[PATCH] [RFC] Rewrite the way we generate debug locations for variables.

[Adrian Prantl]

Hi Alexey,

thanks for tackling this problem!

as for quick gut reactions:
Having some doxygen-style comments for each of the member functions of DbgValueHistoryCalculator would make reviewing this patch a lot easier.

As for testing, in addition to the gdb test suite, there is also the lldb test suite, and if you’re looking for a place to put in more end-to-end debug-info tests, there is the llvm debuginfo-tests repository which runs on both lldb and gdb buildbots. (There are only a handful of tests in there right now, but it doesn’t have to stay that way.) On the llvm side this will need at least a test exercising llc | llvm-dwarfdump.

This is a really naive question, but it will help me understand your decisions better: would it be possible to change codegen to insert a DBG_VALUE into every MC block that post-dominates the basic block that contained the original dbg.declare/value intrinsic?

cheers,
adrian



On Jan 30, 2014, at 8:21, Alexey Samsonov <samsonov at google.com> wrote:

>  clang-formatize the new code
> 
> Hi dblaikie,
> 
> http://llvm-reviews.chandlerc.com/D2658
> 
> CHANGE SINCE LAST DIFF
>  http://llvm-reviews.chandlerc.com/D2658?vs=6773&id=6774#toc
> 
> Files:
>  lib/CodeGen/AsmPrinter/DwarfDebug.cpp
> <D2658.2.patch>