[Lldb-commits] Improvements to PECOFF support
virgile.bello at gmail.com
Mon Jun 17 09:40:54 PDT 2013
Some part of it, yes. Thanks so much for your work! Hope that is OK?
(sorry that I forgot to mention it, I was going to explain that in the much
bigger patch for MingW, partly based on the changes from the windows branch)
My hope is to first get the mingw changes merged upstream (changes are not
so huge, but are still a good base for the MSVC port), so that MSVC-related
changes can then be incrementally merged upstream as well.
So far, I could manage to use LLDB libraries to debug a windows exe
compiled with GCC, including stackframe and dwarf evaluations -- but still
a long way to go.
On Tue, Jun 18, 2013 at 1:32 AM, Carlo Kok <ck at remobjects.com> wrote:
> Op 17-6-2013 18:11, Virgile Bello schreef:
> 3 patches, aiming to improve PE/COFF support:
>> - First patch fix symbol reading (invalid header size from sizeof() ==
>> 20 != 18, and various bugfixes such as invalid skipping of auxiliary
>> symbols, 4 bytes shift from beginning, etc...).
>> - Second patch add image_base to section vmaddr offset so that VM addr
>> is in image_base space.
>> - Third patch add support for DWARF section in PECOFF (taken from ELF
>> counterpart), since they are generated by gcc/clang under windows.
>> This is all part of _much_ bigger patches to come, in order to bring
>> MinGW32 support (patch 2 was necessary to make it backtrace/evaluate
>> variables properly during the port, in case you're wondering).
>> Please let me know if anything seems wrong.
> Is this based on the Windows branch?
> lldb-commits mailing list
> lldb-commits at cs.uiuc.edu
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the lldb-commits