[llvm] r296253 - [DebugInfo] Skip implicit_const attributes when dumping .debug_info. NFC.

Robinson, Paul via llvm-commits llvm-commits at lists.llvm.org
Tue Feb 28 15:37:02 PST 2017



> -----Original Message-----
> From: llvm-commits [mailto:llvm-commits-bounces at lists.llvm.org] On Behalf
> Of Adrian Prantl via llvm-commits
> Sent: Monday, February 27, 2017 11:47 AM
> To: David Blaikie
> Cc: llvm-commits at lists.llvm.org
> Subject: Re: [llvm] r296253 - [DebugInfo] Skip implicit_const attributes
> when dumping .debug_info. NFC.
> 
> 
> > On Feb 27, 2017, at 11:13 AM, David Blaikie <dblaikie at gmail.com> wrote:
> >
> > I tend to think it's more legible if we include the values there, even
> if no bytes were required to produce them. I don't generally look at the
> abbreviations when reading a dump & expect to see the data in the
> debug_info/debug_types output.
> 
> I'm pretty sure we want the implicit attributes to show up in the dump
> (perhaps with a comment that they are implicit in the abbreviation where
> we normally show [DW_FORM...] in the output.
> 
> Tools like llvm-dwarfdump are used by different audiences. Eventually we
> probably want something like a --brief mode (that could even be the
> default like in Darwin's dwarfdump) that just shows the high-level
> contents of the debug_info section without all the low-level details like
> forms and abbreviations, and then we could have a verbose mode that shows
> the actual encoding of the data. Personally, most of the times I'm using
> dwarfdump I typically don't care about how an attribute was encoded.

Shameless plug: There will be a lightning talk about our DIVA tool at
EuroLLVM, which does exactly this kind of abstraction, presenting a
"logical" view of the info.  You can even use it to compare different
versions of DWARF output for the same source.
--paulr

> 
> 
> Side note: I don't think that this commit was really NFC.
> 
> -- adrian
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits


More information about the llvm-commits mailing list