> On Nov 17, 2016, at 1:45 PM, Sam McCall <sammcc...@google.com> wrote: > > On Thu, Nov 17, 2016 at 10:29 PM, Jim Ingham <jing...@apple.com > <mailto:jing...@apple.com>> wrote: > Ah, good catch. That is not right. If there’s inlined function information, > you want to make sure you haven’t gone from one inlined function to another. > But the symbol is always going to be the same in this case. > OK, I've updated the patch to return false in this case. > > Actually, the old code is a overly restrictive by comparing the blocks > directly. What you really want to test is that if either side is in an > inlined function, both sides are still inside the same inlined function > instantiation. We should really compare the ContainingInlinedBlock of the > current blocks if there is inlined info. > Happy to take a stab at this - I'm not too familiar with the structures here > so I'll read a bit and send a new patch if that's OK!
That’s fine. Jim > > This stuff is a PITA to write tests for, too, since it is hard to predict how > the compiler will emit inlined code... > > Jim > > >> On Nov 17, 2016, at 1:13 PM, Sam McCall <sammcc...@google.com >> <mailto:sammcc...@google.com>> wrote: >> >> Thanks. I just noticed, this also changes the behavior (now returns true) if: >> - comp_unit and function both match >> - block->GetInlinedFunctionInfo != nullptr >> - blocks aren't equal >> - symbols are equal >> >> Is this correct? Otherwise I'll fix this case to match the old behavior. >> >> On Thu, Nov 17, 2016 at 10:09 PM, Jim Ingham <jing...@apple.com >> <mailto:jing...@apple.com>> wrote: >> jingham accepted this revision. >> jingham added a comment. >> This revision is now accepted and ready to land. >> >> That way of doing it is fine too. >> >> >> https://reviews.llvm.org/D26804 <https://reviews.llvm.org/D26804> >> >> >> >> > >
_______________________________________________ lldb-commits mailing list lldb-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
