[Lldb-commits] [PATCH] D153840: [LLDB] Fix buffer overflow problem in DWARFExpression::Evaluate.

[Caroline Tice via Phabricator via lldb-commits]

cmtice marked an inline comment as done.
cmtice added a comment.

Hi Jason,

I had been talking more with David, and yes, I had come to the conclusion that you are both right and that this was not the right fix.  I am planning on reverting this, but I am trying to figure out the right fix to replace it with.  I can't share the source that was causing the bug to manifest, because it's in proprietary code, but David is looking at it and I believe he has come to the conclusion that there is a bug in the DWARF code generation -- we were getting a size of 16, which is absolutely not right.  The question is, in the case of bad DWARF being generated, what (if anything) should the LLDB code here be doing? Should we check the size as soon as we read it in, and assert that  it must be <= 8?  Or something else?  Or just leave the LLDB code entirely alone?

What do you (and other reviewers) think is the right thing to do here?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D153840/new/

https://reviews.llvm.org/D153840