 15:59:56 +0000 (18:59 +0300)]
Added OpenVMS-specific information to INSTALL.
Lasse Collin [Tue, 22 Sep 2009 11:03:02 +0000 (14:03 +0300)]
Better fixes for OpenVMS support.
Thanks to Jouk Jansen.
Lasse Collin [Tue, 22 Sep 2009 10:40:19 +0000 (13:40 +0300)]
Avoid non-standard preprocessor construct.
Thanks to Jouk Jansen.
Lasse Collin [Mon, 21 Sep 2009 16:50:09 +0000 (19:50 +0300)]
Make sure that TUKLIB_DOSLIKE doesn't get defined on Cygwin.
Thanks to Charles Wilson.
Lasse Collin [Sat, 19 Sep 2009 06:47:30 +0000 (09:47 +0300)]
Various changes.
Separate a few reusable components from XZ Utils specific
code. The reusable code is now in "tuklib" modules. A few
more could be separated still, e.g. bswap.h.
Fix some bugs in lzmainfo.
Fix physmem and cpucores code on OS/2. Thanks to Elbert Pol
for help.
Add OpenVMS support into physmem. Add a few #ifdefs to ease
building XZ Utils on OpenVMS. Thanks to Jouk Jansen for the
original patch.
Lasse Collin [Tue, 15 Sep 2009 18:07:23 +0000 (21:07 +0300)]
Fix incorrect use of "restrict".
Lasse Collin [Sat, 12 Sep 2009 11:09:17 +0000 (14:09 +0300)]
Fix GCC version check for nothrow attribute.
Lasse Collin [Sat, 12 Sep 2009 11:08:15 +0000 (14:08 +0300)]
Updated THANKS.
Lasse Collin [Sat, 12 Sep 2009 11:07:36 +0000 (14:07 +0300)]
A few grammar fixes.
Thanks to Christian Weisgerber for pointing out some of these.
Lasse Collin [Fri, 11 Sep 2009 14:08:15 +0000 (17:08 +0300)]
Updated THANKS.
Lasse Collin [Fri, 11 Sep 2009 14:06:32 +0000 (17:06 +0300)]
Add PACKAGE_HOMEPAGE to {windows,dos}/config.h to fix build errors.
Lasse Collin [Fri, 11 Sep 2009 07:24:09 +0000 (10:24 +0300)]
Use $(LN_EXEEXT) in symlinks to executables.
This fixes "make install" on operating systems using
a suffix for executables.
Cygwin is treated specially. The symlink names won't have
.exe suffix even though the executables themselves have.
Thanks to Charles Wilson.
Lasse Collin [Fri, 11 Sep 2009 06:25:09 +0000 (09:25 +0300)]
Fix a couple of warnings.
Lasse Collin [Sat, 5 Sep 2009 15:39:21 +0000 (18:39 +0300)]
Add OS/2-specific code to physmem.h.
Also move DJGPP-specific code near the code meant
for other DOS-like systems.
Lasse Collin [Fri, 4 Sep 2009 22:21:15 +0000 (01:21 +0300)]
Updated THANKS.
Lasse Collin [Fri, 4 Sep 2009 22:20:29 +0000 (01:20 +0300)]
Use sysctl() != -1 instead of !sysctl() to check if
the function call succeeded.
NetBSD 4.0 returns positive values on success, but
NetBSD Current and FreeBSD return zero. OpenBSD's
man page doesn't tell what sysctl() returns on
success. All these BSDs return -1 on error.
Thanks to Robert Elz and Thomas Klausner.
Lasse Collin [Wed, 2 Sep 2009 06:43:51 +0000 (09:43 +0300)]
Mention in INSTALL that --enable-small doesn't modify CFLAGS.
Lasse Collin [Tue, 1 Sep 2009 17:40:01 +0000 (20:40 +0300)]
Refactored option parsing.
Lasse Collin [Tue, 1 Sep 2009 17:23:30 +0000 (20:23 +0300)]
Fix options parsing bug in xz.
xz used to reject "xz --lzma2=pb=2," while
"xz --lzma2=pb=2,," worked. Now both work.
Lasse Collin [Tue, 1 Sep 2009 17:20:19 +0000 (20:20 +0300)]
Updated TODO.
Lasse Collin [Mon, 31 Aug 2009 18:59:25 +0000 (21:59 +0300)]
Revert
43f44160b1ddcbf7e5205c37db09b3bebe7226f9
and use a fix that works on all systems using
GNU assembler.
Maybe the assembler code is used e.g. on Solaris x86
but let's worry about it if this doesn't work on it.
Lasse Collin [Sun, 30 Aug 2009 14:29:19 +0000 (17:29 +0300)]
Updated THANKS.
Lasse Collin [Sun, 30 Aug 2009 14:28:52 +0000 (17:28 +0300)]
Add more OS/2 specific info to INSTALL.
Lasse Collin [Sat, 29 Aug 2009 11:43:52 +0000 (14:43 +0300)]
Use even more hackish way to support thousand separators.
Seems that in addition on Windows and DOS, also OpenBSD
lacks support for %'d style printf() format strings.
So far that is the only modern POSIX-like system I know
with this problem, but after this hack, the thousand
separator shouldn't be a problem on any system.
Maybe testing if a format string like %'d produces
reasonable output is invoking undefined behavior on some
systems, but so far all the problematic systems I've tried
just print the raw format string (e.g. %'d prints 'd).
Maybe Autoconf test would have been better, but this
hack works also for cross-compilation, and avoids
recompilation in case the system libc starts to support
the thousand separator.
Lasse Collin [Sat, 29 Aug 2009 10:42:56 +0000 (13:42 +0300)]
Updated THANKS.
Lasse Collin [Sat, 29 Aug 2009 10:39:21 +0000 (13:39 +0300)]
Fix sysctl() usage.
This fixes build on *BSDs and Darwin.
Thanks to Jukka Salmi for the patches.
Richard Koch reported the problem too.
Lasse Collin [Sat, 29 Aug 2009 10:35:23 +0000 (13:35 +0300)]
Fix x86 assembler on GCC 3.
Thanks to Karl Berry.
Lasse Collin [Thu, 27 Aug 2009 15:36:59 +0000 (18:36 +0300)]
"make dist" fixes
Lasse Collin [Thu, 27 Aug 2009 14:08:33 +0000 (17:08 +0300)]
Update xz man page date.