Re: [sage-devel] Re: Race condition when closing tickets
Then open a new ticket. Which is exactly what you would do if the ticket has been switched to closed, so why is it impossible to do when it is positive_review? On Thursday, April 16, 2015 at 7:58:32 AM UTC+2, Jeroen Demeyer wrote: > > On 2015-04-16 01:33, Nils Bruin wrote: > > The whole meaning of "positive review" goes out of > > the window if one still changes the branch on the ticket. > I disagree. It can easily happen that a problem is found after a ticket > was set to positive_review. Perhaps some corner case doesn't quite work. > Or there could be trivial changes like a documentation typo. > > -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Re: Race condition when closing tickets
On 2015-04-16 01:33, Nils Bruin wrote: The whole meaning of "positive review" goes out of the window if one still changes the branch on the ticket. I disagree. It can easily happen that a problem is found after a ticket was set to positive_review. Perhaps some corner case doesn't quite work. Or there could be trivial changes like a documentation typo. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Re: Race condition when closing tickets
On 2015-04-15 23:28, Daniel Krenn wrote: If there is a script for closing tickets, it could be extended to check if the status is still positive_review and if the current branch on the ticket was already (really) merged into develop... Just FYI: that's essentially what I did. My script kept a file with md5sums of the patches (remember those?) which were tested on the buildbot. When the time came to actually close tickets, the md5sums of the patches on the ticket were compared to the stored md5sums. A mismatch would be an error. Jeroen. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Re: Race condition when closing tickets
out of interest, I downloaded
http://trac.sagemath.org/query?status=closed&col=id&col=summary&col=status&col=milestone&col=resolution&col=changetime&col=commit&order=priority
as a .csv file and wrote an ugly script to see whether the commitid (if
present) is merged in 6.7.beta0. I then manually removed those tickets which
where closed yesterday, because that would be the tickets already closed for
6.7.beta1.
Here is the result
ticket commit changed
15017 872e2bc0225b6929ed8cc052a63850c892abe723 07/21/14 03:51:36
15660 d4842426356568815b80323f8fc6afbb80185ba3 01/25/15 14:55:21
15773 81a2ea983e2903b9cc5f2a78a57dd24fe32e48d5 08/31/14 19:45:31
16939 d9214c8c4ce4cbb5b29d207e8ed3bbf4e356b180 03/06/15 20:28:12
17221 2e62790eeec6463562120c204a86ebb323fa7cba 04/14/15 21:44:07
17307 7fa0de34aa713520d0fef2a7065581599dfd7bf3 11/09/14 20:40:11
14880 3b6a841dc5f6210a9c0d70b4d6cf24e34788c228 10/02/13 08:35:34
15599 08be4423f7703847d4da7c6d8cce4bb0902f93ed 03/13/14 04:38:46
16403 9f08c6b859a0fb853f69dbaedc73811740ee3e91 01/25/15 14:56:02
16847 e42ce551e66ff3f854c2424561bc71a075bac67a 08/19/14 12:38:59
16698 ba38202cf55caefb032e1b4519c96b9706577004 09/15/14 16:56:43
17702 daad37297eff79d7ee8c34989840d44f7e3e5105 02/08/15 17:31:25
#17221 (which started this thread) already has a successor ticket #18206; I did
not have a closer look at the others.
And for completeness, here is the ugly script (it assumes that develop is the
current branch):
import csv
import subprocess
with open('/tmp/query.csv', 'r') as csvfile:
tracreader = csv.reader(csvfile, delimiter=',', quotechar='"')
headings_row = next(tracreader)
headings_row[0] = 'id' # override UTF8 markers
for row in tracreader:
data = dict(zip(headings_row, row))
if data['commit'] and data['resolution'] == 'fixed':
try:
output = subprocess.check_output(["git", "branch",
"--contains", data['commit']],
stderr=subprocess.STDOUT).split('\n')
except subprocess.CalledProcessError:
print data['id'], data['commit'], data['changetime']
if "* develop" not in output:
print data['id'], data['commit'], data['changetime']
I am aware of the fact that this would be cleaner if implemented in git trac;
but for a first overview, this seemed to be the quicker approach.
Regards,
CH
Am 2015-04-16 um 05:50 schrieb Volker Braun:
> You are just renaming the step where nobody is supposed to make changes any
> more
> (from positive review to merging). It doesn't solve anything. Eventually a CI
> script will have to be written, and it'll automatically switch from positive
> review to merging.
>
> We could add a mandatory 2-week cooling off period while I'm not merging
> positively-reviewed tickets, if thats what you want. But I'd find that
> annoying.
--
You received this message because you are subscribed to the Google Groups
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to sage-devel+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at http://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Re: Race condition when closing tickets
You are just renaming the step where nobody is supposed to make changes any more (from positive review to merging). It doesn't solve anything. Eventually a CI script will have to be written, and it'll automatically switch from positive review to merging. We could add a mandatory 2-week cooling off period while I'm not merging positively-reviewed tickets, if thats what you want. But I'd find that annoying. On Thursday, April 16, 2015 at 1:33:51 AM UTC+2, Nils Bruin wrote: > > On Wednesday, April 15, 2015 at 2:37:02 PM UTC-7, Volker Braun wrote: >> >> Even then, the correct workflow is to not change tickets after setting >> them to positive review. Having the release manager chase after branches as >> people keep adding stuff after review is not a sustainable workflow. >> > > I completely concur. The whole meaning of "positive review" goes out of > the window if one still changes the branch on the ticket. Obviously, before > someone can change the branch on the ticket, they should set its status to > "needs work" (or at least some other status. Perhaps the act of changing > the branch should automatically remove "positive review" status? However, > even with doing that, there is still a race condition possible. I'm not > quite sure how likely this one is and hence whether we need to protect > against it. The consequences are just an unmerged further change, which > should not be a huge disaster: > > 0. Ticket is at "positive review" > 1. Release Manager (RM) starts testing/merging process > 2. Ticket gets set to "needs work" by author > 3. Branch gets updated by author > 4. Referee (sits next to author, perhaps) agrees with changes and set > ticket to "positive review" (he may have witnessed testing before step 2 > already) > 5. RM finishes testing/merging and if he/she checks at all, would find the > ticket at "positive review" as expected and now moves status to "closed", > but has only merged the old branch. > > The obvious solution is to change step 1 to: > 1'. RM sets status of ticket to "merging" and starts testing/merging > process (this change in status should be made by the same script that > starts the testing merging process, in which case it would not add to the > workload of the RM, only to a more elaborate ticket history) > > I don't think the extra status would even need to be protected so that the > RM can only change it to something else: the flag should be enough. > > > -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Re: Race condition when closing tickets
On Wednesday, April 15, 2015 at 2:37:02 PM UTC-7, Volker Braun wrote: > > Even then, the correct workflow is to not change tickets after setting > them to positive review. Having the release manager chase after branches as > people keep adding stuff after review is not a sustainable workflow. > I completely concur. The whole meaning of "positive review" goes out of the window if one still changes the branch on the ticket. Obviously, before someone can change the branch on the ticket, they should set its status to "needs work" (or at least some other status. Perhaps the act of changing the branch should automatically remove "positive review" status? However, even with doing that, there is still a race condition possible. I'm not quite sure how likely this one is and hence whether we need to protect against it. The consequences are just an unmerged further change, which should not be a huge disaster: 0. Ticket is at "positive review" 1. Release Manager (RM) starts testing/merging process 2. Ticket gets set to "needs work" by author 3. Branch gets updated by author 4. Referee (sits next to author, perhaps) agrees with changes and set ticket to "positive review" (he may have witnessed testing before step 2 already) 5. RM finishes testing/merging and if he/she checks at all, would find the ticket at "positive review" as expected and now moves status to "closed", but has only merged the old branch. The obvious solution is to change step 1 to: 1'. RM sets status of ticket to "merging" and starts testing/merging process (this change in status should be made by the same script that starts the testing merging process, in which case it would not add to the workload of the RM, only to a more elaborate ticket history) I don't think the extra status would even need to be protected so that the RM can only change it to something else: the flag should be enough. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Re: Status of IPython
On Wed, Apr 15, 2015 at 01:07:57PM +0200, Vincent Delecroix wrote: > >With the new display infrastructure we can support multiple UIs without > >much effort. The is no need to remove SageNB any time soon, though I expect > >it'll become less popular. > > It will not until: > - it is possible to import from rest document > - it supports live documentation > Both failed right? For the record: there is some ongoing work by Scott Sievert in this direction here at PyCon 15: https://github.com/ipython/ipython/issues/8319 https://github.com/scottsievert/rst-to-ipynb/tree/master/tests Cheers, Nicolas -- Nicolas M. Thiéry "Isil" http://Nicolas.Thiery.name/ -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Re: Race condition when closing tickets
On Wednesday, April 15, 2015 at 11:28:37 PM UTC+2, Daniel Krenn wrote: > > If there is a script for closing tickets, it could be extended to check > if the status is still positive_review and if the current branch on the > ticket was already (really) merged into develop... > Feel free to send a PR to https://github.com/sagemath/git-trac-command Even then, the correct workflow is to not change tickets after setting them to positive review. Having the release manager chase after branches as people keep adding stuff after review is not a sustainable workflow. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Re: Race condition when closing tickets
Am 2015-04-15 um 22:58 schrieb Simon King: > Would it be really so complicated? I suppose you have a script that > merges into develop and then tests the branches of positively reviewed > ticket (or are you doing it manually?). That script could also contain a > line that blocks the associated ticket (i.e., makes it impossible for a > non-administrator to change the ticket status) while the script is > running. If there is a script for closing tickets, it could be extended to check if the status is still positive_review and if the current branch on the ticket was already (really) merged into develop... Best wishes, Daniel -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
[sage-devel] Re: Race condition when closing tickets
Hi Volker, On 2015-04-15, Volker Braun wrote: >> Can you give a compelling reason why not? I mean, when the ticket branch >> hasn't been merged into the develop branch? > > > Testing imposes a significant delay between merging into the develop branch > and closing the ticket. Aha, you mean the problem is that the process of merging it into the develop branch takes a significant time, so that it is easily possible that the reviewer thinks "well, it hasn't been merged yet, so, we can still work on it"? OK, that's a good reason to discourage withdrawing positive reviews --- unless we have a way to freeze a ticket while the release manager is merging. > Sure we can have a more complicated trac workflow, > but one way or another it'll just boil down to not changing > positively-reviewed tickets unless the release manager finds an issue. Would it be really so complicated? I suppose you have a script that merges into develop and then tests the branches of positively reviewed ticket (or are you doing it manually?). That script could also contain a line that blocks the associated ticket (i.e., makes it impossible for a non-administrator to change the ticket status) while the script is running. Best regards, SImon -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] message to patchbot breeders
Hello, As I have said before, I have no machine other than my laptop, so I only test the patchbot on my own sage installation on that laptop. This means that I test it interactively inside an Ipython Session, and only in plugin_only mode, so that it will not run too long and destroy my sage installation (I hate ATLAS recompilation) This was working when I proposed the 2.3.3 version. But of course I missed that the compilation step would go astray, because I do not use sage -patchbot I am sorry if this causes some inconvenience. I try to do my best to enhance the patchbot, because I care that it works, but maybe I am not competent enough. In the lack of a more qualified person, nevertheless... Frederic Le mercredi 15 avril 2015 21:28:54 UTC+2, Jeroen Demeyer a écrit : > > On 2015-04-15 17:41, Frédéric Chapoton wrote: > > oops ! Thanks a lot. I was not sure which one of the two lines must > > really be first, so I choose the utf8 one.. > > Obvious question: do you actually try the patchbot yourself before > making a new public package? > > -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Re: trac errors out on some tickets
Ok, hopefully it is fixed now. It looks like it is a bug in git (at least the version that is on the trac server), see [1]. [1] http://git.661346.n2.nabble.com/Problems-with-stale-keep-files-on-git-server-td6226633.html On Wed, Apr 15, 2015 at 1:02 PM, R. Andrew Ohana wrote: > There is some permissions issue with the git repository. Working on it... > > On Wed, Apr 15, 2015 at 12:28 PM, Dima Pasechnik > wrote: > >> /tmp or some other filesystem ran out of space? >> >> >> On Wednesday, 15 April 2015 20:09:54 UTC+1, vdelecroix wrote: >>> >>> contamination... >>> >>> http://trac.sagemath.org/ticket/17818 >>> >> -- >> You received this message because you are subscribed to the Google Groups >> "sage-devel" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to sage-devel+unsubscr...@googlegroups.com. >> To post to this group, send email to sage-devel@googlegroups.com. >> Visit this group at http://groups.google.com/group/sage-devel. >> For more options, visit https://groups.google.com/d/optout. >> > > > > -- > Andrew > -- Andrew -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Re: trac errors out on some tickets
There is some permissions issue with the git repository. Working on it... On Wed, Apr 15, 2015 at 12:28 PM, Dima Pasechnik wrote: > /tmp or some other filesystem ran out of space? > > > On Wednesday, 15 April 2015 20:09:54 UTC+1, vdelecroix wrote: >> >> contamination... >> >> http://trac.sagemath.org/ticket/17818 >> > -- > You received this message because you are subscribed to the Google Groups > "sage-devel" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to sage-devel+unsubscr...@googlegroups.com. > To post to this group, send email to sage-devel@googlegroups.com. > Visit this group at http://groups.google.com/group/sage-devel. > For more options, visit https://groups.google.com/d/optout. > -- Andrew -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] message to patchbot breeders
On 2015-04-15 17:41, Frédéric Chapoton wrote: oops ! Thanks a lot. I was not sure which one of the two lines must really be first, so I choose the utf8 one.. Obvious question: do you actually try the patchbot yourself before making a new public package? -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Re: trac errors out on some tickets
/tmp or some other filesystem ran out of space? On Wednesday, 15 April 2015 20:09:54 UTC+1, vdelecroix wrote: > > contamination... > > http://trac.sagemath.org/ticket/17818 > -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Re: trac errors out on some tickets
contamination... http://trac.sagemath.org/ticket/17818 -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Re: trac errors out on some tickets
Just trying to access http://trac.sagemath.org/ticket/16477 gives Trac detected an internal error: OSError: [Errno 2] No such file or directory: '/tmp/tmpafaJGh' Samuel -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Re: trac errors out on some tickets
Just trying to access http://trac.sagemath.org/ticket/16477 gives Trac detected an internal error: OSError: [Errno 2] No such file or directory: '/tmp/tmpafaJGh' -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
[sage-devel] Big bug with the gamma function
Just to let people know about: http://trac.sagemath.org/ticket/18210 A plot of the gamma function which is crashing Sage... I reported the bug on trac, but the person who actually found it is Buck (in CC) who is working on Sage at PyCon. Best Viviane -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
[sage-devel] Re: subdomain for package downloads
2015-04-15 17:54:43 UTC+2, Volker Braun: > > How about packages.sagemath.org > +1 -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
[sage-devel] Re: Race condition when closing tickets
Either way, there needs to be a point where you can't go back and change the branch and expect it to be in the next release. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Re: trac errors out on some tickets
Got the very same error than Darij when trying to submit modifications to some ticket (#5332). I got OSError: [Errno 2] No such file or directory: '/tmp/tmpFn4RT0' On 15/04/15 19:10, Volker Braun wrote: For the record, here is the traceback: File "/usr/local/lib/python2.7/dist-packages/Trac-1.1.1-py2.7.egg/trac/web/main.py", line 497, in _dispatch_request dispatcher.dispatch(req) File "/usr/local/lib/python2.7/dist-packages/Trac-1.1.1-py2.7.egg/trac/web/main.py", line 233, in dispatch content_type) File "/usr/local/lib/python2.7/dist-packages/Trac-1.1.1-py2.7.egg/trac/web/chrome.py", line 991, in render_template stream |= self._filter_stream(req, method, filename, stream, data) File "/usr/local/lib/python2.7/dist-packages/genshi/core.py", line 133, in __or__ return Stream(_ensure(function(self)), serializer=self.serializer) File "/usr/local/lib/python2.7/dist-packages/Trac-1.1.1-py2.7.egg/trac/web/chrome.py", line 1170, in inner data) File "/usr/local/lib/python2.7/dist-packages/sage_trac-0.2-py2.7.egg/sage_trac/ticket_branch.py", line 65, in filter_stream ret = self.get_merge(branch) File "/usr/local/lib/python2.7/dist-packages/sage_trac-0.2-py2.7.egg/sage_trac/git_merger.py", line 25, in get_merge ret = self._merge(commit) File "/usr/local/lib/python2.7/dist-packages/sage_trac-0.2-py2.7.egg/sage_trac/git_merger.py", line 135, in _merge shutil.rmtree(tmpdir) File "/usr/lib/python2.7/shutil.py", line 237, in rmtree onerror(os.listdir, path, sys.exc_info()) File "/usr/lib/python2.7/shutil.py", line 235, in rmtree names = os.listdir(path) On Wednesday, April 15, 2015 at 7:02:15 PM UTC+2, Darij Grinberg wrote: Hi, trac is broken again... but only for some tickets: http://trac.sagemath.org/ticket/11529 http://trac.sagemath.org/ticket/16820 http://trac.sagemath.org/ticket/18001 http://trac.sagemath.org/ticket/18000 It's always this kind of bug: Trac detected an internal error: OSError: [Errno 2] No such file or directory: '/tmp/tmptnACR8' This is now http://trac.sagemath.org/ticket/18209 (works so far!), but I don't know whom to CC. Best regards, Darij -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] __contains__ in Matrix
On Wed, 15 Apr 2015, Vincent Delecroix wrote: I do not like it since a matrix can also be thought as a list of vectors. True. OK with this. sage: m = matrix(3, range(9)) sage: list(m) [(0, 1, 2), (3, 4, 5), (6, 7, 8)] I didn't know this. m.list() gives list of elements. But (just checked) that is documented on .list() so everything is fine. -- Jori Mäntysalo
Re: [sage-devel] __contains__ in Matrix
I do not like it since a matrix can also be thought as a list of
vectors. Which is actually what you get when doing
{{{
sage: m = matrix(3, range(9))
sage: list(m)
[(0, 1, 2), (3, 4, 5), (6, 7, 8)]
}}}
Most of the time, `a in b` should be equivalent to
any(x == b for x in a)
And for matrix `a in m` would be an answer to "is a a row of m?".
Vincent
On 15/04/15 19:26, Jori Mantysalo wrote:
Would it break something if matrix class would have __contains__() so
that one could say for example "42 in M" instead of "42 in M.list()"?
--
You received this message because you are subscribed to the Google Groups
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to sage-devel+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at http://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.
[sage-devel] Re: Race condition when closing tickets
> > > > But you can't turn around and add another feature on a ticket that >> already >> > has positive review. >> > Not another feature, sure. But fixing an error, even a typo (especially if it's a misleading typo like forgetting the word "not") sounds like adequate grounds to revert a positive review. Our testing process should accommodate quality, not vice versa. > Testing imposes a significant delay between merging into the develop > branch and closing the ticket. > My understanding of git is of course weaker than anyone else's, but ... isn't the whole POINT of having a DVCS so that one can test a change and do whatever one wants before actually merging into the develop branch? Again, I would argue that it is far less likely for someone to respond to even a fairly major bug in a "new" ticket than to do a quick fix so that their contribution "gets in" with a ticket that is so so so close to being merged. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
[sage-devel] __contains__ in Matrix
Would it break something if matrix class would have __contains__() so that one could say for example "42 in M" instead of "42 in M.list()"? -- Jori Mäntysalo
[sage-devel] Re: trac errors out on some tickets
For the record, here is the traceback: File "/usr/local/lib/python2.7/dist-packages/Trac-1.1.1-py2.7.egg/trac/web/main.py", line 497, in _dispatch_request dispatcher.dispatch(req) File "/usr/local/lib/python2.7/dist-packages/Trac-1.1.1-py2.7.egg/trac/web/main.py", line 233, in dispatch content_type) File "/usr/local/lib/python2.7/dist-packages/Trac-1.1.1-py2.7.egg/trac/web/chrome.py", line 991, in render_template stream |= self._filter_stream(req, method, filename, stream, data) File "/usr/local/lib/python2.7/dist-packages/genshi/core.py", line 133, in __or__ return Stream(_ensure(function(self)), serializer=self.serializer) File "/usr/local/lib/python2.7/dist-packages/Trac-1.1.1-py2.7.egg/trac/web/chrome.py", line 1170, in inner data) File "/usr/local/lib/python2.7/dist-packages/sage_trac-0.2-py2.7.egg/sage_trac/ticket_branch.py", line 65, in filter_stream ret = self.get_merge(branch) File "/usr/local/lib/python2.7/dist-packages/sage_trac-0.2-py2.7.egg/sage_trac/git_merger.py", line 25, in get_merge ret = self._merge(commit) File "/usr/local/lib/python2.7/dist-packages/sage_trac-0.2-py2.7.egg/sage_trac/git_merger.py", line 135, in _merge shutil.rmtree(tmpdir) File "/usr/lib/python2.7/shutil.py", line 237, in rmtree onerror(os.listdir, path, sys.exc_info()) File "/usr/lib/python2.7/shutil.py", line 235, in rmtree names = os.listdir(path) On Wednesday, April 15, 2015 at 7:02:15 PM UTC+2, Darij Grinberg wrote: > > Hi, > > trac is broken again... but only for some tickets: > http://trac.sagemath.org/ticket/11529 > http://trac.sagemath.org/ticket/16820 > http://trac.sagemath.org/ticket/18001 > http://trac.sagemath.org/ticket/18000 > > It's always this kind of bug: > Trac detected an internal error: > > OSError: [Errno 2] No such file or directory: '/tmp/tmptnACR8' > > This is now http://trac.sagemath.org/ticket/18209 (works so far!), but I > don't know whom to CC. > > Best regards, > Darij > -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
[sage-devel] trac errors out on some tickets
Hi, trac is broken again... but only for some tickets: http://trac.sagemath.org/ticket/11529 http://trac.sagemath.org/ticket/16820 http://trac.sagemath.org/ticket/18001 http://trac.sagemath.org/ticket/18000 It's always this kind of bug: Trac detected an internal error: OSError: [Errno 2] No such file or directory: '/tmp/tmptnACR8' This is now http://trac.sagemath.org/ticket/18209 (works so far!), but I don't know whom to CC. Best regards, Darij -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
[sage-devel] Re: subdomain for package downloads
> > How about packages.sagemath.org > >> Yep, that makes sense. Nathann -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
[sage-devel] Re: Race condition when closing tickets
On Wednesday, April 15, 2015 at 5:44:04 PM UTC+2, Simon King wrote: > > > Its of course fine for the release manager to set something back from > > positive review. > Quod licet Iovi, non licet bovi... > Tautologically true as long as we are talking about a race specifically about Jovi not noticing something in time. > But you can't turn around and add another feature on a ticket that > already > > has positive review. > Can you give a compelling reason why not? I mean, when the ticket branch > hasn't been merged into the develop branch? Testing imposes a significant delay between merging into the develop branch and closing the ticket. Sure we can have a more complicated trac workflow, but one way or another it'll just boil down to not changing positively-reviewed tickets unless the release manager finds an issue. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
[sage-devel] subdomain for package downloads
Harald says: > I think we should introduce a new subdomain for this, to > have more flexibility in the future (e.g. spkgs.sagemath.org, > download.sagemath.org ... ) Any suggestions? -- William (http://wstein.org) -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
[sage-devel] Re: subdomain for package downloads
How about packages.sagemath.org On Wednesday, April 15, 2015 at 5:47:47 PM UTC+2, William wrote: > > Harald says: > > I think we should introduce a new subdomain for this, to > > have more flexibility in the future (e.g. spkgs.sagemath.org, > > download.sagemath.org ... ) > > Any suggestions? > > -- > William (http://wstein.org) > -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
[sage-devel] Re: Race condition when closing tickets
Hi! On 2015-04-15, Volker Braun wrote: > Its of course fine for the release manager to set something back from > positive review. Quod licet Iovi, non licet bovi... > Also, if there is an actual problem with the ticket then > tests should fail Why? The reviewer could notice (too late) that the code introduces a problem for which there is no test, yet. > But you can't turn around and add another feature on a ticket that already > has positive review. Can you give a compelling reason why not? I mean, when the ticket branch hasn't been merged into the develop branch? Of course, there is a race condition: The scenario is that a reviewer attempts to set the review back from "positive" to "needs work", and at the same time the release manager attempts to merge the ticket branch into the develop branch. It should be possible to write a trac plugin (sorry, I can't do it, but I am sure you can) such that * the release manager's merge attempt implies that all attempts to change the status of the ticket are blocked (with a nice message telling to open a new ticket), and * the release manager's merge attempt will automatically be aborted, if the reviewer sets it back to "needs work" a second before the release manager pushed the button. Best regards, Simon -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] message to patchbot breeders
oops ! Thanks a lot. I was not sure which one of the two lines must really be first, so I choose the utf8 one.. I have now remade patchbot-2.3.3 and put the new file ready for upload. Le mercredi 15 avril 2015 17:18:07 UTC+2, Jeroen Demeyer a écrit : > > On 2015-04-15 17:15, Frédéric Chapoton wrote: > > Damn ! Thanks for the feedback. But what can it mean ? > > You messed up the #! line in patchbot.py, it should be the *first* line! > -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] message to patchbot breeders
On 2015-04-15 17:15, Frédéric Chapoton wrote: Damn ! Thanks for the feedback. But what can it mean ? You messed up the #! line in patchbot.py, it should be the *first* line! -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
[sage-devel] Re: Race condition when closing tickets
Its of course fine for the release manager to set something back from positive review. Also, if there is an actual problem with the ticket then tests should fail (if they don't then thats a bug in itself), and you can also switch it back from positive review before I get to the ticket. But you can't turn around and add another feature on a ticket that already has positive review. On Wednesday, April 15, 2015 at 5:05:40 PM UTC+2, kcrisman wrote: > > > > The obvious answer: Don't change a ticket once it is positive review. >> Treat it like "closed" -- if you find a bug open a new ticket. >> >>> >>> > Hmm, but somehow I feel that "socially" that will not always work, unless > such things automatically became blockers. Positive review can often be > wrong - and indeed you yourself will usually unset it to 'needs work' when > the bots tell you there is a test failure, right? > -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] message to patchbot breeders
Damn ! Thanks for the feedback. But what can it mean ?
Le mercredi 15 avril 2015 17:03:47 UTC+2, Jeroen Demeyer a écrit :
>
> Doesn't work:
>
> import import: Unable to open XServer ().
> import import: Unable to open XServer ().
> import import: Unable to open XServer ().
> import import: Unable to open XServer ().
> import import: Unable to open XServer ().
> import import: Unable to open XServer ().
> import import: Unable to open XServer ().
> import import: Unable to open XServer ().
> import import: Unable to open XServer ().
> import import: Unable to open XServer ().
> import import: Unable to open XServer ().
> import import: Unable to open XServer ().
> import import: Unable to open XServer ().
> import import: Unable to open XServer ().
> import import: Unable to open XServer ().
> import import: Unable to open XServer ().
> import import: Unable to open XServer ().
> import import: Unable to open XServer ().
> import import: Unable to open XServer ().
> from: can't read /var/mail/optparse
> from: can't read /var/mail/http_post_file
> from: can't read /var/mail/trac
> /home/patchbot/sage-patchbot/local/bin/patchbot/patchbot.py: line 44:
> syntax error near unexpected token `('
> /home/patchbot/sage-patchbot/local/bin/patchbot/patchbot.py: line 44:
> `from util import (now_str as datetime, prune_pending, do_or_die,'
>
>
--
You received this message because you are subscribed to the Google Groups
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to sage-devel+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at http://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.
[sage-devel] Re: Race condition when closing tickets
The obvious answer: Don't change a ticket once it is positive review. Treat > it like "closed" -- if you find a bug open a new ticket. > >> >> Hmm, but somehow I feel that "socially" that will not always work, unless such things automatically became blockers. Positive review can often be wrong - and indeed you yourself will usually unset it to 'needs work' when the bots tell you there is a test failure, right? -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] message to patchbot breeders
Doesn't work:
import import: Unable to open XServer ().
import import: Unable to open XServer ().
import import: Unable to open XServer ().
import import: Unable to open XServer ().
import import: Unable to open XServer ().
import import: Unable to open XServer ().
import import: Unable to open XServer ().
import import: Unable to open XServer ().
import import: Unable to open XServer ().
import import: Unable to open XServer ().
import import: Unable to open XServer ().
import import: Unable to open XServer ().
import import: Unable to open XServer ().
import import: Unable to open XServer ().
import import: Unable to open XServer ().
import import: Unable to open XServer ().
import import: Unable to open XServer ().
import import: Unable to open XServer ().
import import: Unable to open XServer ().
from: can't read /var/mail/optparse
from: can't read /var/mail/http_post_file
from: can't read /var/mail/trac
/home/patchbot/sage-patchbot/local/bin/patchbot/patchbot.py: line 44:
syntax error near unexpected token `('
/home/patchbot/sage-patchbot/local/bin/patchbot/patchbot.py: line 44:
`from util import (now_str as datetime, prune_pending, do_or_die,'
--
You received this message because you are subscribed to the Google Groups
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to sage-devel+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at http://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.
[sage-devel] message to patchbot breeders
Hello, Ladies and Gentlemen, if your patchbot misbehaves (I am thinking to sage4) or runs a version older than 2.2, then *please consider upgrading to the tamer patchbot 2.3.3* sage -i http://chapoton.perso.math.cnrs.fr/patchbot-2.3.3.spkg By default it only tests safe tickets and uses the develop branch as a base. Robert, could you please upgrade the server ? Cheers, Frederic -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
[sage-devel] Re: selection of translation
On Wednesday, 15 April 2015 11:48:37 UTC+1, François wrote: > > Now that 6.6 is released I finally pushed a feature > in sage-on-gentoo that was requested. Namely only > build an install selected translations. > > Is there an interest to adopt such code in sage > itself? > everything that speeds up documentation building is great! > > François -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
[sage-devel] Re: Sage logo -- Proposal
Some time ago a question about the Sage logo was asked and answered on Ask-Sage. http://ask.sagemath.org/question/7667/geometric-name-of-sage-logo-existence-of-dual/ -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Re: Memory footprint
On Wed, 15 Apr 2015, Volker Braun wrote: A 4GB RAM Stick sets you back about EUR 25. I know, if I have a home-PC on the floor of my room. We have few servers with ECC memory, backups run regularly etc. And Sage as a one small application to run there. If you can't afford that, move the Sage instances to the student computers, presumably they have >= 100 MB RAM. Won't help if they want to share worksheet to teacher (or to some other student). Or use virtual machines on the server, KVM + KSM. Thanks. Going to check if KSM helps. -- Jori Mäntysalo
[sage-devel] Re: Memory footprint
A 4GB RAM Stick sets you back about EUR 25. If you can't afford that, move the Sage instances to the student computers, presumably they have >= 100 MB RAM. Or use virtual machines on the server, KVM + KSM. On Wednesday, April 15, 2015 at 1:14:54 PM UTC+2, Jori Mantysalo wrote: > > What Sage actually loads to memory at startup? It seems that every > instance of Sage 6.6.rc3 (from CUI, no GUI at all) eats almost 100 MB. > > This means that for 30 students we need 3 GB of memory -- for the two > hours they are having exercises at a class. That would be no problem with > some memory ballooning system at virtual servers -- but that we don't > have. Now it is annoying. > > And to make this worse, SageNB seems to left ssh connections (and so, > processes) open even after the studen has stopped worksheets and logged > out. > > -- > Jori Mäntysalo > -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
[sage-devel] Memory footprint
What Sage actually loads to memory at startup? It seems that every instance of Sage 6.6.rc3 (from CUI, no GUI at all) eats almost 100 MB. This means that for 30 students we need 3 GB of memory -- for the two hours they are having exercises at a class. That would be no problem with some memory ballooning system at virtual servers -- but that we don't have. Now it is annoying. And to make this worse, SageNB seems to left ssh connections (and so, processes) open even after the studen has stopped worksheets and logged out. -- Jori Mäntysalo
Re: [sage-devel] Re: Status of IPython
On 15/04/15 13:05, Volker Braun wrote: On Wednesday, April 15, 2015 at 12:55:56 PM UTC+2, Jori Mantysalo wrote: 1) What is the status of multiuser-ipython now? Will it replace sagenb as a GUI for Sage? With the new display infrastructure we can support multiple UIs without much effort. The is no need to remove SageNB any time soon, though I expect it'll become less popular. It will not until: - it is possible to import from rest document - it supports live documentation Both failed right? Vincent -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
[sage-devel] Re: Status of IPython
On Wednesday, April 15, 2015 at 12:55:56 PM UTC+2, Jori Mantysalo wrote: > > 1) What is the status of multiuser-ipython now? Will it replace sagenb as > a GUI for Sage? > With the new display infrastructure we can support multiple UIs without much effort. The is no need to remove SageNB any time soon, though I expect it'll become less popular. 2) How much memory it will eat per user? > Pretty much N * memory for a single user IPython notebook (sage --notebook=ipython) -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
[sage-devel] Status of IPython
1) What is the status of multiuser-ipython now? Will it replace sagenb as a GUI for Sage? 2) How much memory it will eat per user? -- Jori Mäntysalo
[sage-devel] selection of translation
Now that 6.6 is released I finally pushed a feature in sage-on-gentoo that was requested. Namely only build an install selected translations. Is there an interest to adopt such code in sage itself? François -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
