[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: gcc-3.4 roadmap?

Pete Harlan <harlan@artselect.com> writes:

> Thanks for your reply!
> On Thu, Aug 26, 2004 at 01:55:05PM +0200, Goswin von Brederlow wrote:
>> As I see it the sarge-amd64 fork will be gcc-3.3 compiled since
>> gcc-3.4 would need a lot more changes. That means the main sid archive
>> will also stay with gcc-3.3 and follow the normal debian transition to
>> gcc-3.4 over time.
> ...
>> > Does anyone have an educated guess as to when amd64 sid will move to
>> > gcc 3.4?
>> A year after sarge? Maybe faster if we agressively write patches and
>> ask for NMUs.
> This surprises me.  I'm not a Debian developer, but I would have
> guessed that as soon as Sarge were released, the first change to sid
> would be to move to the latest toolchain and kde/gnome/X/whatever,
> breaking as much stuff as possible and then repair and release as
> Etch.

We know everything will break if gcc-3.3 and gcc-3.4 is mixed (more or
less depending on the arch). Also just fixing isn't enough since
people must be able to update from stable at any time. That is the
reason why a there where a lot of library packages libfoo and
libfooc102 created for the gcc-3.2 transition. The same has to happen
now for gcc-3.4 resulting in lifboo or libfooc102 and libfooc104

Complicating this is that if liba depends on libb then first we need a
libbc104 before a libac104 can be uploaded. The longer those depends
chains are the longer it takes and gnome and kde have realy long chains.

> Which would be much more unstable than Debian has been since I've used
> it, which I suppose is why it doesn't work that way.
>> Packages can Build-Depend on gcc-3.4 and compile with it in the
>> "normal" pure64 as some packages (like all the mozilla browsers) are
>> already doing.
> Thanks, I didn't know that.  That's good news.
> --Pete


Reply to: