The SDK for macOS 10.13 adds `futimens` and `utimensat` so our checks
for these symbols may pass. However, the symbols are not available at
runtime on older macOS versions. Instead on macOS we can check for
availability based on the deployment target version.
Issue: #17101
Backport upstream libarchive commit 70f497f456 (As per Cryptographic
Requirements, 2017-03-19). Discard more bytes of the RC4 keystream
to reduce the possibility of non-random bytes.
This was added upstream but does not make sense in CMake because we
may replace the library with a logical target name that will not
work inside a `try_compile`.
Upstream libarchive now defines this macro in two places with the same
logic in both. However, CMake's bundled copy disables this macro, so
we need to update the new location of its definition to be consistent.
OpenSSL 1.1 made some CTX structures opaque. Port our code to use the
structures only through pointers via OpenSSL 1.1 APIs. Use our adaption
layer to make this work with OpenSSL 1.0 and below.
Patch-by: Tomas Mraz <tmraz@redhat.com>
Patch-from: https://bugzilla.redhat.com/1383744
Add private forwarding headers for `openssl/{evp,hmac}.h` to give us a
central place to add adaptation code to work across multiple
incompatible OpenSSL versions. Provide compatibility implementations of
some OpenSSL 1.1 APIs when using older OpenSSL versions.
In commit v3.6.0-rc1~100^2 (libarchive: Drop early use of crypto
library, 2016-05-03) we accidentally dropped the include directory for
OpenSSL as well as the library. Without that, it works only
accidentally when CHECK_CRYPTO happens to choose the OpenSSL
implementation and add the include directory.
Reported-by: Bradley Lowekamp <blowekamp@mail.nih.gov>
We already add the crypto library after the ssl library when using
OpenSSL. Do not add it to the list of libraries individually or we may
end up with the wrong order later. Further investigation will be needed
to see how this change can be reconciled with the upstream need for the
original line.
We will now manage the libarchive source tree updates using
the `Utilities/Scripts/update-libarchive.bash` script. Drop
the README that covered the old method.
Port upstream LibArchive commit "compute string pointers after
concatenation" (2014-09-25) and commit "Move variables to top of
function for non-C99 compilers" (2014-11-15) to our CMake copy.
Otherwise we may compute a pointer to memory that is about to be freed
and then compute a bad size to give to CryptGenRandom.
Inspired-by: Tim Kientzle <kientzle@gmail.com>
When testing CMAKE_<LANG>_COMPILER_ID values, do not explicitly
dereference or quote the variable. We want if() to auto-dereference the
variable and not its value. Also replace MATCHES with STREQUAL where
equivalent.
Newer HP-UX versions (e.g. 11.31) also offers F_SETTIMES, but with a different
struct for the timing information.
Suggested-by: Eric Berge <ericmberge@gmail.com>
67f5f0a9 libarchive: Use _snprintf on Windows, not snprintf
fcfbb0a9 libarchive: Drop LIBARCHIVE_ADDITIONAL_LIBS, CMake does not need it
61a649d9 libarchive: Update README-CMake.txt for new snapshot
4533560c Merge branch 'libarchive-upstream' into update-libarchive
44d6b82f libarchive: Disable all whitespace checks in third-party code
37f225b7 libarchive 3.1.2-246-ga5a5d28b (reduced)
6ab7c326 libarchive: Avoid left-shift overflow of signed integer
In libarchive/archive_write_set_format_zip.c there are two calls to
archive_le32enc whose second argument is of the form
archive_entry_mode(zip->entry) << 16
However, the return type from archive_entry_mode may be a signed integer
so the shift may overflow. Since the second argument of archive_le32enc
expects uint32_t anyway, simply cast to that prior to shifting.
In archive_entry.h refer to "BY_HANDLE_FILE_INFORMATION *" using
"struct _BY_HANDLE_FILE_INFORMATION *" to avoid requiring the
includer to already have <windows.h>.
Drop ENABLE_TAR and ENABLE_CPIO related options because we do not build
these command-line tools. Drop ENABLE_TEST and ENABLE_COVERAGE options
because we do not build the tests.
Resolve conflicts in favor of the upstream side where possible.
Resolve a logical conflict in archive_windows.h where the upstream
port to Watcom was done slightly differently from ours.
Port upstream commit 6cf33c93 (Issue 320: Rewrite (again) to avoid
the left shift that CLang dislikes so much, 2013-12-07) into CMake.
Inspired-by: Tim Kientzle <kientzle@freebsd.org>
Port upstream commit 533e8fda (Rework the sign-extension to avoid
left-shift of an explicit negative number, 2013-06-29) into CMake.
Inspired-by: Tim Kientzle <kientzle@freebsd.org>
Since commit 7d47c693 (Drop compatibility with CMake < 2.4, 2013-10-08)
we no longer need to use the configure_file IMMEDIATE option to support
compatibility modes less than 2.0.
Apple distributes their own Clang build with their own version numbers
that differ from upstream Clang. Use the __apple_build_version__ symbol
to identify the Apple Clang compiler and report the Apple Build Version
as the fourth version component in CMAKE_<LANG>_COMPILER_VERSION. Add
Compiler/AppleClang-<lang> and Platform/Darwin-AppleClang-<lang> modules
that simply include the upstream equivalents.
Fix comparisons of CMAKE_<LANG>_COMPILER_ID to Clang in CMake's own
source and tests to account for AppleClang.
Ensure CMAKE_DATA_DIR, CMAKE_DOC_DIR, and CMAKE_MAN_DIR are always
relative paths in CMake code, and set defaults accordingly. Use the
install() command instead of install_files() and install_targets().
This is more modern and also avoids stripping of the first character
from user-specified destinations.
While at it, fix the default destinations reported in the bootstrap
help.
Avoid requiring CMake 2.8.6 for CMakePushCheckState or CMake 2.8.8 for
CMakeExpandImportedTargets. Drop the custom versions of CMake modules
CheckCSource(Compiles|Runs) because we do not use the SAFESEH option
anyway.
Follow up change from commit ffa6faa4 (libarchive: Include cm_zlib.h to
get zlib used by CMake, 2011-12-20) for new includes of zlib.h in
updated libarchive.
The archive_string_conv type sc variable already freed via free(sc) on
the other hand in second line we are tyring to free its subset via
free(sc->from_charset) this will cause a problem because we couldn't
reach sc after first release.
Reviewed-by: Igor Murzov <e-mail@date.by>
caught by clang's -fsanitize=shift. A small unsigned int was
promoted, according to C's regular promotion rules, to a signed
int, it was then left shifted. This sometimes pushed a 1 into
the sign bit, which is undefined behaviour. Fixed by using
unsigned temporaries.
Update the requirement specified in the top-level CMakeLists.txt file.
Drop the special-case minimum required version of 2.8.0 because the new
minimum subsumes it.
Revert commit 6c611c6b (libarchive: Restore CMake 2.6.3 as minimum
version, 2012-01-05) since our requirement now subsumes libarchive's.
The CMake TarTest fails with the error
mbsnrtowcs.c:116: __mbsnrtowcs: Assertion
`status == GCONV_OK || status != GCONV_EMPTY_INPUT ||
status == GCONV_ILLEGAL_INPUT || status == GCONV_INCOMPLETE_INPUT ||
status == GCONV_FULL_OUTPUT' failed.
on very old glibc versions. Work around the problem by pretending that
mbsnrtowcs does not exist. Libarchive will fall back to mbrtowc.
Use the approach originally used in commit f91b3c1d (Add options to
build with system utility libraries, 2006-10-19) for all other
third-party libraries. Create a "cm_bzlib.h" header wrapper that
robustly includes the header from the bzip2 library chosen for the CMake
build (either builtin or system version). Include the header wrapper
anywhere we need the API provided by <bzlib.h>.
Upstream libarchive now requires CMake 2.8 to get the newer add_test
functionality. Since we do not build libarchive's tests we do not
need the requirement.
We cannot suppress PGI compiler warnings completely because even with
the "-w" flag the compiler still writes a message containing "compilation
completed with warnings" to stderr.
A warning is triggered by expressions like
test ? NULL : ptr_to_const_char
test ? ".." : ptr_to_const_char
that the PGI compiler handles incorrectly. It chooses the pointer type
of the first option (either void* or char*) and warns about conversion
of the second without a cast. Flip the expression logic to
!test ? ptr_to_const_char : NULL
!test ? ptr_to_const_char : ".."
to help the compiler choose the proper result type.
