* New testing upstream release
* Relative call in the chroot without proc failed.
See: fix-an-issue-in-chroot-witout-proc.diff
* Bring back lldb-link-atomic.diff to make sure lldb builds under
powerpc
* Also limit the number of archs for liblldb-dev
* Fix the bad declaration on the lldb desactivation
* Also disable lldb under powerpc
* Hopefully, fix lldb under Kfreebsd-* (thanks to Ed Maste if it works)
* Add the Ocaml ABI dependency (Closes: #731344)
* Disable LLDB also for ia64, mips & mipsel
* New testing upstream release
* 0047-version-name.diff ocamldoc.diff removed (applied upstream)
* r600 is now compiled by default (remove the configure arg)
* Remove the usage of --with-c-include-dirs, --with-cxx-include-root,
--with-cxx-include-arch and --with-cxx-include-64bit-dir
It was blocking the automatic detection of the path of clang.
In particular in the context of the usage of -target.
However, it does not completely fix the detection of the i386 C++ path.
See the next item.
(Closes: #729933)
* Bring back the path to libstdc++ under i386. Still not fixed upstream
(Closes: #730857)
* Define also MAXPATHLEN in Path.inc for HURD.
* Silent the trillion of warnings in the LLDB Python wrapper (swig generated)
See silent-swig-warning.diff
* Silent some i386 tests failing (it is expected)
See silent-MCJIIT-tests.diff
* Make lldb 3.4 also conflict with 3.5 (Closes: #730163)
* Make python-clang 3.4 also conflict with 3.5 (Closes: #730164)
* Port LLVM to mips64el. Thanks to YunQiang Su. Initially done for
3.3 and ported on the 3.4 (Closes: #730808)
* If we get an unexpected pass, do not break the tests
do-not-fail-on-unexpected-pass.diff (I am disabling some tests)
* Fix the path detection of the objective h headers.
* Also add usr/lib/llvm-3.4/lib/clang/3.4/include =>
usr/lib/clang/3.4/include symlink to simplify the path detection
* Force the build to gcc 4.8... gcc 4.6 used on some Debian archs does not
support some C++ features.
* Fail the build when llvm tests are failing under amd64 + i386. More to come.
* Fix a libclang.so.1 issue during the clang tests
* Improve the patch 23-strlcpy_strlcat_warning_removed.diff
(also remove the tests)
* Make lldb 3.4 also conflict with 3.5 (Closes: #730163)
* Make python-clang 3.4 also conflict with 3.5 (Closes: #730164)
* Remove usr/lib/llvm-3.4/build/autoconf/LICENSE.TXT
* silent warning "manpage-has-useless-whatis-entry" in lldb-3.4
* silent warning "package-name-doesnt-match-sonames libclang1"
* Refresh patch kfreebsd_v2.diff to, maybe, fix lldb build under kfreebsd.
Thanks to Ed Maste for the patch.
* New testing upstream release
* kfreebsd.diff removed. Applied upstream
* Remove patch 0046-Revert-Patch-to-set-is_stmt-a-little-better-for-prol.patch
Useless now and missleading
* Branch from llvm-toolchain-snapshot
* Standards-Version updated to 3.9.5
--with-cxx-include-arch and --with-cxx-include-64bit-dir
It was blocking the automatic detection of the path of clang.
In particular in the context of the usage of -target.
However, it does not completely fix the detection of the i386 C++ path.
See the next item.
(Closes: #729933)
* Bring back the path to libstdc++ under i386. Still not fixed upstream
(Closes: #730857)
* silent warning "manpage-has-useless-whatis-entry" in lldb-3.4
* silent warning "package-name-doesnt-match-sonames libclang1"
* Refresh patch kfreebsd_v2.diff to, maybe, fix lldb build under kfreebsd.
Thanks to Ed Maste for the patch.
