Re: Tcl/Tk plans for Squeeze
On Sat, Mar 13, 2010 at 08:54:33PM +0100, Francesco P. Lovergine wrote:
> > > Dropping 8.3 will require migrating old strict rev-dep packages
> > > to the new policy and possibly patching for using a modern Tcl (default
> > > or 8.4 at least). Eventually, some packages will be dropped, because too ancient.
> > > Let them to go.
> > What's the status of dropping 8.3 aka what still needs to be done to
> > make that happen?
> > Cheers
> > Luk
> One possibility could be movig tcl-defaults to 8.5 and disable threading on 8.4.
> Than dropping 8.3. Of course the transition should be more elaborate than that
> in order to avoid breakages for many packages, but it could be a basis of
> work. We will return with a more complete plans in a few, I hope...
> Of course if someone could fix hppa problems with multi-threading it would be great.
Ok, we found an agreement about a tentative plan for Tcl/Tk. I'm
submitting this plan to you in order to understand if there is the time
and will to allow this transition in squeeze.
At the very first stage, we will submit a tcltk-default with 8.5 in
*experimental* in order to allow all default dependent packages to use 8.5
(ideally by just binNMU). The 8.4 should be then built without
multi-threading, and all packages currently depending on tcl/expect 8.3 should
instead use 8.4 as such. All that could be done in experimental and by
NMUs. Iff all key packages had no problems on all archs, we could
consider uploading onto sid and then dropping 8.3. The staging area is
the only way to avoid unexpected breakages in sid ATM. Of course, we
could start with staging with or without RMs agreement, but the final
acceptance for sid will depend on freezing time and RM team decisions.
In the meantime we will do our home work in experimental.
Note that the list in  is wider than necessary, because automagically
computed by following rdeps and bdeps.
Francesco P. Lovergine