clone 584272 -1
reassign -1 myspell-hu
severity -1 critical
found -1 0.99.4-1.3
retile -1 dictionary broken, makes hunspell crash (and e.g. enchant/pidgin)
reassign 584272 libhunspell-1.2-0
retitle 584272 crashes with malformed dictionaries
severity 584272 important
block -1 by 476677
thanks
Hi,
On Wed, Jun 09, 2010 at 12:02:14AM +0200, Rene Engelhard wrote:
> On Tue, Jun 08, 2010 at 10:30:02PM +0200, Paul Chany wrote:
> > > On Tue, Jun 08, 2010 at 09:06:21PM +0200, Paul Chany wrote:
> > >> Rene Engelhard <r...@debian.org> writes:
> > >>
> > >> > On Sun, Jun 06, 2010 at 07:35:11PM +0100, Iulian Udrea wrote:
> > >> >> 1259.08 <ari> iulian: the problem is in hunspell
> > >>
> > >> > I doubt that it's a problem of hunspell - but if it is it's a
> > >> > dictionary bug.
> > >> > Paul, I see you use hu_HU as locale, are you also using the hungarian
> > >> > myspell/hunspell dictionary with enchant? If yes, which one? Does it
> > >> > work if you exchange them? Or remove the one you use=
> > >>
> > >> I have installed hunspell version 1.2.11-1 but I haven't installed
> > >> hunspell-hu.
> > >
> > > And myspell-hu?
> >
> > I have installed myspell-hu version 0.99.4-1.3
>
> Aha. So myspell-hu is buggy.
For that case, for reference, we had a discussion on IRC yesterday
about that:
17:50 < _rene_> ari: looks like a pattern bug to me
17:50 < _rene_> eh
17:50 < _rene_> s/pattern/dictionary/
17:51 -!- nthykier [~nthyk...@217.237.166.202] has quit [Ping timeout: 480
seconds]
17:52 -!- guus [g...@234-166.surfsnel.dsl.internl.net] has quit [Quit: Changing
server]
17:52 < ari> right
17:59 < _rene_> bu still this is just a HUNSPELL_WARNING(), not a crash which
should prevent pidgin from doing something
17:59 < _rene_> so I am still not convinced it's a hunspell bug. yes, the dict
probably is buggy and hunspell complains, but...
18:00 < _rene_> Paul sent no bt with anything hunspellish in it
18:01 < ron> after its done warning, does it return a NULLish (or just invalid)
something back to pidgin?
18:02 < ron> error handling isn't among bigstar's strong points
[...]
host closed the connection]
18:05 < _rene_> ron: -1 afaics if at all, but that only if malloc() fails. See
http://www.google.com/codesearch/p?hl=de#L2WY3uEXPxg/trunk/Hunspell/src/hunspell/hashmgr.cxx&q=bad%20flagvector&sa=N&cd=2&ct=rc
18:05 < zumbi> is BTS slowish/dead?
18:08 < _rene_> ron: line 505 ff
18:09 < ron> _rene_: yeah, found it, but that's the most fucked up interface
ever. I can't use the normal browser text search in it. yay goog
18:09 < ron> so what if len == 0 ?
18:09 < ron> or at least gets turned into it after being 1
18:11 -!- marga [~ma...@190.55.191.246] has quit [Ping timeout: 480 seconds]
18:11 -!- marga (Margarita Manterola) [~ma...@190.55.191.246] has joined
#debian-devel
18:12 -!- sabdfl [~sab...@loquat.canonical.com] has quit [Quit: Leaving.]
18:13 < _rene_> ron: yeah, that could be..
18:14 < ron> _rene_: or any odd length maybe ... len / 2 * sizeof(short)
isn't big enough to hold len if anything tries
18:15 < ron> dumping the value of len when it fails doesn't sound like useless
information to follow in any case
18:16 < _rene_> yeah, please write to that bug, will forward that upstream when
we know which dict he uses
18:16 < ron> ari: ^
18:18 < ari> i really have no idea at this point if the warning is related to
the crash at all
18:19 < ari> frankly the rest of the replies are just confusing
18:21 -!- sabdfl (Mark Shuttleworth) [~sab...@loquat.canonical.com] has joined
#debian-devel
18:22 < _rene_> ari: ack
18:22 < _rene_> ron: you can't do? ;)
Cloning the bug to myspell-hu, making it critical as it "breaks unrelated
software" and downgrading this one to important. Will talk with upstream...
Grüße/Regards,
René
--
.''`. René Engelhard -- Debian GNU/Linux Developer
: :' : http://www.debian.org | http://people.debian.org/~rene/
`. `' r...@debian.org | GnuPG-Key ID: D03E3E70
`- Fingerprint: E12D EA46 7506 70CF A960 801D 0AA0 4571 D03E 3E70
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
