Previous discussions about recording accurate code ranges in psymtabs?

Previous discussions about recording accurate code ranges in psymtabs?

[Joel Brobecker]

I was looking at the current code and trying to figure out how to
modify GDB to record accurate code ranges in its psymtabs instead
of the textlow/texthigh pair. It looks like it's going to be a major
change, so I'm trying to find a path that would allow for a transition
rather than a gigantic patch.

Daniel mentioned that this has been discussed before. Any pointer to
these discussions? I found some discussions, but they seemed to make
a decision to move the opposite way: "Remove live range support"...

-- 
Joel

Re: Previous discussions about recording accurate code ranges in psymtabs?

[Elena Zannoni]

Joel Brobecker writes:
 > I was looking at the current code and trying to figure out how to
 > modify GDB to record accurate code ranges in its psymtabs instead
 > of the textlow/texthigh pair. It looks like it's going to be a major
 > change, so I'm trying to find a path that would allow for a transition
 > rather than a gigantic patch.
 > 
 > Daniel mentioned that this has been discussed before. Any pointer to
 > these discussions? I found some discussions, but they seemed to make
 > a decision to move the opposite way: "Remove live range support"...
 > 
 > -- 
 > Joel

Yes, there were patches from Peter Sorfa (or Sofra) from Caldera.  Jim
started also a branch a couple of years back.  I don't have references
handy, sorry.  And yes, i need to reply to a lot of mail (I was really
busy these past 2-3 days), will be back on line shortly.

elena

Re: Previous discussions about recording accurate code ranges in psymtabs?

[Daniel Jacobowitz]

On Wed, Nov 19, 2003 at 10:39:04AM -0800, Joel Brobecker wrote:
> I was looking at the current code and trying to figure out how to
> modify GDB to record accurate code ranges in its psymtabs instead
> of the textlow/texthigh pair. It looks like it's going to be a major
> change, so I'm trying to find a path that would allow for a transition
> rather than a gigantic patch.
> 
> Daniel mentioned that this has been discussed before. Any pointer to
> these discussions? I found some discussions, but they seemed to make
> a decision to move the opposite way: "Remove live range support"...

Jim Blandy posted a proposed interface for describing the ranges, but I
can't find the subject now.  And live ranges are something completely
different :)  More like .debug_loc location lists, really.

-- 
Daniel Jacobowitz
MontaVista Software                         Debian GNU/Linux Developer