It hadn't crossed our radar yet, no. However, the opportunity to drop a 
dependency is always good. 

I'm currently involved in a substantial rewrite of our dwarf logic, so I would 
suggest waiting until I am done with that. I'll keep an eye on making our code 
as library-independent as possible. 

In the meantime, do you have any pointers to documentation on libdw? 

Drew

On Aug 20, 2012, at 6:45 PM, Josh Stone <[email protected]> wrote:

> Hi folks,
> 
> Has anyone in the project ever considered moving from libdwarf to
> elfutils' libdw?
> 
> My perspective is admittedly skewed, but in the land of RHEL+Fedora
> you're already using libelf from elfutils, so using libdw too seems like
> a good choice.  Plus, while libdwarf is available on Fedora, it's not in
> any RHEL 4-6, so removing an external dependency would be good.  That
> RHEL maintenance also impacts how well the libraries will support new
> DWARF versions and extensions in RHEL and Fedora, like the new dwz
> compression:
> 
> https://fedoraproject.org/wiki/Features/DwarfCompressor
> http://www.dwarfstd.org/ShowIssue.php?issue=120604.1
> 
> AFAIK elfutils is well supported on most Linux distros.  But I'm not
> sure about completely different OSes using DWARF (I think not), so your
> opinions on which ones matter to you and how they fare here would be
> appreciated.
> 
> The APIs of libdwarf and libdw are similar, but I think not quite
> possible as drop-in replacements.  So while a port would probably be
> straightforward, there would need to be some kind of shim if you want to
> maintain support for both libraries.
> 
> I'm interested to do this work if it is deemed welcome.  Please let me
> know what you think and how you'd like to proceed.
> 
> Thanks,
> Josh
> _______________________________________________
> Dyninst-api mailing list
> [email protected]
> https://lists.cs.wisc.edu/mailman/listinfo/dyninst-api

--
Andrew Bernat
Paradyn Project
[email protected]
http://www.cs.wisc.edu/~bernat




_______________________________________________
Dyninst-api mailing list
[email protected]
https://lists.cs.wisc.edu/mailman/listinfo/dyninst-api

Reply via email to