Skip to Content.
Sympa Menu

sm-sorcery - Re: [SM-Sorcery] splitting BUILD

sm-sorcery AT lists.ibiblio.org

Subject: Discussion of Sorcery related topics

List archive

Chronological Thread  
  • From: Casey Harkins <charkins AT upl.cs.wisc.edu>
  • To: Andrew <afrayedknot AT thefrayedknot.armory.com>
  • Cc: SM Sorcery <sm-sorcery AT lists.ibiblio.org>
  • Subject: Re: [SM-Sorcery] splitting BUILD
  • Date: Wed, 10 Mar 2004 21:21:22 -0600 (CST)


I just checked in test, stable, games and z-rejected and see only 28
spells that use POST_BUILD. Only sybase-ase in z-rejected uses both
POST_BUILD and POST_INSTALL. The POST_INSTALL in sybase-ase only echo's a
message that the spell was installed to /opt and will not be tracked.

So do we even need POST_BUILD? Could the 28 spells that use it just mv it
to POST_INSTALL?

This does seem like a good time to bring up another idea, which I was
going to propose post-1.0. There has been talk now and then about using a
shared sorcery cache for installing multiple machines, or using the cache
files on the iso to install the basesystem individually instead of using a
monolithic image.tar.bz2. For this to work reliably, most of the stuff
done in POST_INSTALL should be done upon resurrecting a spell. However,
there are a few instances[0] where the POST_INSTALL uses the
SOURCE_DIRECTORY to do something. This obviously wouldn't be available
when resurrecting.

Perhaps POST_INSTALL could be used for any changes that should be done
upon resurrecting, with the requirement that it should not reference
SOURCE_DIRECTORY. POST_BUILD could then be used in the few (12 in stable)
where the SOURCE_DIRECTORY is used. The only change in sorcery would be to
run POST_INSTALL on resurrection.

-casey



On Wed, 10 Mar 2004, Andrew wrote:
> >
> > If it's truly backwards compatible (which it sounds like it is) I see no
> > reason
> > not to try it as cleanup is our goal, and it sounds like this would do
> > some
> > cleanup (separating some of the libraries into lib files) along with
> > fixing a
> > few of the filed bugs.
> Its totally backwards compatible. The code looks for INSTALL or COMPILE,
> and if they exist follows the new behavior, otherwise it reverts to the
> old.
>
> One issue is that POST_BUILD and POST_INSTALL are now both run after
> INSTALL (which is the last half of BUILD). This presents a slight
> inconsistancy, as POST_INSTALL now runs after POST_BUILD rather than
> immediatly after INSTALL like one would thing. So a global renaming
> of all the POST_BUILD/POST_INSTALL files might be in order. But theres
> certainly no reason to not allow that file to still work during a
> transition period (just as with BUILD). Does that make sense?
>
> What should POST_BUILD be renamed to? Or should we leave it with its
> contradictory name?
>
> -Andrew
>
> --
> ________________________________________________________________________
> |Andrew D. Stitt | astitt AT sourcemage.org |
> |irc: afrayedknot | afrayedknot AT t.armory.com |
> |aim: thefrayedknot or iteratorplusplus | |
> |Grimoire Guru ham/smgl, Porting Team Lead | ftp://t.armory.com |
> |Author and Maintainer of Prometheus | |
> -------------------------------------------------------------------------
> _______________________________________________
> SM-Sorcery mailing list
> SM-Sorcery AT lists.ibiblio.org
> http://lists.ibiblio.org/mailman/listinfo/sm-sorcery
>




Archive powered by MHonArc 2.6.24.

Top of Page