gtk+-2.8.14 version bump

[sofar]

I suggest that we do the following:

- keep glib-2 in moonbase as is
- all: test by recompiling most glib-2 dependent software
- report on possible issues

In the future, I suggest that whenever high impact changes are imminent, that we send out an RFT message to the lunar-dev mailinglist before the change is made. Something like :

Subject: RTF: lin -w 10.0 glib-2

should notify and alert all devs that they are required to do testing on it and report back - only this way we can make sure that:

- the module was actually tested by people
- you get feedback on it

maybe it will make this type of upgrades easier.

As for the glib-2 issue - there is no proof to roll the module back - the only thing to do is verification - thus more testing. And feeding it back. 

Auke

On Thu, 9 Mar 2006 14:02:23 -0600, Terry Chan <tpchan at comcast.net> wrote:
> I have been running with glib-2.10.0 for 2 days straight on this lunar box
> with no segfaults (excluding the one time I tried that piece of crap, AKA
> galculator). I've now been running glib-2.10.1 for over 24 hrs straight,
> again
> with no segfaults. Compiling many modules in moonbase all from inside
> xfce4.
> 
> I'm still waiting for some real posts of verifiable problems with
> glib-2.10.1.
> Problems that are severe enough to justify holding off on gtk+-2.8.14.
> 
> Terry Chan
> -----------------------------------------------------------
> On Thu, Mar 09, 2006 at 09:44:16AM +0100, Stefan Wold wrote:
>> The thing is that the problem doesn't start occuring right away, so a
>> quick start and stop of a program won't reveal the crashes. It seems you
>> have to be running a few hours before the problem occur. After that the
>> apps just refuse to start. At that point restarting the computer or
>> downgrading glib and rebuilding pango seem to be the only solution. I'm
>> not saying glib-2.10.x is the full cause of this, it could be
> glib-2.10.x
>> in conjuction with gtk and pango.
>>
>> Downgrading glib did however solve my crashes immediately.
>>
> _______________________________________________
> Lunar-dev mailing list
> Lunar-dev at lunar-linux.org
> http://foo-projects.org/mailman/listinfo/lunar-dev