[Lldb-commits] [lldb] [lldb] Don't allocate an implausible PT_NOTE in FindBuidIdInCoreMemory (PR #236999)

[Jacob Lalonde via lldb-commits]

Jlalond wrote:

I'm not sure if this fixes the real problem. A full core should be equal to or greater than the sum of all the file_sz. (A VMA with no faulted pages can have a mem_sz > file_sz and the debugger should handle that as a range of all zeros correctly).

However when they're truncated this does not apply, and I think what we need to do is when loading in different segments, only load the segments up until we exceed the file-length. Which is what I think is actually happening here, I'd be very surprised is the program headers were actually garbage, but instead that the core was truncated.

https://github.com/llvm/llvm-project/pull/236999