Bugzilla Bug 195543

can we get a newer pan into stable please ?  the 0.14.x series and older blow
compared to the newer 0.13x

Looks ok to me.

Hi arches, please stabilize this package. Target keywords are :
alpha amd64 hppa ppc ppc64 sparc x86

it'll allow for a little clean up right after that, thanks in advance.

seems like this new pan is missing quite a few features from the old one (such
as loading the number of unread articles from groups without loading the whole
group!)

ppc and ppc64 done

x86 stable

alpha is not marked stable, we'll pass for now

fails tests on sparc:

PASS test #66 (numbers-test.cc, 199)
PASS: numbers-test
/bin/sh: line 4: 26729 Bus error               ${dir}$tst
FAIL: scorefile-test
PASS test #1 (text-massager-test.cc, 23)
PASS test #2 (text-massager-test.cc, 29)
PASS test #3 (text-massager-test.cc, 35)

Lovely test suite. Stable for HPPA.

Marked stable on amd64.

@sparc, any chance we can get this forward, some gcc/gtk/glib version would be
appreciated. If you can get a trace of the crash in tests, I'm sure we could
help upstream fix it. Besides is the app working at all ?

There is a coding error in quark.h, thus:
At line 89, we have
const char * data (s.str);

Then at line 100, we have:
hash  += get16bits (data);

 And at line 42, we have:
#define get16bits(d) (*((const uint16_t *) (d)))

This is PROHIBITED on sparc because a 16-bit data item must be aligned (mod
16), but in general, of course, a (char) is aligned only (mod 8).

Thus, sparc cannot go stable with this sort of coding in the source, because in
general there is no reason to believe pan would work.

For bookkeeping, the test that fails is:
scorefile-test

thanks for the explanation. Now I need to go stab upstream, yay :)

adding upstream report for tracking.

There's now a patch in upstream-SVN, which will be 0.133.  And a request to
test on the upstream bug.  I just updated to svn-r350 here and see it in the
changelog.  It's running, but I'm on amd64 so that's expected.  A sparc test of
either the patch or current svn would be appreciated.

(BTW, if anyone wants an svn.eclass based pan-9999 live-svn ebuild, that's what
I'm running locally.  Just ask.  It'd be nice to have it in the tree, masked of
course, but...)

(BTW 2, current svn and therefore 0.133 merge or otherwise eliminate the need
for all four current Gentoo pan-0.132-rX patches, bringing upstream up to date
with current gcc and glib, killing that buffer overflow, and fixing the pixbuff
thing too.)

Duncan

send the live ebuild my way.

(In reply to comment #12)
> A sparc test of either the patch or current svn would be appreciated.

All tests pass now, I also left comment on the upstream bug.

net-news/gnome-team: Can you bump to 0.133, please?

closing since all arches that could stabilize this did

@sparc, see you in the 0.133 stabilization bug :)