Okay, weird. So the DICompileUnit wouldn't even get written out to bitcode, IIUC. LGTM then.
> On 2016-Apr-06, at 14:31, Adrian Prantl <apra...@apple.com> wrote: > > aprantl added a comment. > > Duncan: There already exists a verifier check that ensures each DICompileUnit > is listed in llvm.dbg.cu. > This didn't break existing users of sample-based profiling, because the > Verifier cannot *find* the orphaned DICompileUnits. After the DISubprogram / > DICompileUnit pointer reversal, however, this will be possible and the > verification will fail. > > > http://reviews.llvm.org/D18808 > > > _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
