Public SourceMage Discussion List

[Andrew <afrayedknot AT thefrayedknot.armory.com>]

On Mon, Sep 26, 2005 at 04:42:20PM -0500, Jeremy Blosser (emrys) wrote:
> On Sep 26, Ladislav Hagara [ladislav.hagara AT unob.cz] wrote:
> > >
> > >
> > >> root@matrix:~# summon -d enlightenment
> > >> Downloading source file enlightenment-0.17-cvs.tar.bz2 for spell
> > >> enlightenment
> > >> Previous source found. Extracting package 
> > >> enlightenment-0.17-cvs.tar.bz2...
> > >> Running cvs update...
> > >> cvs [update aborted]: end of file from server (consult above messages 
> > >> if any)
> > >>
> > > ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > >
> > >> Creating package enlightenment-0.17-cvs.tar.bz2...
> > >>
> > >
> > > CVS is dying there too, looks like stable sorcery is just paying less
> > > attention to the exit status from CVS. IMO that would be a bug itself,
> > > since the checkout may not be complete...
> > >
> > 
> > enlightenment is extracting, running cvs, cvs is unavailable, 
> > enlightenment is packaging (the same as before).
> 
> In this case yes, but it would also be possible for cvs update to get
> halfway done and then die.  In either case sorcery would only have one exit
> code of failure, should it continue when it gets an error code as though it
> completed successfully?
> 
> > I have used it in this way from 2004-12-15. I have problems with 
> > cvs.sourceforge.net quite often but stable sorcery is great and never 
> > had problem with enlightenment sources.
> 
> You'd likely get the same CVS error right now if you tried to run cvs
> update manually.  Just because that error condition has not in the past
> caused a problem doesn't mean ignoring it is the right thing to do.
> 
> All that having been said, I don't know if test sorcery caring doing what
> it does now is intentional or not.  But I don't think it should ignore an
> error code from the cvs update command.

Stable sorcery did not check the return code, it assumes cvs always works.
Thats a bug. That and several other things were what prompted the major
re-factoring of download code.

Im open to suggestions on a better compromise behavior for failed scm
checkouts.

However, hropping the exit code on the floor is absolutely the wrong
thing to do. That people think thats someone "okay" is quite distressing.

On the other point:

It might be worthwhile to not look at fallback mirrors for non-tarball
downloads.

-Andrew

-- 
_________________________________________________________________________
| Andrew D. Stitt  | acedit at armory.com    | astitt at sourcemage.org | 
| irc: afrayedknot | Sorcery Team Lead       |      ftp://t.armory.com/ |
| 1024D/D39B096C   | 76E4 728A 04EE 62B2 A09A 96D7 4D9E 239B D39B 096C  |
-------------------------------------------------------------------------