Re: Q: Estimated date for 2.16 release ?

[Daniel Jacobowitz <drow at false dot org>]

On Sun, Jan 30, 2005 at 09:10:52PM -0500, Hans-Peter Nilsson wrote:
> On Thu, 27 Jan 2005, Daniel Jacobowitz wrote:
> > On Thu, Jan 27, 2005 at 10:56:30AM -0500, Hans-Peter Nilsson wrote:
> > > On Thu, 27 Jan 2005, Ian Lance Taylor wrote:
> > > > Hans-Peter Nilsson <hp@bitrange.com> writes:
> > > > > BTW, was the intl/ directory about to be added to binutils?
> > > > > Maybe that and the fallout should be done before the release.
> > > > > Or maybe better done after branching?
> > > > The binutils releases already have an intl directory, don't they?
> > > > They used to, and it seems to me that they still do.
> > >
> > > Oh right you are, problem misdiagnosed.  Still a problem,
> > > though, and one that would be good to fix before the next
> > > release.
> >
> > It's more complicated than you think :-)
> 
> Uh, remind me: what did I think?  (I made no claim about
> simplicity above. ;-)
> 
> FWIW, one workaround is to remove the intl directory from the
> combined sources.  Perhaps better than upgrading binutils to
> autoconf 2.59 at this time (which would seem necessary in order
> to sync the intl machinery).

Anyway, it sounds like this is specifically related to combined builds,
and therefore not an issue for 2.16; if you try to combine released
toolchain components rather than HEAD, this is the least of your
problems :-)

> Anyway, I've seen a problem with the stabs linker-warning
> machinery (see libgloss/libnosys/warning.h) and a.out; there's a
> SEGV which valgrind says is an invalid access, after allocated
> memory.  Valgrind also changes the behavior and instead of the
> linker warnings, the symbols are undefined.  PR and testcase
> later.

I believe that you've fixed this since the above message.  Am I
remembering rightly?

-- 
Daniel Jacobowitz
CodeSourcery, LLC