Perl 6 - the future is here, just unevenly distributed

IRC log for #webwork, 2012-09-07

| Channels | #webwork index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:18 Paul_Pearson joined #webwork
00:35 aubreyja joined #webwork
01:04 mgage_ I'd call ww2.5.2 a beta release at best right now.  That's what it's not in openwebwork/webwork2 just yet. I am trying to get ww2.5.1.1 into openwebwork  it's about ready for release I think.  Probably this weekend because I'll have some time to handle issues if they come up.
01:06 mgage_ ionparticle:  would having the ownership of logs/debug.log reset when you visit the admin course solve the problem you are seeing?  I've just handled by hand up until now.
01:20 ionparticle the problem is that $WeBWorK::Constants::WEBWORK_DIRECTORY is set by one of the apache configuration files
01:20 ionparticle if you run something like bin/addcourse
01:20 ionparticle it doesn't go through the Apache stack, so that variable doesn't get set
01:21 ionparticle and it complains every time
01:23 mgage_ actually it happens using things from the web as well (although the errors just accumulate in the log) -- we've inserted some "testing" code in the admin course so that certain aspects of the database and the directory structure get checked when ever you visit the admin course -- that way the code isn't run too often but it is semi automatic as long as you remember to visit the admin "course" before you release the site into the wild..
01:23 mgage_ that help handle your use case?  or should we look for something more robust?
01:25 mgage_ actually that variable needs to be set in your environment  WEBWORK_ROOT if you are running scripts from the command line.  the script needs some method for determining where the .conf files live
01:25 mgage_ but the ownership of the log file is a separate problem.
01:25 ionparticle I think that might be a different error from this one
01:26 ionparticle this one doesn't show up ever if I'm visiting the site from the browser
01:26 ionparticle ownership isn't a problem
01:26 mgage_ ok -- I've reread your original post more carefully.  -- you are write it's a separate problem.
01:26 ionparticle since the directories and the http user is all the same
01:27 mgage_ yeah -- but that log file gets set when the http process starts up and the ownership is still root :-)
01:28 ionparticle nah, I'm using the dev setup
01:28 ionparticle so I start the http process with bin/wwapa
01:28 ionparticle so I start the http process with bin/wwapach2ctl
01:28 mgage_ however for your problem many scripts require that WEBWORK_ROOT be set which has the same content as the WEBWOR_DIRECTORY variable
01:29 mgage_ addcourse is kind of an old script so it might not have had that safety check built in to it.
01:29 ionparticle WEBWORK_ROOT is set
01:29 ionparticle and it is checked in addcourse
01:29 mgage_ so can you tell where the error gets thrown?
01:30 ionparticle eval "use lib '$webwork_dir/lib'"; die $@ if $@;
01:31 ionparticle it's just a warning though, so execution continue
01:32 mgage_ sure it's not the use CourseEnvironment evaluation?  That would make more sense.  I don't see how the first eval could inquire about WEBWORK_DIRECTORY
01:32 ionparticle oh, you're right
01:32 mgage_ I'll be you could fix it by adding eval "use WeBWorK::Constants"  to the script
01:32 ionparticle ok, let me try
01:33 mgage_ be _. bet
01:34 ionparticle still there
01:34 mgage_ ok -- guess it really has to be looked at then
01:34 ionparticle no worries
01:34 ionparticle it's only a problem if debug is on
01:35 ionparticle and if you don't want to hard code the directory to the debug file
01:35 mgage_ over time there were about three different methods for finding the home directory -- I've combined some of them, but I guess there are still some left out there.  Unfortunately you have to do it in away that doesn't cause to much backward incompatibility
01:39 mgage_ guess two   setting  $webwork::Constants::WEBWORK_DIRECTORY = $webwork_dir;
01:39 mgage_ in the script might solve this
01:39 ionparticle yeah, it does, at least something similar
01:40 ionparticle but it seemed a bit weird to me, so thought I'd ask
01:40 ionparticle $WeBWorK::Constants::WEBWORK_DIRECTORY = $ENV{WEBWORK_ROOT};
01:40 ionparticle before the eval "use..." clauses fixes it
01:40 mgage_ don't know if that is the best solution however   -- if you look at some of the WebworkWebservice files and their initialization you'll see some struggle with the same problem.
01:42 mgage_ no -- I think that's the best solution for now -- but maybe adjusting the way that constant is set in Constants.pm might avoid having to find kludge fixes in other files.  The basic problem is that there is no one entry point to these files  -- using the apache initialization files is the closest we've been able to come.
01:43 ionparticle ok
01:48 ionparticle only fix I can think of in Constants.pm
01:49 ionparticle is using the location of Constants.pm itself as an anchor point
01:49 ionparticle since presumably, Constants.pm always has to be in webwork2/lib/WeBWorK
01:50 ionparticle so if we take the value of __FILE__ in Constants.pm
01:50 mgage_ that was not an assumption we wanted to make originally
01:50 ionparticle and remove lib/WeBWorK/Constants.pm
01:50 ionparticle then it's webwork root
01:50 ionparticle lol
01:50 ionparticle ah
01:50 mgage_ yes or use ../../
01:50 ionparticle that too
01:50 aubreyja hola
01:51 mgage_ hi jason
01:51 ionparticle hi
01:51 aubreyja ionparticle: i saw your note about blackboard - there's also this: http://webwork.maa.org/viewvc/bbplugin/trunk/BuildingBlock_Missouri/WebWrks2_2/
01:52 mgage_ I was wondering about setting it to $ENV{{WEBWORK_ROOT} if it wasn't already defined.
01:52 ionparticle yep, we've tried that
01:52 ionparticle doesn't work with our Blackboard version
01:52 aubreyja it doesn't work with ours either :), but there's actually progress being made now
01:53 ionparticle lol
01:53 mgage_ it would have the advantage of handling grades as well as single sign-in   ---- once it works
01:55 aubreyja bill's lit stuff might to that too -I'm sure it does sso, but maybe it does grade book communication too?  I've been wanting to try it out
01:55 aubreyja "Bill's LTI stuff might do that too"
01:55 ionparticle we're actually banking on LTI here too
01:56 ionparticle the hope is that if we ditch Blackboard later on, wouldn't have to do extra development
01:56 mgage_ they hadn't gotten the gradebook to work yet with LTI and Canvas at Utah (as of Mathfest) -- but it sounded like that was Canvas's issue
01:57 aubreyja I admire the preparation for ditching blackboard.  You know - just in case :)
01:57 ionparticle lol
01:57 mgage_ I'd like to have someone work on LTI and  moodle/webwork -- that should be relatively easy test case
01:57 aubreyja good point
01:58 mgage_ it would be interesting to see if it works better than our current bridge1 -- or if there are things that are to be desired
01:58 aubreyja actually, there is a group here called MORE-Net that is interested in possibly using moodle+LTI+webwork
01:59 aubreyja they use moodle and do a lot for high schools, and are interested no in the fact that I'm hosting a lot of HS's
01:59 mgage_ so we already have bridge1 as a stop-gap measure and as a model for doing the LTI interface.
02:00 ionparticle does it support any of the BasicLTI extension protocols?
02:00 ionparticle is bridge1 in the webworkdev github?
02:01 aubreyja not sure - I know very little about LTI - but Bill Wheeler would know
02:01 aubreyja the code, by the way, could use a second set of eyes, so if you guys are interested in it feel free to make improvements
02:02 ionparticle I'm interested, but can't spare the time
02:02 mgage_ bridge1 webservice has been in WeBWorK for 4 or 5 years now.        the moodle mod is in https://github.com/openwebwork/wwmoodle
02:03 mgage_ use wwassignment5 for moodle2 and wwassignment4 for moodle1.9.x
02:03 ionparticle we're using a custom LTI solution in Webwork for course creation/enrolment, and sso
02:04 mgage_ you have to enable the webservice (SOAP) in the webwork.apache2-config
02:05 aubreyja how custom is the course creation and enrollment stuff? We could use something like that
02:05 aubreyja BTW Mie, I talked to Shivram - he has a very interesting project
02:05 ionparticle bridge1 is webservices?
02:07 mgage_ bridge1 uses webservices -- but it connects webwork to moodle as an assignment type. bridge 2 (about 80% operational) would connect webwork to the moodle/quiz mod so that it grades math homework problems but the presentation is done entirely within moodle. -- if we finish that you could use moodle without ever realizing that webwork is running behind the scenes.
02:07 ionparticle ahhh
02:07 ionparticle cool
02:07 mgage_ it works a little like STACK if you are familiar with that.
02:07 ionparticle no clue
02:08 ionparticle lol
02:08 ionparticle https://github.com/ubc/webwork2
02:08 ionparticle our version of webwork
02:08 mgage_ you can look at https://math.webwork.rochester.edu/moodle/ for an example -- you can get in as guest for now
02:09 ionparticle lib/WebworkBridge is where most of the custom code live
02:09 mgage_ cool  -- ty
02:10 ionparticle we use the B"asic LTI tools" Building Block to talk to it
02:10 ionparticle "Basic LTI tools"
02:10 ionparticle there's a similar one for Moodle
02:10 ionparticle similar plugin*
02:11 mgage_ do you have a reference for the docs for LTI tools Building Block?
02:11 mgage_ that would be a useful reference
02:11 ionparticle unfortunately, it's very tailored to our data set
02:11 ionparticle so it's pretty custom
02:12 aubreyja The blackboard basic LTI tools building block?
02:12 aubreyja I think it's standard issue now, right?
02:13 ionparticle the standard one doesn't support the extension protocols
02:13 ionparticle this one is one of the OSCELOT projects
02:13 ionparticle trying to find the link
02:14 ionparticle http://www.spvsoftwareproducts.com/bb/basiclti/
02:14 mgage_ ty
02:16 aubreyja Mike, btw, you missed our call today!
02:17 mgage_ yes -- I was in class :-)  -- sent a note to vicki but I guess she missed it too
02:17 mgage_ I'm bridging to CS this semester which means this week at least I'm taking 4 classes on Tue and Thurs :-)
02:17 aubreyja Nothing much happened, but Davide joined us for the first time
02:18 aubreyja sheesh
02:18 mgage_ I'll be more sympathetic to the bewildered looks on students during the first week of class :-0
02:19 aubreyja anything interesting?
02:19 mgage_ firday is less cluttered so I'll be spending some time on some ww projects -- perhaps we can get caught up then?
02:19 mgage_ well -- the data mining course and the software engineering course have the most potential for applications to webwork.
02:20 aubreyja cool
02:20 mgage_ I'm also sitting in on an algorithms course which is just fun (but basically math) and a complier/languages course
02:21 mgage_ in most of these cases I know between 1/3 and 2/3 of the material but since I'm self taught and never studied this stuff systematically I have some big holes and I'm not always aware of what they are
02:21 aubreyja ah, yeah, I could see myself really digging those and forgetting the practical courses :)
02:21 mgage_ it's been fun hanging around the CS dept all day
02:22 aubreyja I had that experience sitting in on a statistics course once
02:22 mgage_ anyway -- it's different and interesting at least so far.
02:22 aubreyja sounds great
02:23 aubreyja here comes obama… :)
02:23 mgage_ will you have anytime tomorrow -- perhaps for facetime?  we should catch up on top level plans for Fitchburg  and some other stuff.
02:23 aubreyja Sure - I teach until 12, but after that I can meet
02:23 mgage_ k -- guess I'll listen with full attention for awhile.  ttyl
02:23 aubreyja ttyl
02:24 aubreyja ttyl ionparticle
02:31 rbeezer joined #webwork
08:45 djun joined #webwork
12:32 aubreyja joined #webwork
13:53 aubreyja joined #webwork
14:21 Paul_Pearson joined #webwork
15:55 Paul_Pearson joined #webwork
15:56 eee-blt joined #webwork
15:58 eee-blt Q: how can I have the TeX string of a Formula print with a fixed number of significant digits on each fp coefficient?
15:58 eee-blt btw, webwork is a great piece of software, thanks!
16:17 goehle joined #webwork
16:18 goehle hey mgage
16:21 goehle or mgage_
17:23 djun joined #webwork
17:28 mgage__ joined #webwork
17:30 mgage__ aubreyja: hi -- I'm going to have a seminar in about 30 minutes (2pm EDT) but I'll try to get hold of you after that at about 3:15pm EDT
17:30 goehle hey
17:30 mgage__ hi geoff
17:30 goehle I just wanted to ask if my latest changes made the pull request work
17:31 goehle I fixed a bunch of bugs and added a couple of little new features
17:31 goehle no comment emails yet though
17:31 mgage__ yes it did
17:32 goehle good
17:33 mgage__ yeah -- august 22 right?  fixed initialzation stuff with problemGrader?
17:33 goehle right
17:35 mgage__ yep -- it's in there.  you can look at openwebwork/webwork2-dev and click on the commits tab -- you'll see your commits there (back a ways -- I've been busy trying to get everything gathered together and included)
17:35 mgage__ brb
17:35 goehle is it on hosted2?
17:45 goehle because I'm still getting the initialization issues there
17:49 Paul_Pearson joined #webwork
18:05 djun joined #webwork
18:09 aubreyja_ joined #webwork
18:09 mgage no -- it's not on hosted2 yet.  I didn't want to change that until I would have some time to watch it
18:19 CIA-47 joined #webwork
18:23 mgage goehle: ok hosted2 is updated now
18:24 mgage__ joined #webwork
18:29 goehle looks like it works now
18:29 goehle :)
18:29 goehle do you want me to try to get a comment email system up before 1.5.2 is finished
18:30 goehle ?
18:30 mgage sure -- I thought I made a comment when I accepted your commit and that that message would be sent back to you when I did the commit.  Is that not true?
18:33 goehle ?
18:34 goehle It might be there but I didn't get a notification
18:35 djun joined #webwork
18:55 ionparticle joined #webwork
19:17 mgage left #webwork
19:17 mgage joined #webwork
19:25 goehle I can't find your comment anywhere
19:25 goehle O.o
19:26 mgage__ joined #webwork
19:27 mgage__ aubreyja: hi
19:29 goehle when will 1.5.2 be finsihed?
19:29 goehle (it helps to have a deadline ... )
19:31 mgage__ 2.5.2 has the features in it that I want (minus bug fixes) -- basically it has essayQuestions -- so it requires another table upgrade.  Robin Cruz and Rob Beezer are using it (Robin on her machine and Rob on hosted2) -- if nothing big goes wrong I'll release it as it is.
19:32 mgage__ It's hard to set deadlines since at the moment I can predict when it will be used enough to be considered tested.
19:32 mgage__ can't
19:32 mgage__ 2.5.3 is not ready at all -- if there are no emergencies I'll work on that this weekend and try to get the whiteboard feature compatible with everything else.
19:47 mgage__ goehe:  can you do me a favor -- if you are not using the latest ww2.5.2   could you add a fake student, put in an illegal address (e.g.   jane@     ) and then try to send email to the student?
19:47 mgage__ I think I fixed a bug in the latest 2.5.2 that has been around for several years.
19:51 goehle sure
19:52 goehle email sent on a server running 2.5.1
19:52 goehle A message with the subject line
19:52 goehle TestCourse notice
19:52 goehle has been sent to
19:52 goehle 1 recipient(s) in the class TestCourse.
19:52 goehle There were 0 message(s) that could not be delivered.
19:53 goehle thats what I got back
19:53 goehle which seems like a bug
19:53 mgage__ yep -- it never actually reports when it can't send something -- I didn't see how the original code could possibly have worked in any environment.
19:53 mgage__ just verifying that I was right
19:55 goehle nice
19:55 mgage__ yep -- no idea how this bug has survived this long -- I'm pretty sure it's five years old
19:56 goehle probably just dismissed as user error when a student get an email
19:57 mgage__ probably -- since I use the moodle front end in classes I haven't been using the email feature of webwork -- but others use it.
19:58 goehle I do use it
19:58 goehle but my students are imported using a script
19:58 djun The Centre for Teaching, Learning and Technology at UBC is using 2.5.2 for production hosting of several dozen courses, thousands of students
19:58 goehle so their email addresses are pretty solid
19:58 goehle yeesh
19:58 djun I think the perception is that this is the latest stable release
19:58 mgage__ yes -- as long as there are no errors you don't see anything
19:59 djun I've discovered a bunch of problems already in the last 24 hours
20:00 mgage__ just make sure that you have either the version from webwork2-dev or else that you have pulled the version from webwork2 later this morning.  -- the version in openwebwork/webwork2 master was not up and had a VERSION with 2.5.2 in it that it should not have had.
20:01 mgage__ If you want to be safe then update to openwebwork/webwork2   ww2.5.1.1 (the 2.5.1 version with bug fixes)  -- if you also want essay questions then update to the ww2.5.2 branch in openwebwork/webwork2
20:02 mgage__ you will need to upgrade courses going from ww2.5.1.1 to 2.5.1
20:03 ionparticle ah
20:03 djun hi ionparticle
20:03 mgage__ things got a little hairy over the summer particularly after the second code camp when I was a way for a week immediately afterward.  It's taken a while to get all of the various features produced in the two code camps organized into a sequence of releases.  I think that is pretty much done now, although since it's had only light testing there may still be a few glitches -- one I'm pretty sure is still there is the
20:03 mgage__ OPL-update
20:03 ionparticle hi djun
20:04 ionparticle we pulled openwebwork/webwork2 15 days ago
20:04 ionparticle for the update
20:04 djun was just going to ask :)
20:05 mgage__ it's rewritten from NPL-update but I think there might be a few issues still -- Robin Cruz has been having trouble with it -- I'll get to it later tonight if no-one else does.
20:05 ionparticle so if I pull now, we'll get a more stable version?
20:06 djun i have a quick bug fix I'm going to push to fix http://bugs.webwork.maa.org/show_bug.cgi?id=2043
20:06 mgage__ yes -- except that unless you plan to use the essayQuestions you might pull from the ww2.5.1.1 branch (which is also the master branch).  I think what you have now is a buggy version of 2.5.1.1
20:07 djun mike: re OPL update - this is the version Jason & I worked on, yes?
20:08 mgage__ yes -- I believe so -- with additional corrections from ion particle
20:08 ionparticle yep, pull from master
20:08 djun joined #webwork
20:08 mgage__ yes -- I believe so -- with additional corrections from ion particle
20:09 djun i've been spread way too thin over the past few weeks :(
20:09 ionparticle I know the feeling
20:10 djun just getting back to some of the winona code
20:10 ionparticle lol
20:10 mgage__ I understand -- me too.  I think I'll get caught up a little this weekend.  By the way I gave Paul and JohnJones commit privileges to webwork-open-problem-library --- put them on the curate team -- so that they can help get the bugs fixed.
20:11 djun that's a good idea.
20:11 mgage__ I'd actually prefer to work there since I think git will support the editorial board concept a bit better than svn -- I'm still investigating the facilities that github makes possible
20:12 djun does JJ hang out in IRC?
20:12 mgage__ users can still use svn to download -- they just have to point to a different repo -- the one on github instead of the one on webwork.maa.org
20:12 mgage__ not often -- I can't remember if he's ever been on.  Paul does
20:13 djun maybe we can encourage him. It would be nice to have a place to go for discussions
20:13 mgage__ yes
20:18 djun mike, what are your plans for the various branches? Are 2.5.1.1 and 2.5.2 getting merged?
20:18 djun ionparticle, are you guys planning to continue pulling from ww2.5.2?
20:19 mgage__ No 2.5.1.1 is the fixed version of 2.5.1.  To update
20:19 ionparticle was going to pull master/2.5.1.1
20:19 ionparticle but if there's a need for 2.5.2, we can do that
20:20 mgage__ to 2.5.2 requires that you update all the courses since there is a new table for essay answers so I plan to keep that separate.
20:20 ionparticle ah
20:20 ionparticle I'd prefer doing that at the end of term instead, then
20:20 ionparticle lol
20:21 mgage__ 2.5.3 has the white board which I isn't closed to finished yet.  Aaron was working from a two year old version of Problem.pm and both his fork and the main fork have had substantial changes since that time so it will take a while to merge the changes.
20:21 djun i feel that this should really be a minor revision bump. 2.6.x?
20:22 djun more than three digits in a version number make me uncomfortable
20:23 mgage__ well 2.5.1 to 2.5.2 is a minor revision bump, etc. -- the habit has been pretty conservative in the sense that even pretty big substantial changes have just moved the third digit -- at least to date.  I don't care much how the versions are numbered as long as they are increasing :-)
20:24 mgage__ the 2.5.1 ->2.5.2 -> 2.5.3  each adds a new feature and will require the courses to be upgraded.
20:24 djun there's a pretty widely used conversion of major.minor.patch
20:25 mgage__ making 2.5.3  2.6.0 would be ok with me -- and no one is using it yet so it wouldn't be confusing.
20:27 djun I vote for that :)
20:30 mgage__ k -- I'll do that then.
20:30 ionparticle djun, any major showstopper bugs you've noticed in prod?
20:31 djun there is a bug (uninitialized string variable) in pg/macros/unionTables.pl that causes an angry red error message
20:31 mgage__ you can tone down the red in the .css if you wish :-)
20:32 mgage__ I haven't begun to tweak that
20:32 djun there's also a bug in includePGtext (BEGIN_SOLUTION isn't handled)
20:33 djun mike: i've just started looking again at HTML/CSS, hope to get to some of these points :)
20:33 ionparticle ok
20:34 djun I'm going to make some quick fixes and make a pull request.  So, I'll branch from master/2.5.1.1...
20:35 djun gimme a couple of hours :)
20:35 djun heading off to get some lunch
20:45 mgage__ djun branch from webwork2-dev/2.5.1.1 not from webwork2/2.5.1.1  I want all of the developers and all development to take place in webwork2-dev.  I'm hoping this will mean that we don't get accidental updates to branches in webwork2  -- once we are all used to git we might be able to dispense with having a webwork2 and a webwork2-dev  but let's see how this goes for a while.
20:46 mgage__ I'm sure I fixed the BEGIN_SOLUTION problem but it might have gotten lost in my hard disk crash.
20:55 djun joined #webwork
21:05 ionparticle joined #webwork
22:08 aubreyja joined #webwork
22:51 mgage__ joined #webwork
23:16 ionparticle is there a reason that attempted score and recorded scores are implemented differently?
23:17 ionparticle when a student take a gateway qui
23:17 ionparticle quiz*
23:18 mgage_ that's a question for Gavin LaRose I think     glarose@umich.edu   he implemented the gateway quizzes.
23:19 ionparticle thanks
23:20 djun joined #webwork

| Channels | #webwork index | Today | | Search | Google Search | Plain-Text | summary