Skip to Content.
Sympa Menu

sm-discuss - Re: [SM-Discuss] libpng, xorg-modular, and GCC

sm-discuss AT lists.ibiblio.org

Subject: Public SourceMage Discussion List

List archive

Chronological Thread  
  • From: David Kowis <dkowis AT shlrm.org>
  • To: sm-discuss AT lists.ibiblio.org
  • Subject: Re: [SM-Discuss] libpng, xorg-modular, and GCC
  • Date: Mon, 16 May 2011 13:50:27 -0500

On 05/14/2011 09:27 AM, Ladislav Hagara wrote:
> devel-libpng contains libpng 1.4.5 (I was updated it on 2010-12-10).
> Last 1.4 branch has already version 1.4.7.
> There is alredy 1.5 branch and the last stable libpng is 1.5.2.
>
> I wrote to the sm-discuss list [1], none conclusion, nobody is
> interesting. :-(

I'd posit that if the spells that depend on it, build and install, maybe
even work correctly, that it's good to be integrated. If that number is
too great, then we can opt for the most commonly used ones. Lacking that
data, go with half of them.

>
>
> devel-xorg-modular (X.org 1.10) is OK. I updated X.Org to 7.6 before
> Xmas and sent email to sm-discuss [2] mentioned problem with
> libpthread-stubs. Nobody is interesting, nobody merged (yea, some
> individual spells were merged). :-(

Difficult to test this one, unless we do a build of it, and ensure that
it works. Like all the X functionality works. Perhaps as people test it,
they should write up what they tested, and what the expected result is,
so we'll have a test plan for next time.

>
>
> devel-gcc is a evidence that if we want to stop updating spells it is
> sufficient to create new branch. Last version update of gcc in test
> grimoire was on 2010-04-01 into version 4.4.3. ;-( Nobody is really
> interesting. :-(

I agree with a later email. Fix the version and do the merge with
merged-gcc. Either build an automated test to ensure that GCC builds
things as we expect, or run the GCC test cases that come with it.
Obviously, not on every install, but for this test install, we should.

>
>
> devel-gnome-3-0-0, I had created it. I updated all gnome spells. Nobody
> is interesting, nobody touched this branch. Instead there are fixes of
> deprecated gnome spells in test grimoire and neverending discussion if
> we will support gnome2. Gnome2 is dead! Gnome3 is working for me. We can
> merge it, of course we will have to also fix several conflicts.

I'm not sure how to deal with this, as it appears that Gnome 2 will be
in some level of demand for a while after gnome 3 takes over. Perhaps
splitting grimoires would be a good thing to do at this point.

>
>
> devel-firefox4, instead of merging it we provide old firefox version
> with security bugs in test. I updated firefox and xulrunner to 4.0.1 and
> 2.0.1 in devel-firefox4, how many tested? :-(

I don't want firefox4 yet. I'd like to have some way of keeping
firefox3. That's just my personal opinion.

>
>
> Also we have devel-udev branch. I have updated udev to version 168 with
> changes in init.d. How many tested? The problem is with /run directory.
> Should it be part of smgl-fhs spell or should be created by udev spell?
> Some developers strictly disaprove that /run was on their boxes. We need
> conclusion.

I say put run in smgl-fhs. The 4k you're going to lose by having an
extra directory you don't use isn't worth the pain. Keeping it simple. I
think, regarding the systemd discussions, that more things are going to
use /run because /var/run may not exist early in the boot cycle, and it
gives a good spot for things like mtab.

>
>
> I am really upset with last Source Mage development. How many active
> developers do we have? How it is possible that core developers use half
> of year old sorcery? Core developers don't update their boxes. Instead
> of using upstream they trying to fix deprecated spell versions. Last
> months I really think about new Source Mage fork. :-(

Priorities change. People want different things than they used to. We
need to automate testing, so that the manual labor is removed. Those
kinds of things make it really difficult for me to use SMGL for
everything. I cannot update my box once a month, because it often
results in several hours of me fixing things. I don't have time for that
anymore. That's the simple truth.

I think we need to split up grimoires, have a core grimoire that we
slowly build tests into, and we have a defined testing process, as
automated as possible, to ensure that every bit of software in that core
grimoire is stable. Obviously, that will be released far slower, but I
think it's a really good idea.

I'll do some more thinking and planning to try to get some of these
issues solved.

Sorry for my absence, I haven't been as involved as I'd wanted to be as
Real Life came up and bit my bum.

David

Attachment: signature.asc
Description: OpenPGP digital signature




Archive powered by MHonArc 2.6.24.

Top of Page