Bugzilla Bug 179645

sci-biology/tree-puzzle :0
  5.1 -> nothing on default-linux/ppc/ppc32/2006.1 (unstable)
  5.1 -> nothing on default-linux/ppc/ppc32/2007.0 (unstable)
  5.1 -> nothing on default-linux/ppc/ppc32/2007.0/desktop (unstable)
  5.1 -> nothing on default-linux/ppc/ppc32/2007.0/G3 (unstable)
  5.1 -> nothing on default-linux/ppc/ppc32/2007.0/G3/Pegasos (unstable)
  5.1 -> nothing on default-linux/ppc/ppc32/2007.0/G4 (unstable)
  5.1 -> nothing on default-linux/ppc/ppc32/2007.0/G4/Pegasos (unstable)
  5.1 -> nothing on default-linux/ppc/ppc32/2007.0/desktop/G3 (unstable)
  5.1 -> nothing on default-linux/ppc/ppc32/2007.0/desktop/G3/Pegasos
(unstable)
  5.1 -> nothing on default-linux/ppc/ppc32/2007.0/desktop/G4 (unstable)
  5.1 -> nothing on default-linux/ppc/ppc32/2007.0/desktop/G4/Pegasos
(unstable)
  5.1 -> nothing on default-linux/ppc/ppc64/2007.0/32bit-userland (unstable)
  5.1 -> nothing on default-darwin/macos/10.4 (stable)
  5.1 -> nothing on hardened/ppc (unstable)
  5.1 -> nothing on selinux/ppc/ppc32/2006.1/G3 (unstable)
  5.1 -> nothing on selinux/ppc/ppc32/2006.1/G4 (unstable)

Apparent cause:

  21 May 2007; Olivier Fisette <ribosome@gentoo.org>
  -tree-puzzle-5.1.ebuild:
  Pruned old version.

5.1 was the last ~ppc version.

This is indeed my mistake, and I have no clue as to why ppc was dropped when I
added version 5.2. I suggest marking version 5.2 ~ppc rather than bringing back
version 5.1.

To test Tree-Puzzle, start an interactive session with "puzzle -randseed=1".
Choose /usr/share/tree-puzzle/data/atp6.a as the input file. Do not change any
option; use defaults. The analysis should be very fast. 3 output files will be
created in the working directory. Compare these with the attached reference
results. The only differences should be timestamps and computation times.

Created an attachment (id=120322) [details]
outdist

Created an attachment (id=120323) [details]
outfile

Created an attachment (id=120325) [details]
outtree

It fails its own tests:
=====================
14 of 17 tests failed
=====================

Your interactive test checks out fine though.

(In reply to comment #5)
> It fails its own tests:
> =====================
> 14 of 17 tests failed
> =====================
> 
> Your interactive test checks out fine though.
> 

I would have been surprised if it had worked. ;-)

The test suite is broken as it uses diff to compare results where round-up
errors are likely. I think the test I provided is enough, but if you want to
use the upstream-provided tests, I am afraid you will have to manually check
the output of "make check"... 

Good enough for me. ~ppc keyword added.