Re: Meeting Agenda 2008-08-06

[Mark Wielaard <mark at klomp dot org>]

On Thu, 2008-08-07 at 17:42 +0200, Mark Wielaard wrote:
> On Thu, 2008-08-07 at 11:06 -0400, Andrew Cagney wrote:
> > Tom Tromey wrote:
> > > * Andrew said that our existing dwarf-reading tools are insufficient.
> > >   Andrew, I forgot to ask at the meeting -- what specifically do you
> > >   need that we can't readily do now?
> > >   
> > readelf, and clones, are too raw.  It is very hard to meaningfully map 
> > the information back to the source.
> 
> There are the dwarves tools:
>         
>         dwarves is a set of tools Iâve been writing that use the DWARF
>         debugging information inserted in ELF binaries by compiler such
>         as GCC , used by well known debuggers such as GDB, and more
>         recent ones such as systemtap. Utilities in the dwarves suite
>         include pahole, that can be used to find alignment holes in
>         structs and classes in languages such as C, C++, but not limited
>         to these, and other information such as CPU cacheline alignment,
>         helping pack those structures to achieve more cache hits,
>         codiff, a diff like tool to compare the effects changes in
>         source code generate on the resulting binaries, pfunct, that can
>         be used to find all sorts of information about functions,
>         inlines, decisions made by the compiler about inlining, etc, to
>         be continued.

That quote was from Arnaldo Carvalho de Melo btw.
He writes more about dwarves at http://oops.ghostprotocols.net:81/blog
Dwarves itself doesn't seem to have a homepage at the moment.

> http://gnu.wildebeest.org/diary/2007/07/10/the-7-dwarves/
> https://ols2006.108.redhat.com/2007/Reprints/melo-Reprint.pdf

Cheers,

Mark