freezing at computing transaction

[Rehan Khan]

Imported: http://tracker.labix.org/issue12
Reason for Import: Patch Review, Issue Review

msg1200 (view) Author: kindjal Date: 2007-10-04.15:12:22

As per msg1192, this appears to imply we get stuck in the while loop beginning
at line 282 of smart/transaction.py. In my experience, the symptom described
can last a very long time, over an hour. I can't say that it's an infinite loop
though.

In my case, I'm using a repository of packages that include many of my own. I
wonder if the symptom is caused by some special set of upgrade paths and
dependencies, and since most people are using "standard" distributions (RedHat,
Centos, etc.), then most people aren't seeing this. It's just those few of us
with lots of our own RPMs that trigger this.

Perhaps the logic that leads to the queue.append() in this code needs some added
conditions? Or perhaps a loop could be detected and a warning issued? Any
thoughts?

msg1192 (view) Author: mxsscott Date: 2007-09-26.21:09:36

I'm seeing the same issue with smart 0.50, this time on Centos 4.4.

strace is showing a similar mremap as previously commented. The following are
backtraces with gdb and python. I'm seeing it every few days, so if any more
information is required let me know.

gdb:
bt
#0 0x002f87e7 in PyList_GetSlice () from /usr/lib/libpython2.3.so.1.0
#1 0x002f8e35 in PyList_SetSlice () from /usr/lib/libpython2.3.so.1.0
#2 0x00304961 in PyCFunction_Call () from /usr/lib/libpython2.3.so.1.0
#3 0x0033760c in _PyEval_SliceIndex () from /usr/lib/libpython2.3.so.1.0
#4 0x00338a7b in _PyEval_SliceIndex () from /usr/lib/libpython2.3.so.1.0
#5 0x00338a7b in _PyEval_SliceIndex () from /usr/lib/libpython2.3.so.1.0
#6 0x00338a7b in _PyEval_SliceIndex () from /usr/lib/libpython2.3.so.1.0
#7 0x00339196 in PyEval_EvalCodeEx () from /usr/lib/libpython2.3.so.1.0
#8 0x00337e89 in _PyEval_SliceIndex () from /usr/lib/libpython2.3.so.1.0
#9 0x00338a7b in _PyEval_SliceIndex () from /usr/lib/libpython2.3.so.1.0
#10 0x00339196 in PyEval_EvalCodeEx () from /usr/lib/libpython2.3.so.1.0
#11 0x0033945d in PyEval_EvalCode () from /usr/lib/libpython2.3.so.1.0
#12 0x00352917 in PyErr_Display () from /usr/lib/libpython2.3.so.1.0
#13 0x00353a42 in PyRun_SimpleFileExFlags () from /usr/lib/libpython2.3.so.1.0
#14 0x00354a94 in PyRun_AnyFileExFlags () from /usr/lib/libpython2.3.so.1.0
#15 0x0035978e in Py_Main () from /usr/lib/libpython2.3.so.1.0
#16 0x080485b2 in main ()

(funky code to get python bt)
call PyRun_SimpleString("import sys, traceback;
sys.stderr=open('/tmp/tb','w',0); traceback.print_stack()")
file /tmp/tb:
  File "/usr/bin/smart", line 194, in ?
    main(sys.argv[1:])
  File "/usr/bin/smart", line 167, in main
    exitcode = iface.run(opts.command, opts.argv)
  File "/usr/lib/python2.3/site-packages/smart/interface.py", line 53, in run
    result = _command.main(self._ctrl, opts)
  File "/usr/lib/python2.3/site-packages/smart/commands/upgrade.py", line 152,
in main
    trans.run()
  File "/usr/lib/python2.3/site-packages/smart/transaction.py", line 1463, in run
    self._policy.runStarting()
  File "/usr/lib/python2.3/site-packages/smart/transaction.py", line 283, in
runStarting
    path = queue.pop(0)
  File "<string>", line 1, in ?

msg1158 (view) Author: kindjal Date: 2007-06-25.18:07:16

For the record, I have seen this again in smart 0.50.

smart upgrade gets stuck, stracing it shows the following line repeating
over and over again:

mremap(0xb68ea000, 790528, 790528, MREMAP_MAYMOVE) = 0xb68ea000
mremap(0xb68ea000, 790528, 790528, MREMAP_MAYMOVE) = 0xb68ea000
mremap(0xb68ea000, 790528, 790528, MREMAP_MAYMOVE) = 0xb68ea000
mremap(0xb68ea000, 790528, 790528, MREMAP_MAYMOVE) = 0xb68ea000

Reading the mremap man page, this looks like essentially a "noop". The
old and new sizes are identical, and the returned address is the same as
the starting one.

I have to kill -9 the smart process, and re-running "smart upgrade" does
the same thing with a different memory location. When I see it, it will repeat
several times in a row, but if I reboot and try again, the issue does not repeat.

msg1091 (view) Author: niemeyer Date: 2007-02-13.15:14:10

This issue hasn't happened for a while, after the recent improvements on 0.50.
I'm closing it.

msg1069 (view) Author: niemeyer Date: 2007-01-15.22:45:55

> I'll give another try this week-end. I'll try to have a deeper
> understanding on how it should work and what might cause the problem.

That'd be fantastic, but if you manage to at least save the
system state in a moment the problem is reproducible, we can
debug it and fix it.

msg1068 (view) Author: mathieu.clabaut Date: 2007-01-15.22:37:00

Oh, bad news....
I'll give another try this week-end. I'll try to have a deeper understanding
on how it should work and what might cause the problem.

Anyway, thank you for your investigations. I'll keep you informed of any
progress on my side.

-mathieu

On 1/15/07, Gustavo Niemeyer at Labix Tracker <email address hidden> wrote:
>
>
> Gustavo Niemeyer <email address hidden> added the comment:
>
> Mathieu, I got your files, and tried to reproduce the issue you mention by
> running the upgrade operation. Smart presents only 26 packages to
> upgrade, so
> I'm unable to reproduce the problem you mention with the presented data.
>
> I'm going to consider this issue as resolved since there are no further
> reports
> of that kind of problem, and the provided information isn't sufficient.
>
> ----------
> assignedto: tromer -> niemeyer
>
> _______________________________________
> Labix issue tracker <email address hidden>
> <http://tracker.labix.org/issue12>
> _______________________________________
>

msg1067 (view) Author: niemeyer Date: 2007-01-15.18:50:26

Mathieu, I got your files, and tried to reproduce the issue you mention by
running the upgrade operation. Smart presents only 26 packages to upgrade, so
I'm unable to reproduce the problem you mention with the presented data.

I'm going to consider this issue as resolved since there are no further reports
of that kind of problem, and the provided information isn't sufficient.

msg1023 (view) Author: mathieu.clabaut Date: 2006-12-08.20:01:59

I give some days ago another try to upgrade form FC5 to FC6 with the SVN
revision r829, and it failed with a bunch of errors like :
  erreur: file /usr/share/man/man1/lsb_release.1.gz from install of
redhat-lsb-3.1-11 conflicts with file from package redhat-lsb-3.0-9.2

The rpm database (/var/lib/rpm) and smart setup (/var/lib/smart) are uploaded to
http://mathieu.clabaut.free.fr/smart/

Revision 706 of smart does not finish when calculating transaction (it takes for
ever, I've interupted calculation after several hours)...

I later tried to upgrade some package from FC5 to FC5 (no more FC6) with smart
HEAD (r836) and it gives me the following errors :

ImageMagick-c++-6.2.5.4-4.2.1.fc5.3 requiert ImageMagick = 6.2.5.4-4.2.1.fc5.3
git-svn-1.4.1-1.fc5 requiert git-core = 1.4.1-1.fc5
gitk-1.4.1-1.fc5 requiert git-core = 1.4.1-1.fc5
git-email-1.4.1-1.fc5 requiert git-core = 1.4.1-1.fc5
git-cvs-1.4.1-1.fc5 requiert git-core = 1.4.1-1.fc5

Tell me if I can help in anything to resolve this issue.

msg917 (view) Author: niemeyer Date: 2006-11-11.17:27:06

Mathieu, would it be possible for you to upload your /var/lib/rpm and
/var/lib/smart directories to somewhere? Notice that if there's any
content in /var/lib/smart/packages, this may be removed, but
/var/lib/smart/channels is important.

Thanks in advance,

msg916 (view) Author: niemeyer Date: 2006-11-11.17:19:08

> Retrying a fc5 -> fc6 with last SVN trunc (r810) in text mode, all package
> were downloaded and the smart failed with the following error :
> Committing transaction...
> error: valgrind-callgrind-0.10.1-1.1 requires valgrind = 1:3.1.0
>
> After checking, I've neither valgring nor valgrind-callgrind installed on
> the FC5...

This doesn't say much. It'd be nice to inspect these packages and check
which of them are in the transaction and why. Can you please issue

  smart query --show-all valgrind-callgrind

And confirm that it is in the selected transaction, and that valgrind
isn't?

Thanks,

msg915 (view) Author: mathieu.clabaut Date: 2006-11-11.07:00:33

Retrying a fc5 -> fc6 with last SVN trunc (r810) in text mode, all package
were downloaded and the smart failed with the following error :
Committing transaction...
error: valgrind-callgrind-0.10.1-1.1 requires valgrind = 1:3.1.0

After checking, I've neither valgring nor valgrind-callgrind installed on
the FC5...

-mathieu

On 11/10/06, Mathieu Clabaut <email address hidden> wrote:
>
> Mmmmm, I started it with --gui only.... Is it a bad thing to do ?
>
> On 11/10/06, Gustavo Niemeyer at Labix Tracker < <email address hidden>> wrote:
> >
> >
> > Gustavo Niemeyer < <email address hidden>> added the comment:
> >
> > > I just tried fc5 -> fc6 on x86_64, and the calculation wents ok in
> > > less than 4 minutes ! Yeepee !
> >
> > Awesome!
> >
> > > But ! (Yes ,there's a but), asking to do the upgrade results i nthe
> > > following backtrace :
> >
> > Oops..
> >
> > Please try to run it without --yes (or -y). We'll make sure that this
> > problem is not present in the final version.
> >
> > _______________________________________
> > Labix issue tracker < <email address hidden>>
> > <http://tracker.labix.org/issue12>
> > _______________________________________
> >
>
>

msg914 (view) Author: mathieu.clabaut Date: 2006-11-10.17:55:23

Mmmmm, I started it with --gui only.... Is it a bad thing to do ?

On 11/10/06, Gustavo Niemeyer at Labix Tracker <email address hidden> wrote:
>
>
> Gustavo Niemeyer <email address hidden> added the comment:
>
> > I just tried fc5 -> fc6 on x86_64, and the calculation wents ok in
> > less than 4 minutes ! Yeepee !
>
> Awesome!
>
> > But ! (Yes ,there's a but), asking to do the upgrade results i nthe
> > following backtrace :
>
> Oops..
>
> Please try to run it without --yes (or -y). We'll make sure that this
> problem is not present in the final version.
>
> _______________________________________
> Labix issue tracker <email address hidden>
> <http://tracker.labix.org/issue12>
> _______________________________________
>

msg913 (view) Author: niemeyer Date: 2006-11-10.01:08:48

> I just tried fc5 -> fc6 on x86_64, and the calculation wents ok in
> less than 4 minutes ! Yeepee !

Awesome!

> But ! (Yes ,there's a but), asking to do the upgrade results i nthe
> following backtrace :

Oops..

Please try to run it without --yes (or -y). We'll make sure that this
problem is not present in the final version.

msg912 (view) Author: mathieu.clabaut Date: 2006-11-09.20:41:26

I just tried fc5 -> fc6 on x86_64, and the calculation wents ok in less than
4 minutes ! Yeepee !

But ! (Yes ,there's a but), asking to do the upgrade results i nthe
following backtrace :

Traceback (most recent call last):

  File
"/usr/lib64/python2.4/site-packages/smart/interfaces/gtk/interactive.py",
line 171, in callback
    exec code in globals

  File "<callback>", line 1, in ?

  File
"/usr/lib64/python2.4/site-packages/smart/interfaces/gtk/interactive.py",
line 457, in upgradeAll
    self.applyChanges(confirm=not emptychangeset)

  File
"/usr/lib64/python2.4/site-packages/smart/interfaces/gtk/interactive.py",
line 415, in applyChanges
    if self._ctrl.commitTransaction(transaction, confirm=confirm):

  File "/usr/lib64/python2.4/site-packages/smart/control.py", line 505, in
commitTransaction
    return self.commitChangeSet(trans.getChangeSet(), caching, confirm)

  File "/usr/lib64/python2.4/site-packages/smart/control.py", line 512, in
commitChangeSet
    iface.showChangeSet(changeset)

  File "/usr/lib64/python2.4/site-packages/smart/__init__.py", line 76, in
__getattr__
    return getattr(self.object, attr)

AttributeError: 'GtkInteractiveInterface' object has no attribute
'showChangeSet'

On 11/9/06, Gustavo Niemeyer at Labix Tracker <email address hidden> wrote:
>
>
> Gustavo Niemeyer <email address hidden> added the comment:
>
> > fc4->fc5, fc4->fc6, fc4->devel, fc5->fc6, all calculate updates and
> > upgrades in a few minutes on a vm with 512M, using standard repos +
> > Freshrpms and Dries. Much quicker than with 0.42.
>
> That's fantastic! Thanks for these tests Dave.
>
> _______________________________________
> Labix issue tracker <email address hidden>
> <http://tracker.labix.org/issue12>
> _______________________________________
>

msg909 (view) Author: niemeyer Date: 2006-11-08.23:06:08

> fc4->fc5, fc4->fc6, fc4->devel, fc5->fc6, all calculate updates and
> upgrades in a few minutes on a vm with 512M, using standard repos +
> Freshrpms and Dries. Much quicker than with 0.42.

That's fantastic! Thanks for these tests Dave.

msg905 (view) Author: dfarning Date: 2006-11-08.21:38:58

test 0.50rc1

fc4->fc5, fc4->fc6, fc4->devel, fc5->fc6, all calculate updates and upgrades
in a few minutes on a vm with 512M, using standard repos + Freshrpms and Dries.
Much quicker than with 0.42.

I have not been able to see if the upgrades complete corectly due to bandwidth
limitations.

Dave

msg886 (view) Author: netmask Date: 2006-11-07.17:44:08

Please, test with latest version (0.50rc1). We already received some reports of
this issue being fixed.

msg881 (view) Author: tromer Date: 2006-11-06.23:28:28

Should be fixed in current subversion trunk. It may still take a few minutes on
tough upgrades, but will usually run exponentially faster (literally) than before.

msg880 (view) Author: ittayd Date: 2006-11-06.12:51:13

works good so far. 3 minutes to compute the transaction.

msg874 (view) Author: tromer Date: 2006-11-06.10:04:44

Try the attached patch, which improves the transaction calculation algorithm.
For best effect, make sure you have non-identical channel priorities.

(This combined patch includes some unrelated changes; these are minor harmless.)

msg872 (view) Author: ittayd Date: 2006-11-02.16:25:19

this issue happens to me also trying to upgrade from fc4 to fc6.

 > smart channel --show
[core]
type = rpm-md
name = Fedora Core 6
priority = 100
baseurl = http://download.fedora.redhat.com/pub/fedora/linux/core/6/i386/os/

[dries]
type = rpm-md
name = Dries
priority = 100
baseurl = http://apt.sw.be/dries/fedora/fc6/i386/dries/RPMS

[livna-stable]
type = rpm-md
name = Livna Stable
priority = 10
baseurl = http://rpm.livna.org/fedora/6/i386

[kde-redhat-stable]
type = rpm-md
name = Kde Stable
priority = 10
baseurl = http://apt.kde-redhat.org/apt/kde-redhat/fedora/6/i386/stable/

[extras]
type = rpm-md
name = Fedora Extras 6
priority = 100
baseurl = http://download.fedora.redhat.com/pub/fedora/linux/extras/6/i386

[updates]
type = rpm-md
name = Fedora Updates 6
priority = 100
baseurl = http://download.fedora.redhat.com/pub/fedora/linux/core/updates/6/i386/

[rpm-db]
type = rpm-sys
name = RPM Database
priority = 100

[freshrpms]
type = rpm-md
name = Freshrpms
priority = 100
baseurl = http://ayo.freshrpms.net/fedora/linux/6/i386/freshrpms/

msg572 (view) Author: tomee Date: 2006-06-19.11:37:36

The same happens to me.
A few times smart showed upgrades withis seconds, and other times (with cooker
added and now -- don't know why) would freeze.
The LD_ thing doesn't work, python complaining about libpthreads (in fact no
command works with that variable set...)

The files you asked for will be available from http://tomee.kadu.net/smart.tar.gz

Please let me know when you download them cause I don't want to keep them there.

msg353 (view) Author: mathieu.clabaut Date: 2006-02-10.21:46:57

I tried a FC3 to FC4 upgrade with smart.
after installing fedora-release-4-2.noarch.rpm, I updated my repository list to
 match the FC4 version and then did a :
  smart update
which went ok, and a
  smart upgrade --stepped --explain

And now, 49 hours after, smart is still running without more informations than :
 "Calcul de la transaction..." (in french).

Is this an expected behaviour (i.e. smart can not handle such a big upgrade) ?
If not, how may I help to dig further (I'm totally python ignorant)..

msg339 (view) Author: risto.kankkunen Date: 2006-02-07.19:36:08

I did some investigating and found out it's Transaction._pending() in
/usr/lib/python2.4/site-packages/smart/transaction.py that's taking all the
time. I added some diagnostic prints to show call count, recursion depth and the
length of each parameter. When trying to update FF 1.5 from Fedora Core 4
Development, the log starts:

   1 [000] _pending(len(chageset)=6, len(locked)=13, len(pending)=6)

and 10 minutes later:

9014 [020] _pending(len(chageset)=347, len(locked)=345, len(pending)=1)
9015 [018] _pending(len(chageset)=342, len(locked)=338, len(pending)=1)

I tried the same operation with yum, and in 1 minute 13 seconds it complained
about 13 missing dependencies.

Next I tried installing FF 1.5 from nrpms.net/production with yum. In 11s it
offered to update firefox and cairo, pango and pango-devel. smart again started
to consume all CPU. After 2 minutes:

1752 [026] _pending(len(chageset)=310, len(locked)=381, len(pending)=2)
1753 [026] _pending(len(chageset)=318, len(locked)=387, len(pending)=1)
1754 [026] _pending(len(chageset)=318, len(locked)=387, len(pending)=1)
1755 [025] _pending(len(chageset)=305, len(locked)=374, len(pending)=3)

Pretty deep recursion there...

Hope this helps. I'm not familiar with the code base and don't know, if I have
time to delve deeper...

msg336 (view) Author: risto.kankkunen Date: 2006-02-05.20:17:31

Forgot to mention: my smart version is smart-0.41-26.rhfc4.at and kernel version
is 2.6.15-1.1830_FC4.

msg335 (view) Author: risto.kankkunen Date: 2006-02-05.20:15:02

I just migrated my yum settings (18 channels) to smart and did an upgrade after
a couple weeks of non-use. Smart handled the upgrade beautifully downloading
almost 300 MB of files at my full 1MB/s speed using its clever concurrent
downloading.

However, now it seems I cannot upgrade or install anything. I always end up into
a CPU-consuming loop. E.g. trying to install Firefox 1.5 from FC 4 Development
channel ties all the CPU. Stracing the system shows either:

     0.000075 [00d27402] futex(0xa80db88, FUTEX_WAKE, 1) = 0
     0.000071 [00d27402] futex(0xa80db88, FUTEX_WAKE, 1) = 0
     0.000069 [00d27402] futex(0xa80db88, FUTEX_WAKE, 1) = 0

or on other occasions:

     0.136168 [00b7d402] brk(0xad61000) = 0xad61000
     0.074208 [00b7d402] brk(0xad82000) = 0xad82000
     0.048184 [00b7d402] brk(0xad7f000) = 0xad7f000

It doesn't matter, whether I use the GUI or not. I'm not short on memory (512 MB
RAM, smart taking about 100 MB, no swapping) and the CPU should be fast enough
(Athlon XP 2000+).

I tried the LD_ASSUME_KERNEL setting from message msg101, but that just produced
loads of warning about undefined symbols.

Smart seems really promising, but at the moment I cannot use it at all.

msg216 (view) Author: netmask Date: 2006-01-09.13:36:20

Thanks for your reporting.

msg215 (view) Author: roos Date: 2006-01-09.11:35:56

Me too. 0.41 seems to do the trick, whatever it is :)
/roos

Citerar Jon at Labix Tracker <email address hidden>:

>
> Jon <email address hidden> added the comment:
>
> I have also installed 0.41 and have not had any trouble at all.
> Thanks!
>
> _______________________________________
> Labix issue tracker <email address hidden>
> <http://tracker.labix.org/issue12>
> _______________________________________
>

msg214 (view) Author: jon Date: 2006-01-09.01:39:05

I have also installed 0.41 and have not had any trouble at all.
Thanks!

msg207 (view) Author: Gary Date: 2006-01-04.05:22:53

I have installed and run 0.41. I am satisfied the the "freeze aat computing transaction" problem has been fully corrected. Thanks! ------- Original Message ------ Subject: [issue12] freezing at computing transactionTo:<email address hidden>, <email address hidden>, <email address hidden>, <email address hidden>, <email address hidden>, <email address hidden>:"Mauricio Teixeira (netmask) at Labix Tracker" Date: Mon, 02 Jan 2006 19:16:27 +0000Mauricio Teixeira (netmask) added the comment:Have you tried with current smart release (0.41)?----------nosy: +netmask_______________________________________Labix issue tracker _______________________________________

---------------------------------
Yahoo! Photos
 Ring in the New Year with Photo Calendars. Add photos, events, holidays, whatever.

msg204 (view) Author: roos Date: 2006-01-03.01:01:39

No!
I will in a few days though, ill get back to you...

Quoting "Mauricio Teixeira (netmask) at Labix Tracker"
<email address hidden>:

>
> Mauricio Teixeira (netmask) <email address hidden> added the
> comment:
>
> Have you tried with current smart release (0.41)?
>
> ----------
> nosy: +netmask
>
> _______________________________________
> Labix issue tracker <email address hidden>
> <http://tracker.labix.org/issue12>
> _______________________________________
>

msg202 (view) Author: netmask Date: 2006-01-02.19:16:26

Have you tried with current smart release (0.41)?

msg106 (view) Author: Gary Date: 2005-11-10.02:38:26

My first use of Smart 0.40-8 required an update of about 400 Mb using a 56k
modem running the gui from the Suse menu. Download was fine and at my best
download speed (yeah, go ahead and laugh). No changes from included channels. I
experienced the same problem described in this thread during computing
transactions. After waiting about an hour I killed the application (would not
respond otherwise) and restarted. It took up where it left off, then hung
again: I killed. This process repeated about 10 times before finishing (hey,
I'm impressed that it could pick up like that). Since then with normal daily
upgrades (small < 40 Mb) performance has been flawless in this area (currently
0.40-11).

msg101 (view) Author: jon Date: 2005-11-05.19:37:59

I've had some similar issues and I suspected threading.

Using the following paradigm resolved all of the problems I was having:

LD_ASSUME_KERNEL="2.4.0" <invoke smart here>
example:
LD_ASSUME_KERNEL="2.4.0" smart upgrade

msg47 (view) Author: quintesse Date: 2005-10-10.11:01:33

Might this nto be related to the fact that when there are a lot of packages to
consider smart just takes a very long time to calculate the upgrade? It has
happened to me several times as well, try upgrading FC3 to FC4 using smart for
example. After leaving it running for the night I cancelled it in the morning,
ran apt-get to do the upgrade and ran smart afterwards for the finishing touch.
Some kind of indication that it is still busy and does not hang (even better
would be some kind of progress bar of course if possible) might be nice though.

msg33 (view) Author: niemeyer Date: 2005-10-05.19:10:07

I need more information to debug this.

Can you please upload your /var/lib/smart/config and /var/lib/rpm to a public
place I could download? Notice that it may be a bit large.

Thanks.

msg27 (view) Author: roos Date: 2005-09-30.09:25:13

My smart seem to freeze for up to an hour when computing transactions. Then
suddenly it starts the installation. Sometimes after seconds, sometimes after
many, many minutes.

History
Date User Action Args
2007-10-04 15:12:23 kindjal set messages: + msg1200
2007-09-26 21:11:58 mxsscott set nosy: + roos, quintesse, jon, Gary, risto.kankkunen, ittayd, tomee, dfarning, tromer, kindjal
2007-09-26 21:09:42 mxsscott set nosy: + mxsscott, - roos, quintesse, tromer, jon, Gary, tomee, dfarning, risto.kankkunen, kindjal, ittayd
messages: + msg1192
2007-06-25 18:07:19 kindjal set status: resolved -> chatting
nosy: + kindjal
messages: + msg1158
2007-02-13 15:14:12 niemeyer set status: chatting -> resolved
messages: + msg1091
2007-01-15 22:45:57 niemeyer set messages: + msg1069
2007-01-15 22:37:04 mathieu.clabaut set files: + unnamed
status: resolved -> chatting
messages: + msg1068
2007-01-15 18:52:49 niemeyer set status: need-info -> resolved
2007-01-15 18:50:31 niemeyer set assignedto: tromer -> niemeyer
messages: + msg1067
2006-12-08 20:02:02 mathieu.clabaut set messages: + msg1023
2006-11-11 17:27:10 niemeyer set messages: + msg917
2006-11-11 17:19:09 niemeyer set messages: + msg916
2006-11-11 12:46:09 netmask set files: - unnamed
2006-11-11 12:46:03 netmask set files: - unnamed
2006-11-11 12:45:56 netmask set files: - unnamed
2006-11-11 12:45:43 netmask set files: - unnamed
2006-11-11 07:00:39 mathieu.clabaut set files: + unnamed
messages: + msg915
2006-11-10 17:55:26 mathieu.clabaut set files: + unnamed
messages: + msg914
2006-11-10 01:08:49 niemeyer set messages: + msg913
2006-11-09 20:41:27 mathieu.clabaut set files: + unnamed
messages: + msg912
2006-11-08 23:06:07 niemeyer set messages: + msg909
2006-11-08 21:39:01 dfarning set nosy: + dfarning
messages: + msg905
2006-11-08 10:31:29 netmask set assignedto: tromer
2006-11-07 17:44:12 netmask set status: testing -> need-info
messages: + msg886
2006-11-06 23:28:30 tromer set status: chatting -> testing
messages: + msg881
2006-11-06 12:51:14 ittayd set nosy: roos, quintesse, tromer, mathieu.clabaut, jon, niemeyer, Gary, tomee, netmask, risto.kankkunen, ittayd
messages: + msg880
2006-11-06 10:14:57 tromer link issue127 superseder
2006-11-06 10:09:14 tromer link issue111 superseder
2006-11-06 10:04:45 tromer set files: + forkingsmart-4-combined.patch
nosy: + tromer
messages: + msg874
2006-11-06 10:00:37 tromer link issue237 superseder
2006-11-06 10:00:29 tromer link issue82 superseder
2006-11-06 10:00:21 tromer link issue122 superseder
2006-11-02 16:25:21 ittayd set nosy: + ittayd
messages: + msg872
2006-06-19 11:37:37 tomee set nosy: + tomee
messages: + msg572
2006-02-10 21:46:59 mathieu.clabaut set nosy: + mathieu.clabaut
messages: + msg353
2006-02-07 19:36:11 risto.kankkunen set messages: + msg339
2006-02-05 20:17:32 risto.kankkunen set status: chatting
nosy: niemeyer, roos, quintesse, netmask, Gary, risto.kankkunen, jon
messages: + msg336
2006-02-05 20:15:27 risto.kankkunen set priority: bug -> critical
status: resolved -> (no value)
messages: + msg335
nosy: + risto.kankkunen
2006-01-09 13:36:24 netmask set status: need-info -> resolved
messages: + msg216
2006-01-09 11:35:57 roos set messages: + msg215
2006-01-09 01:39:06 jon set nosy: roos, quintesse, jon, niemeyer, Gary, netmask
messages: + msg214
2006-01-04 05:22:55 Gary set files: + unnamed
messages: + msg207
2006-01-03 01:01:41 roos set messages: + msg204
2006-01-02 19:16:27 netmask set nosy: + netmask
messages: + msg202
2005-11-10 02:38:34 Gary set nosy: + Gary
messages: + msg106
2005-11-05 19:38:00 jon set nosy: + jon
messages: + msg101
2005-10-10 11:01:34 quintesse set nosy: + quintesse
messages: + msg47
2005-10-05 19:10:09 niemeyer set project: smart
status: unread -> need-info
messages: + msg33
nosy: + niemeyer
2005-09-30 09:25:14 roos create

[Rehan Khan (rasker)]

Imported from the old bugtracker : http://tracker.labix.org/issue12

This issue has not been reported since at least smart 0.50.

Reason for import: Continue monitoring to see if the issue persists.

[Gustavo Niemeyer (niemeyer)]

If this is seen again, please open a new bug. The original issue has been
partly dealt with, but taking a long time in edge cases may still be seen due
to the nature of the algorithm.