[RFC] fullname attribute for GDB/MI stack frames

[Daniel Jacobowitz]

On Wed, May 04, 2005 at 07:05:40AM +0300, Eli Zaretskii wrote:
> > Date: Tue, 3 May 2005 18:13:40 -0400
> > From: Christopher Faylor <me@cgf.cx>
> > 
> > Wouldn't it make more sense to fix the fullpath machinery?
> 
> The fullpath machinery is fine, it doesn't need any fixing.
> 
> Please, let's not break something that works just because some test
> case doesn't seem 100% correct.  Let's keep things in the right
> perspective here.  The fullpath machinery was invented to let GDB
> process file names for its internal purposes, and it does its job
> well, because for those purposes, treating d:foo as absolute file
> names is _exactly_ right.

Huh?  That's not true.  xfullpath was created to avoid canonicalizing
certain file symlinks, not to change anything having to do with
absolute paths or directory links.  Are you talking about some
different function?

-- 
Daniel Jacobowitz
CodeSourcery, LLC