These ports are failing on build cluster for some time. I tried reproducing
it locally and it turned out to be a strange heisenbug - the compiler crashes
when compiling different unrelated files and restarting the build makes the
problem go away.
I wasn't able to extract a minimal reproducing testcase, so didn't contact
upstream.
Mark BROKEN for now and hope for a new compiler version to work.
Reported by: pkg-fallout
- Update all the consumers to use USES=tex
- USE_TEX=yes is the old way of writing USES=tex which has been removed
and replaced in all ports
- Almost all of the USE_TEX features remains unchanged
- Some consumers had the same variables defined both in the mk
infrastructure and also in the ports which have been removed from the
ports as those are redundant.
In case any of the consumers are failing to build please make sure that
the nexessary USES=tex is there. Unlike previous USE_TEX=yes will no
longer load the required VARS for tex and related dependencies.
Reviewed by: portmgr
Approved by: portmgr (blanket)
During an exp-run for llvm 15 (see bug 265425), it turned out that
lang/gjs failed to build with clang and libc++ 15:
In file included from ../test/gjs-tests.cpp:11:
In file included from /usr/include/c++/v1/random:1689:
In file included from /usr/include/c++/v1/__random/discrete_distribution.h:18:
In file included from /usr/include/c++/v1/numeric:167:
In file included from /usr/include/c++/v1/functional:515:
In file included from /usr/include/c++/v1/__functional/boyer_moore_searcher.h:24:
In file included from /usr/include/c++/v1/array:127:
In file included from /usr/include/c++/v1/algorithm:1851:
In file included from /usr/include/c++/v1/__algorithm/ranges_sample.h:13:
In file included from /usr/include/c++/v1/__algorithm/sample.h:18:
/usr/include/c++/v1/__random/uniform_int_distribution.h:162:5: error: static assertion failed due to requirement '__libcpp_random_is_valid_inttype<char32_t>::value': IntType must be a supported integer type
static_assert(__libcpp_random_is_valid_inttype<_IntType>::value, "IntType must be a supported integer type");
^ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../test/gjs-tests.cpp:71:16: note: in instantiation of template class 'std::uniform_int_distribution<char32_t>' requested here
return std::uniform_int_distribution<T>(lowest_value)(gen);
^
../test/gjs-tests.cpp:942:30: note: in instantiation of function template specialization 'Gjs::Test::get_random_number<char32_t>' requested here
char32_t random_char32 = get_random_number<char32_t>();
^
This is because std::uniform_int_distribution is only defined for strict
integer types, such as short, int, etc. Use
std::uniform_int_distribution<uint32_t> instead, and cast the result
back to a char32_t.
PR: 268329
Approved by: portmgr (tcberner)
MFH: 2022Q4
During an exp-run for llvm 15 (see bug 265425), it turned out that
lang/mono failed to build with clang 15, on i386:
mini-runtime.c:806:24: error: incompatible pointer to integer conversion assigning to 'guint32' (aka 'unsigned int') from 'MonoLMF *' (aka 'struct MonoLMF *') [-Wint-conversion]
ext->lmf.previous_lmf = *lmf_addr;
^ ~~~~~~~~~
mini-runtime.c:808:24: error: incompatible pointer to integer conversion assigning to 'guint32' (aka 'unsigned int') from 'gpointer' (aka 'void *') [-Wint-conversion]
ext->lmf.previous_lmf = (gpointer)(((gssize)ext->lmf.previous_lmf) | 2);
^ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
As noted, this is because on i386 the previous_lmf field is declared as
guint32, while other architectures use gpointer.
Upstream changed the field to gpointer in
<https://github.com/mono/mono/pull/8308/commits/1c43a8476b5>, which got
merged into main in <https://github.com/mono/mono/commit/384c4a4e2ea>.
PR: 268321
Approved by: portmgr (tcberner)
MFH: 2022Q4
* Python 3.7 will reach its End-of-Life on 2023-06-27. Deprecate it and
set the expiration date accordingly.
Approved by: kai (python, maintainer)
Differential Revision: https://reviews.freebsd.org/D37587
Since GNAT is now installed in a specific prefix, we can simplify by
removing the programs suffix: this will avoid the usage of binary
aliases, only setting the PATH is necessary.
OpenCL drivers in clover lack headers and unlikely to be usable
outside of ocl-icd which uses dlopen(3) to find the drivers via
${PREFIX}/etc/OpenCL/vendors/*.icd (or OCL_ICD_VENDORS).
Approved by: manu (on dev-commits-ports-main@ list)
Gnat runs along with a gcc compiled with Ada, but we don’t want neither
modify lang/gcc12 nor conflict with it. Therefore we have to install it
in a separate directory.
After https://cgit.freebsd.org/src/commit/?id=be39466a1035, there is no
need anymore to patch the perl5 Makefiles to add
"-Wl,--allow-multiple-definition" to the linker flags. They should now
link with lld 15 and later without any issues.
PR: 265516
Fixes: 8c69fc20e6
MFH: 2022Q4
This Ada compiler is built from Gcc-12, with help from Stephane Carrez
(Ada France).
The first time it uses lang/gcc6-aux as a bootstrap, but it can produce
a bootstrap for the next time, either for this version or for the next
ones.
Yet to be done:
- plug it into Mk/Uses/ada.mk
- build more bootstrapping assets
(ATM they exist only for amd64 on FreeBSD 13 and 14)
- add a test target
- try to cross-compile for more platforms?
KDE Gear 22.12 is Here!
Thursday, 8 December 2022
KDE Gear ⚙️ is back with exciting new features, performance boosts, and
bugfixes for all your favorite KDE apps!
In this release: Kate extends a warm welcome, Dolphin offers you more
choices, and a lot of apps serve up hamburgers galore!
Announcement: https://kde.org/announcements/gear/22.12.0/
When the both options are on, post-install-DOCS-on target is called
after post-install-CAPIDOCS-on has completed. And in the former files
under ${WRKSRC}/doc/ are copied to ${STAGEDIR}${RUBY_DOCDIR}. This
means files under ${WRKSRC}/doc/capi are copied to
${STAGEDIR}${RUBY_DOCDIR}/capi again. And of course
${WRKSRC}/doc/capi/.timestamp is included in them. As a result, even
though ${STAGEDIR}${RUBY_DOCDIR}/capi/.timestamp is removed in
post-install-CAPIDOCS-on target, it is detected as orphaned file at
check-plist phase. So fix it by adding 3rd argument to
${COPYTREE_SHARE} command so .timestamp is excluded from files to be
copied.
PR: 267775
Approved by: meta (ruby@)
Fixes: b49bdb1fe4 lang/ruby32: Simplify CAPIDOCS handling
When the both options are on, post-install-DOCS-on target is called
after post-install-CAPIDOCS-on has completed. And in the former files
under ${WRKSRC}/doc/ are copied to ${STAGEDIR}${RUBY_DOCDIR}. This
means files under ${WRKSRC}/doc/capi are copied to
${STAGEDIR}${RUBY_DOCDIR}/capi again. And of course
${WRKSRC}/doc/capi/.timestamp is included in them. As a result, even
though ${STAGEDIR}${RUBY_DOCDIR}/capi/.timestamp is removed in
post-install-CAPIDOCS-on target, it is detected as orphaned file at
check-plist phase. So fix it by adding 3rd argument to
${COPYTREE_SHARE} command so .timestamp is excluded from files to be
copied.
PR: 267775
Approved by: meta (ruby@)
Fixes: 6021649ff0 lang/ruby31: Simplify CAPIDOCS handling
When the both options are on, post-install-DOCS-on target is called
after post-install-CAPIDOCS-on has completed. And in the former files
under ${WRKSRC}/doc/ are copied to ${STAGEDIR}${RUBY_DOCDIR}. This
means files under ${WRKSRC}/doc/capi are copied to
${STAGEDIR}${RUBY_DOCDIR}/capi again. And of course
${WRKSRC}/doc/capi/.timestamp is included in them. As a result, even
though ${STAGEDIR}${RUBY_DOCDIR}/capi/.timestamp is removed in
post-install-CAPIDOCS-on target, it is detected as orphaned file at
check-plist phase. So fix it by adding 3rd argument to
${COPYTREE_SHARE} command so .timestamp is excluded from files to be
copied.
PR: 267775
Approved by: meta (ruby@)
Fixes: 93e3548c57 lang/ruby30: Simplify CAPIDOCS handling
When the both options are on, post-install-DOCS-on target is called
after post-install-CAPIDOCS-on has completed. And in the former files
under ${WRKSRC}/doc/ are copied to ${STAGEDIR}${RUBY_DOCDIR}. This
means files under ${WRKSRC}/doc/capi are copied to
${STAGEDIR}${RUBY_DOCDIR}/capi again. And of course
${WRKSRC}/doc/capi/.timestamp is included in them. As a result, even
though ${STAGEDIR}${RUBY_DOCDIR}/capi/.timestamp is removed in
post-install-CAPIDOCS-on target, it is detected as orphaned file at
check-plist phase. So fix it by adding 3rd argument to
${COPYTREE_SHARE} command so .timestamp is excluded from files to be
copied.
PR: 267775
Approved by: meta (ruby@)
Fixes: bdbe2a2f40 lang/ruby27: Simplify CAPIDOCS handling
Not every mesa ports provides shared libraries intented for dynamic uses
so stop running ldconfig script for those ports.
Sponsored by: Beckhoff Automation GmbH & Co. KG
