[gentoo-dev] eapi1 bug/pkgcore sucks thread [was EAPI-2 - Let's get it started]

[Brian Harring]

On Wed, Jun 11, 2008 at 07:00:16PM +0100, David Leverton wrote:
> On Thursday 12 June 2008 02:46:03 Jim Ramsay wrote:
> > David Leverton <levertond@××××××××××.com> wrote:
> > > Since at least one ebuild has already been modified specifically to
> > > work around the bug, I'd say it's pretty real.
> >
> > For those of us trying to play along at home, which one is this?
>
> http://tinyurl.com/4w4t69

Few things I'll note about this stupid, stupid mess- looks of it, 
paludis folk have known about this for a while.  In other words, folk 
bitching about 'improving' QA intentionally sat on a bug for the sake 
of mocking, bug which according to them ebuild devs have supposedly 
worked around (yet to see it, but it's viable).

Useful to the whole, I'm sure.  Same folk in control of PMS for those 
playing the home game, politics over QA seemingly.

So what was the bug?  Aside from having to walk the full eapi-1 bugs,
(ebuild referenced wasn't of use), majority of which actually *is* 
tested in pkgcore (unlike portage which makes one wonder why pkgcore 
is targeted), the fault is a simple defaulting of an unset var being 
missed in implementing an undocumented spec (honestly, where is eapi1 
spec?).

Literally, the BS of the last day all comes down to inability to state 
the following:

=== modified file 'pkgcore/bin/ebuild-env/ebuild-functions.sh'
--- pkgcore/bin/ebuild-env/ebuild-functions.sh	2007-11-12 01:17:24 +0000
+++ pkgcore/bin/ebuild-env/ebuild-functions.sh	2008-06-11 22:24:16 +0000
@@ -236,7 +236,7 @@ src_compile
 {
     if [ "${EAPI:-0}" == 0 ] ; then
         [ -x ./configure ] && econf
-    elif [ -x ${ECONF_SOURCE}/configure ]; then
+    elif [ -x ${ECONF_SOURCE:-.}/configure ]; then
         econf || die "econf failed"
     fi
     if [ -f Makefile ] || [ -f GNUmakefile ] || [ -f makefile ]; then


Bit of a dumb bug, but it occurs unfortunately.  And yes, bash bits 
aren't currently tested since they're going to be completely ripped 
out and replaced (in the process shifting where/how it's accessed).

Why the exherbo/paludis/PMS folk decided to go this route to report, 
I'm not quite sure aside from assuming they're just griefers.

Regardless, fixed, released as 0.4.7.4, and in the tree.

Cheers
~harring