[PATCH] D11324: [PM/AA] Replace the only use of the AliasAnalysis::deleteValue API (in GlobalsModRef) with CallbackVHs that trigger the same behavior.
jbrecka at apple.com
Wed Jul 22 03:51:55 PDT 2015
Same happens on the OS X buildbot: http://lab.llvm.org:8080/green/job/clang-stage1-cmake-RA-incremental_build/12400/console <http://lab.llvm.org:8080/green/job/clang-stage1-cmake-RA-incremental_build/12400/console>
> On 22 Jul 2015, at 12:39, Kim Gräsman <kim.grasman at gmail.com> wrote:
> kimgr added inline comments.
> Comment at: llvm/trunk/lib/Analysis/IPA/GlobalsModRef.cpp:106-114
> @@ -105,2 +105,11 @@
> + /// Handle to clear this analysis on deletion of values.
> + struct DeletionCallbackHandle;
> + /// List of callbacks for globals being tracked by this analysis. Note that
> + /// these objects are quite large, but we only anticipate having one per
> + /// global tracked by this analysis. There are numerous optimizations we
> + /// could perform to the memory utilization here if this becomes a problem.
> + std::list<DeletionCallbackHandle> Handles;
> This doesn't compile on my FreeBSD 10.1 rig;
> In file included from /home/kimgr/code/llvm-trunk/llvm/lib/Analysis/IPA/GlobalsModRef.cpp:32:
> /usr/include/c++/v1/list:227:9: error: field has incomplete type '<anonymous>::GlobalsModRef::DeletionCallbackHandle'
> _Tp __value_;
> /usr/include/c++/v1/type_traits:1994:27: note: in instantiation of template class 'std::__1::__list_node<<anonymous>::GlobalsModRef::DeletionCallbackHandle, void *>' requested here
> decltype(__is_constructible_test(declval<_Tp>(), declval<_Args>()...))
> $ clang --version
> FreeBSD clang version 3.4.1 (tags/RELEASE_34/dot1-final 208032) 20140512
> Target: x86_64-unknown-freebsd10.1
> Thread model: posix
> Not sure which libc++ version came with this system. I've confirmed that moving the full `DeletionCallbackHandle` definition up in place of the forward decl resolves the problem.
> The Clang version is getting old, but this is FreeBSD's latest production release, so it would be nice if it could be supported.
> rL LLVM
> llvm-commits mailing list
> llvm-commits at cs.uiuc.edu
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the llvm-commits