[prev in list] [next in list] [prev in thread] [next in thread] 

List:       gcc
Subject:    Re: IA32 bootstrap error
From:       Daniel Jacobowitz <drow () mvista ! com>
Date:       2001-12-31 3:30:55
[Download RAW message or body]

On Sun, Dec 30, 2001 at 06:41:25PM -0500, Richard Kenner wrote:
>     > Do you really mean to require using an unreleased GDB in order to debug
>     > programs produced by GCC?
> 
>     Yes, in order to debug an unreleased version of gcc.
>     What's the problem?
> 
> For one thing, it means that people who want to test the new version
> of GCC will also have to test that new version of GDB, which may not
> be something they want to do.

That's why the patch was moved to a stable release branch, from which
AFAIK we are still spinning snapshots, a month ago.

> For another, using the cvs head of a package can require knowlege of
> that package's internals to fix any bugs you run into.  GCC knowlege
> is quite specialized, as is GDB knowlege.  The number of people who
> know both is *very* small and the number who know both *well* is
> probably very close to zero.
>  
>     It is expected that gdb will release a version that supports the
>     features once gcc is released (IE soon after).
> 
> And what if they don't?  Is the GCC release schedule to be held hostage
> to the GDB release schedule or are we going to tell people "oh, by the way,
> if you you want to debug programs produced by GCC 3.1, you'll have to
> pick up this untested GDB"?

It should be irrelevant.  We plan for a GDB 5.1.1 release to be rolled
mid-January, or possibly early February - i.e. months before GCC 3.1. 
I hope to have C++ debugging back to a reasonable point by then,
barring whatever additional DWARF2 work I am actively doing on the
main GDB trunk.

-- 
Daniel Jacobowitz                           Carnegie Mellon University
MontaVista Software                         Debian GNU/Linux Developer
[prev in list] [next in list] [prev in thread] [next in thread]