Re: [sage-devel] Busted maxima interface - anyone got a clue?
On 09/ 1/10 02:19 AM, Alex Ghitza wrote: On Wed, 01 Sep 2010 00:07:43 +0100, "Dr. David Kirkby" wrote: It's nice to know that running the tests 100 times and reporting the failures managed to unearth a bug. I'm not sure I would call this a bug, but it is definitely something that's very easy to overlook, and hard to detect without running the tests a large number of times. Best, Alex Oh well, I'm disappointed then! But in a way, the fact this test did call Maxima by mistake, means it generated two failures, which gives us more chance of finding the underlying cause for the busted maxima interface. http://trac.sagemath.org/sage_trac/ticket/8772 From what it said on the trac ticket, it appears only to happen on parallel testing. Though I don't know if anyone has run enough serial tests to be sure of this. I see the bug in 4% of cases in parallel parallel testing 100 times. I did start testing 100 serial cases, in different directories, but I messed up and run out of swap space. I'll give that another go later. Dave -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
Re: [sage-devel] Busted maxima interface - anyone got a clue?
On Tue, 31 Aug 2010 16:11:50 -0700, Mike Hansen wrote: > I did that this morning, although apparently it was only enough to get > the server back up and working -- not enough to upload files. I > removed some old logs and you should be able to upload patches again. > I'll look into moving it to a spot with more room once I have some > time. OK, it now worked for me. Best, Alex -- Alex Ghitza -- http://aghitza.org/ Lecturer in Mathematics -- The University of Melbourne -- Australia -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
Re: [sage-devel] Busted maxima interface - anyone got a clue?
On Wed, 01 Sep 2010 00:07:43 +0100, "Dr. David Kirkby" wrote: > It's nice to know that running the tests 100 times and reporting the failures > managed to unearth a bug. I'm not sure I would call this a bug, but it is definitely something that's very easy to overlook, and hard to detect without running the tests a large number of times. Best, Alex -- Alex Ghitza -- http://aghitza.org/ Lecturer in Mathematics -- The University of Melbourne -- Australia -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
Re: [sage-devel] Busted maxima interface - anyone got a clue?
Hi Mike, On Tue, 31 Aug 2010 16:11:50 -0700, Mike Hansen wrote: > I did that this morning, although apparently it was only enough to get > the server back up and working -- not enough to upload files. I > removed some old logs and you should be able to upload patches again. > I'll look into moving it to a spot with more room once I have some > time. Thanks for looking into this. I have tried again, and failed again with the same error message: Trac detected an internal error: OSError: [Errno 28] No space left on device: '/var/trac/sage_trac/attachments/ticket/9843/trac_9843.patch' This particular patch is not very time-sensitive so I can wait :) Best, Alex -- Alex Ghitza -- http://aghitza.org/ Lecturer in Mathematics -- The University of Melbourne -- Australia -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
Re: [sage-devel] Busted maxima interface - anyone got a clue?
On Tue, Aug 31, 2010 at 4:07 PM, Dr. David Kirkby wrote: > As you say, trac is down - it looks like /var is full on whatever machine > runs the server. /var is often used for log files. Perhaps someone with root > access on whatever machine it is could clean up /var. I did that this morning, although apparently it was only enough to get the server back up and working -- not enough to upload files. I removed some old logs and you should be able to upload patches again. I'll look into moving it to a spot with more room once I have some time. --Mike -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
Re: [sage-devel] Busted maxima interface - anyone got a clue?
On 08/31/10 05:56 PM, Dr. David Kirkby wrote: As some of you know, I run the long doctests 100 times on my Sun Ultra 27 on sage 4.5.3.alpha0. I should have said it was on 4.5.3.alpha2. The last of the 32-bit Solaris x86 / OpenSolaris fixes were merged in 4.5.3.alpha2, so that's the first "release" of sage that passes all doc tests - some of the time at least! Dave -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
Re: [sage-devel] Busted maxima interface - anyone got a clue?
On 08/31/10 11:13 PM, Alex Ghitza wrote: On Tue, 31 Aug 2010 17:56:35 +0100, "Dr. David Kirkby" wrote: Looking at the results from 100 runs, there was one or more failures on 11 of 100 times. Of the 11 runs where one or more doctests failed, 4 of them devel/sage-main/sage/modular/overconvergent/weightspace.py (twice) devel/sage/sage/tests/benchmark.py (once) devel/sage/sage/calculus/desolvers.py (once) were caused by this bug http://trac.sagemath.org/sage_trac/ticket/8772 discussed at http://groups.google.com/group/sage-devel/browse_thread/thread/3b43147e44324c25 It is marked as "critical", and has been open 4 months, but it seems to have got no attention at all. Perhaps someone who knows a bit about the pexpect/maxima interface might take a look. Unfortunately, I'm not one that knows enough about the Maxima interface to fix this. However, I do know that modular/overconvergent/weightspace.py has no reason to use Maxima. As soon as the trac server is working again, I'll have a patch for this small issue up at http://trac.sagemath.org/sage_trac/ticket/9843 Best, Alex It's nice to know that running the tests 100 times and reporting the failures managed to unearth a bug. As you say, trac is down - it looks like /var is full on whatever machine runs the server. /var is often used for log files. Perhaps someone with root access on whatever machine it is could clean up /var. Dave -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
Re: [sage-devel] Busted maxima interface - anyone got a clue?
On Tue, 31 Aug 2010 17:56:35 +0100, "Dr. David Kirkby" wrote: > Looking at the results from 100 runs, there was one or more failures on 11 of > 100 times. Of the 11 runs where one or more doctests failed, 4 of them > > > devel/sage-main/sage/modular/overconvergent/weightspace.py (twice) > devel/sage/sage/tests/benchmark.py (once) > devel/sage/sage/calculus/desolvers.py (once) > > were caused by this bug > > http://trac.sagemath.org/sage_trac/ticket/8772 > > discussed at > > http://groups.google.com/group/sage-devel/browse_thread/thread/3b43147e44324c25 > > It is marked as "critical", and has been open 4 months, but it seems to have > got > no attention at all. > > Perhaps someone who knows a bit about the pexpect/maxima interface might take > a > look. Unfortunately, I'm not one that knows enough about the Maxima interface to fix this. However, I do know that modular/overconvergent/weightspace.py has no reason to use Maxima. As soon as the trac server is working again, I'll have a patch for this small issue up at http://trac.sagemath.org/sage_trac/ticket/9843 Best, Alex -- Alex Ghitza -- http://aghitza.org/ Lecturer in Mathematics -- The University of Melbourne -- Australia -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
[sage-devel] Busted maxima interface - anyone got a clue?
As some of you know, I run the long doctests 100 times on my Sun Ultra 27 on sage 4.5.3.alpha0. Looking at the results from 100 runs, there was one or more failures on 11 of 100 times. Of the 11 runs where one or more doctests failed, 4 of them devel/sage-main/sage/modular/overconvergent/weightspace.py (twice) devel/sage/sage/tests/benchmark.py (once) devel/sage/sage/calculus/desolvers.py (once) were caused by this bug http://trac.sagemath.org/sage_trac/ticket/8772 discussed at http://groups.google.com/group/sage-devel/browse_thread/thread/3b43147e44324c25 It is marked as "critical", and has been open 4 months, but it seems to have got no attention at all. Perhaps someone who knows a bit about the pexpect/maxima interface might take a look. Dave -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org