Skip to Content.
Sympa Menu

sm-commit - Re: [SM-Commit] PERFORCE change 80208 by Daniel Goller for review

sm-commit AT lists.ibiblio.org

Subject: Source Mage code commit list

List archive

Chronological Thread  
  • From: Jeremy Blosser <jblosser-smgl AT firinn.org>
  • To: SM-Commit Daemon <sm-commit AT lists.ibiblio.org>
  • Subject: Re: [SM-Commit] PERFORCE change 80208 by Daniel Goller for review
  • Date: Mon, 5 Jun 2006 14:10:53 -0500

On Jun 05, Eric Sandall [eric AT sandall.us] wrote:
> Jeremy Blosser wrote:
> > On Jun 05, Eric Sandall [eric AT sandall.us] wrote:
> >> Perforce Review Daemon wrote:
> >>> Differences ...
> >> <snip>
> >>> @@ -1,3 +1,8 @@
> >>> +2006-06-04 Daniel Goller <morfic AT sourcemage.org>
> >>> + * DEPENDS: remove -stl/-no-stl flags from optional stlport depend to
> >>> avoid turning off STL support
> >>> + * BUILD: add -stl to configure to ensure building with STL support,
> >>> + fixes build errors with celestia and scribus that depend on it
> >>> +
> >>> 2006-05-22 Treeve Jelbert <treeve01 AT pi.be>
> >>> * DETAILS: Updated to 4.1.3
> >>> * BUILD: remove invoke_gcc ( done in default_pre_build)
> >> People will probably complain about this, but I think you should add
> >> PATCHLEVEL in DETAILS so that qt4 is recompiled with this option so STL
> >> is enabled for celestia and scribus, otherwise people will have to just
> >> 'know' they need to recompile to get these fixes rather than having it
> >> automagic.
> >
> > If it weren't qt, maybe, but that thing takes forever. Not having this
> > change does not break any existing installs, if they were "broken" before
> > only then do they need it.
> >
> > This one is probably on the border but because it is and because qt is
> > huge
> > I'm going to argue "no, let the local admin manage it themselves".
>
> Yes, qt is huge and takes a while to compile, but IMO it's better to
> guarantee that everyone's box is working without users having to know if
> they are affected or not (some may just notice scribus or celestia fail
> and not know why, perhaps this also affects other packages?).

This is *exactly* the argument "other distros" make for pre-configuring
things for people: "It's better to guarantee that everyone's box is working
without users having to know if they are affected or not". It seems a
double standard to me that we make sure we let them make their own
decisions on *how* to build things but don't mind being aggressive in
telling them *when* to build things.

> In the end it's up to the gurus (and Arwed) to make the call, but I'd
> rather have my CPU do some extra work to make sure I will have no
> problems later on, you know, when we've all forgotten about this bug.

Quite a few people won't be able to afford the cost of rebuilding qt out of
the blue, and so they'll skip that update for now. The farther we go in
this direction the less likely people are going to be to able to hold in
their heads what updates are really needed for their boxes, and we start to
lose "easy and straightforward to maintain" as a distro feature.

Attachment: pgp9Zuo7lc2Jr.pgp
Description: PGP signature




Archive powered by MHonArc 2.6.24.

Top of Page