http://bugs.winehq.org/show_bug.cgi?id=22384
Ken Thomases ken@codeweavers.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |ken@codeweavers.com
--- Comment #2 from Ken Thomases ken@codeweavers.com 2010-04-19 17:10:23 --- (In reply to comment #1)
I don't have much to add. I added support for stabs because that's what is useful to me and CodeWeavers. There's no argument against DWARF support implicit in that decision.
I know next to nothing about DWARF. From what I can tell, Charles, your summary seems about correct, and suggests that adding support for DWARF will be much more complicated than stabs was, or than DWARF is on Linux.
Apparently, GDB will find the .dSYM bundle via a search heuristic. It will also find DWARF info in the original .o files, provided that they are in their original location. We may want DbgHelp to do the same. Unfortunately, GDB makes use of an Apple-private framework to do the searching.
A bit of experimentation shows that the paths to the .o files are referenced via stabs-style entries in the executable (assuming they aren't stripped). So, it may be possible to use the current stabs support to find the .o files, then import the DWARF info from them. Unfortunately, that requires some address mapping.
I found this article from an Apple engineer who helped design and implement their DWARF support: http://wiki.dwarfstd.org/index.php?title=Apple%27s_%22Lazy%22_DWARF_Scheme