In mydist the point is to check using the file list from the Git
repository. In dist-hook it is to check that the TARBALL_IGNORE
patterns work when the .git dir or the "git" command aren't available.
Refuse to create a distribution tarball if license issues are found.
This helps in spotting files that lack SPDX license identifier
and which haven't been explicitly white listed either. The script
requires the .git directory to be present as only the files that
are in the Git repository are checked.
XZ Utils isn't FSFE REUSE compliant for now.
The po4a directory is in EXTRA_DIST and thus all files there
are included in the package. .gitignore doesn't belong in the
package so keep that file out of the po4a directory.
Now the test scripts detect both
#define HAVE_DECODER_ARM
#define HAVE_DECODER_ARM 1
as support for the ARM filter without confusing it with these:
#define HAVE_DECODER_ARM64
#define HAVE_DECODER_ARM64 1
Previously only the ones ending with " 1" were accepted for
the macros where this kind of confusion was possible.
This should help with Meson support because Meson's built-in
features produce config.h entries that are either
#define FOO 1
#define FOO 0
or:
#define FOO
#undef FOO
The former method has a benefit that one can use "#if FOO" and -Wundef
will catch if a #define is missing (for example, it helps catching
typos). But XZ Utils has to use the latter since it has been
convenient with Autoconf's default behavior.[*] While it's easy to
emulate the Autoconf style (#define FOO 1 vs. no #define at all)
in Meson, it results in clumsy code. Thus it's better to change
the few places in the tests where this difference matters.
[*] While most checks in Autoconf default to the second style above,
a few things use the first style (like AC_CHECK_DECLS). The mix
of both styles is the most confusing as one has to remember which
macro needs #ifdef and which #if. Currently HAVE_VISIBILITY is
only such config.h entry that is 1 or 0. It comes unmodified
from Gnulib's visibility.m4.
Add a new optional argument to specify the directory of the xz and
xzdec executables.
If ../config.h doesn't exist, assume that all encoders and decoders
are available.
Add a new optional second argument: directory of the xz and xzdec
executables. This is need with the CMake build where the binaries
end up in the top-level build directory.
If ../config.h doesn't exist, assume that all encoders and decoders
are available. This will make this script usable from CMake in the
most common build configuration.
NOTE: Since the existence of ../config.h is checked, the working
directory of the test script must be a subdir in the build tree!
Otherwise ../config.h would look outside the build tree.
Use the default check type instead of forcing CRC32 or CRC64.
Now the script doesn't need to check if CRC64 is available.
This is a bit hacky since the scripts grep config.h to know which
features were built but the CMake build doesn't create config.h.
So instead those test scripts will be run only when all relevant
features have been enabled.
The code makes aligned 16-byte reads which may read up to 15 bytes
before the beginning or past the end of the buffer if the buffer
is misaligned. The unneeded bytes are then ignored. It cannot cross
page boundaries and thus cannot cause access violations.
This inherently trips address sanitizer which was already disabled
with __attribute__((__no_sanitize_address__)). However, it also
trips memory sanitizer if the extra bytes are uninitialized because
memory sanitizer doesn't see that those bytes then get ignored by
byte shuffling in the xmm registers.
The plan is to change the code so that all sanitizers pass but it's
not finished yet (performance shouldn't get worse) so as a temporary
measure to keep OSS Fuzz happy, the CLMUL CRC is now disabled even
though I think think the code is fine to use (and easy enough to review
the memory accesses in it too).
This is closer to what it was before the --filtersX support was added,
just extended to support for scaling all filter chains. The method
before this commit was an extended version of the original too but
it was done in a more complex way for no clear reason. In case of
an error, the complex version printed fewer informative messages
(a good thing) but it's not a sigificant benefit.
In the limit is too low even for single-threaded mode, the required
amount of memory is now reported like in 5.4.x instead of like in
5.5.1alpha - 5.6.1 which showed the original non-scaled usage. It
had been a FIXME in the old code but it's not clear what message
makes the most sense.
Fixes: 5f0c5a0438
The convention is that
lzma_filter filters[LZMA_FILTERS_MAX + 1];
contains the filters of a single filter chain.
It was so here as well before the commit
d6af7f3470.
It changes "filters" to a ten-element array of filter chains.
It's clearer to call this array-of-arrays "chains".
This also renames "filter_idx" to "chain_idx" which is used
as an index as in chains[chain_idx].
opt_mode == MODE_COMPRESS isn't possible when HAVE_ENCODERS isn't
defined. Thus, when *encoding*, the message about *decoder* memory
usage is possible to show only when both encoder and decoder have
been built.
Since the message is shown only at V_DEBUG, skip the memusage
calculation if verbosity level isn't high enough.
Fixes: 5f0c5a0438
Since 5.3.5beta, "xz --lzma2=mf=bt4,nice=2" works even though bt4 needs
at least nice=4. It is rounded up internally by liblzma when needed.
Fixes: 5cd9f0df78
Lasse Collin