Perl 6 - the future is here, just unevenly distributed

IRC log for #webwork, 2012-09-21

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

All times shown according to UTC.

Time Nick Message
01:40 rbeezer joined #webwork
02:29 yoavfreund joined #webwork
04:28 CIA-43 joined #webwork
04:28 ionparticle joined #webwork
11:22 mgage joined #webwork
11:33 mgage joined #webwork
11:36 mgage joined #webwork
11:52 Paul_Pearson joined #webwork
12:25 goehle joined #webwork
12:54 mgage joined #webwork
12:54 mgage 'morning
12:54 mgage my "off" day -- I may even have some time to chat
13:42 mgage aubreyja:
13:57 aubreyja Hi - just got in - just a sec
13:58 mgage no hurry
14:06 eee-blt hi
14:06 mgage hi
14:07 eee-blt when i'm writing new ww questions with a graphical component
14:07 eee-blt i am finding the server is serving up the same png even when the code
14:07 eee-blt changes
14:08 eee-blt is there a server-side setting that can stop this?
14:08 mgage it's the caching effect -- either the browser or the server (or both ) is keeping the images.
14:08 eee-blt yes, it is the server
14:08 eee-blt can i turn that off
14:08 mgage try this: http://webwork.maa.org/moodle/mod/forum/search.php?search=refresh+&id=3
14:09 aubreyja Hi - back
14:09 mgage I'm thinking of sending a copy of the email I sent to Farr to this guy Henderson at WPI who writes assistments.  what do you think?
14:10 mgage did you get a chance to read the NYT magazine article?
14:11 aubreyja I think I saw the one you're referring to - about expert systems and tutoring?
14:11 mgage yeah -- http://www.nytimes.com/2012/09/16/magazine/how-computerized-tutors-are-learning-to-teach-humans.html?pagewanted=5&_r=1&ref=magazine
14:11 aubreyja haven't read it though - exam week!
14:11 mgage :-)
14:11 aubreyja it's over though
14:12 eee-blt @mgage, thanks, i'll try that, but i am a bit confused because i am using insertGraph...
14:12 mgage sorry the guy's name is heffernan
14:12 aubreyja I chatted yesterday with yoavfreund about setting up webdav access to his course templates directory so he could easily add and possibly version problem files in his course.
14:13 aubreyja I set it up easily enough, but I think a github like ssh key git-access set up would be better
14:13 mgage I saw that -- that's an interesting idea to add git versioning to local problem authoring
14:14 aubreyja brb
14:14 mgage eee-bit: I think if you put  $refreshCachedImages=1;  somewhere near the top of your problem -- the caching issue will go away.  Once the problem is in production you can comment it out since usually you are not changing images on student problems while they work on them
14:17 eee-blt @mgage: thanks, that worked when i put it near the top of the problem
14:18 mgage great -- which school are you at?
14:18 eee-blt i am at central michigan
14:19 mgage thanks for hanging out in the irc -- it's helping to build critical mass
14:19 eee-blt thanks to you folks for providing this
14:20 mgage np -- it's beginning to gain steam over the last 6 months after coasting for awhile
14:21 eee-blt btw, i git vc my problems and used emacs-w3m plus some customizations of cperl-mode to interact with our webwork server
14:21 eee-blt is there a place to share the emacs stuff?
14:22 mgage yes -- the way we are developing is for you to push your code to a github account under your name.  pull  webwork2-dev and pg-dev from github.com/openwebwork   into your account
14:23 mgage make changes, add files and so forth.  Others can see your github and use your code experimentally -- when you/we etc. feel it's ready for prime time you send a pull request to openwebwork and we'll pull it into your code.
14:23 mgage if it's stand alone scripts it might belong in webwork2/clients or webwork2/bin   -- or perhaps we need another subdirectory
14:24 eee-blt ok, nice, i'll look into this when i get a few spare moments
14:24 mgage great -- look forward to your help
14:24 eee-blt maybe webwork2/emacs ?
14:24 mgage or authoring   authoringtools?
14:25 eee-blt sure, that sounds like a more agnostic setup
14:25 mgage have you looked at the clients directory?  -- that allows external editing via the webservice -- I have it hooked up to BBedit -- it's probably even easier to hook that up to emacs
14:26 rbeezer joined #webwork
14:26 aubreyja back
14:28 eee-blt no, i haven't looked at any of the webwork code, but i guess i should before doing more hacking.
14:28 mgage the client stuff is fairly independent -- look at the README and renderProblem.pl
14:29 mgage the README might be a little behind the code itself
14:29 aubreyja eee-bit: we could keep emacs stuff in, for example, our admintools repo
14:29 mgage ye -- that's another possibility
14:29 aubreyja or, if it goes with the client stuff, maybe just under clients/...?  Or you could make up another directory :)
14:29 eee-blt i have to go mark exams, bye for now
14:30 aubreyja ttyl
14:30 mgage ttyl
14:31 aubreyja ooh, just reading about git-shell and setting up ssh access - I think this is absolutely do-able
14:32 aubreyja we could provide yoavfreund and others git access to their templates directory without having to provide shell access
14:32 aubreyja just like github
14:32 aubreyja that's better than webdav mainly because of authentication
14:34 aubreyja maybe we can even let people upload ssh keys to their courses themselves...
14:39 mgage that sounds very good.  -- another thing to look into is some method such as clients that makes it easier to edit on your own machine while using the web for rendering -- then you can set up git or mecurial or whatever you like for local version control
14:40 Paul_Pearson joined #webwork
14:40 mgage hi paul
14:41 aubreyja ah, but unfortunately ssh access is blocked except through the vpn….
14:50 aubreyja ok, here's a goal: general git access to /opt/webwork/courses/<course_name> as a git repo over https at, say, git.courses.webwork.maa.org/<course_name> with read and write access limited to professors in the course using the Apache::Authen::WeBWorK module
15:02 mgage headed for school --I'll be off line for a while
15:02 mgage Paul_Pearson:  have you had a chance to look at the Statics.pm file that made it into openwebwork/webwork2-dev  ww2.5.1.1 branch?  Is that what you wanted?
15:57 Paul_Pearson joined #webwork
16:19 Paul_Pearson Hi Mike.
16:19 Paul_Pearson Looking at
16:19 Paul_Pearson https://github.com/openwebwork/webwork2-dev/tree/ww2.5.1.1/lib/WeBWorK/ContentGenerator/Instructor
16:19 Paul_Pearson I don't see any recent changes.
16:19 Paul_Pearson I'm probably doing something wrong trying to use git...
16:20 Paul_Pearson What I see is
16:20 Paul_Pearson Stats.pm 23 days ago Fix x-axis label position (hopefully) [paultpearson]
16:20 Paul_Pearson Stats_old.pm a month ago Commit copy of Stats_old.pm [mgage]
16:20 Paul_Pearson and I do not see any other Stat* files.
16:33 mgage_ joined #webwork
16:36 mgage_ hi
16:40 Paul_Pearson hi mike
16:40 Paul_Pearson how busy are you?
16:40 mgage_ not too busy at the moment
16:40 Paul_Pearson did you get my IRC post about Stats.pm?
16:41 Paul_Pearson Here it is again:
16:41 Paul_Pearson 12:19:28 PM - Paul_Pearson: Hi Mike.
16:41 Paul_Pearson 12:19:28 PM - Paul_Pearson: Looking at
16:41 Paul_Pearson 12:19:28 PM - Paul_Pearson: https://github.com/openwebwork/webwork2-dev/tree/ww2.5.1.1/lib/WeBWorK/ContentGenerator/Instructor
16:41 Paul_Pearson 12:19:28 PM - Paul_Pearson: I don't see any recent changes.
16:41 Paul_Pearson 12:19:28 PM - Paul_Pearson: I'm probably doing something wrong trying to use git...
16:42 Paul_Pearson 12:20:10 PM - Paul_Pearson: What I see is
16:42 Paul_Pearson 12:20:10 PM - Paul_Pearson: Stats.pm 23 days ago Fix x-axis label position (hopefully) [paultpearson]
16:42 Paul_Pearson 12:20:10 PM - Paul_Pearson: Stats_old.pm a month ago Commit copy of Stats_old.pm [mgage]
16:42 Paul_Pearson 12:20:26 PM - Paul_Pearson: and I do not see any other Stat* files.
16:42 mgage_ I was switching irc's so I missed it
16:45 mgage_ i think I may have been updating mgage/webwork2-dev -- I'm still getting used to this as well.
16:50 Paul_Pearson A cursory glance at
16:50 Paul_Pearson https://github.com/mgage/webwork2-dev/blob/master/lib/WeBWorK/ContentGenerator/Instructor/Stats.pm
16:50 mgage_ see if you see the changes now
16:51 Paul_Pearson makes makes me think that my changes made it in to your branch.
16:51 mgage_ ok -- I think I've pushed those same changes into webwork2-dev ww2.5.1.1   I'll wait until this weekend to propagate them further
16:53 Paul_Pearson https://github.com/openwebwork/webwork2-dev/blob/ww2.5.1.1/lib/WeBWorK/ContentGenerator/Instructor/Stats.pm
16:53 Paul_Pearson looks good :)
16:53 mgage_ k -- thanks for the updates.
16:54 mgage_ paul -- would you be available/ have an interest in the code camp at Fitchburg Oct 12-14?
16:54 mgage_ I'm a little late getting this organized
16:55 aubreyja Hi - I just about have git push/pull working for courses
16:55 mgage_ cool
17:02 mgage_ lunch break --- back in a bit
17:24 Paul_Pearson1 joined #webwork
17:24 mgage_ joined #webwork
17:24 Paul_Pearson1 Hi Mike,
17:24 mgage_ hi
17:24 Paul_Pearson1 The code camp in October won't work well for me.
17:24 Paul_Pearson1 Not surprisingly, I have plenty of other things to do!
17:25 mgage_ I figured that but I wanted to try anyway
17:27 Paul_Pearson1 I do want to get webwork workin on my laptop to be able to make some small contributions.
17:27 Paul_Pearson1 For example, I think that there are upgrades to the Stats and Student Progress pages that ought to be done.
17:28 Paul_Pearson1 And that I could do these bit by bit.
17:29 mgage_ what kind of laptop?
17:30 Paul_Pearson1 hp pavilion with windows 7
17:30 Paul_Pearson1 I just installed Ubuntu 12.0.4 LTS under Virtual Box
17:30 mgage_ k -- your best bet might be jason's install script
17:31 mgage_ I have a dmg file that works for macs
17:31 mgage_ ask jason about how to use his install script -- it will do most of the work
17:32 Paul_Pearson1 And got everything except for the OPL working (I think that the OPL is the only thing not working).
17:32 mgage_ do you have that updated to the latest in webwork2-dev  ww2.5.1.1?
17:33 Paul_Pearson1 As of last night, yes.  webwork2-dev ww2.5.1.1.
17:34 mgage_ k -- I'll have Jason look at it then -- I'm pretty sure this bug was fixed (it's just renaming things from npl to opl -- but I'm not sure where the fixed script is
17:34 mgage_ aubreyja: around?
17:34 aubreyja yep, on phone with Rajiv
17:34 mgage_ kk  -- have some questions when you are done
17:39 CIA-43 WeBWorK-NPL: jj * r3064 /trunk/NationalProblemLibrary/UVA-Stew5e/setUVA-Stew5e-C02S08-Derivatives/2-8-24.pg: Fixed bug 2438, allowing more answers to be marked right.
17:41 Paul_Pearson1 jj: are you making changes to the NPL or the OPL?  On IRC, the changes appear to be made to the NPL.
17:53 aubreyja hi
17:53 aubreyja mgage back
17:54 mgage_ hi jason -- two things -- can you help Paul with installing ww on his ubuntu virtual machine with your install script?
17:54 aubreyja well, I haven't updated it yet for the new config files...
17:55 aubreyja However, I can probably help debug the problems
17:55 mgage_ and secondly -- could you trouble shoot the OPL script -- the error paul is getting is just a missed name (npl to opl) and I thought we had fixed that
17:55 aubreyja yes, that should be fixed...
17:55 mgage_ could you see if the fix is in ww2.5.1.1 or if not where it is?
17:56 mgage_ Paul_Pearson: jason's install script isn't quite updated yet but he can help with debugging
17:56 aubreyja Paul_Pearson1 - what do you have set for the $problemLibrary{version} in defaults.config?
17:56 mgage_ good point -- seminar coming up -- I'll be back in an hour or so.
17:57 aubreyja ok, ttyl
18:35 djun joined #webwork
18:54 Paul_Pearson1 Hi Jason.
18:54 Paul_Pearson1 I had  $problemLibrary{version} = "2"; in defaults.config.
18:57 aubreyja Hi paul - try changing that to 2.5 and running the script
18:57 aubreyja mgage_ around?
19:23 mgage_ joined #webwork
19:24 djun joined #webwork
19:51 aubreyja mgage: around?
20:02 mgage_ joined #webwork
20:02 mgage_ hi -- back
20:10 aubreyja Hi - I'm liking the idea of setting up courses as git repos.  Then maybe editing a library problem is a fork in some sense  - maybe we can keep origin/version info that way on local edits
20:12 goehle could be an interesting way to do model courses too
20:13 goehle you could fork them
20:13 goehle and then have model courses for specific universities
20:13 goehle which could be forked for specific teachers
20:13 goehle and then theoretically upgrades made to the problems or setup of the main model would propigate
20:14 aubreyja yeah, and maybe we could redo our file manager so that edits from within the file manager end up recorded as git changes when editing from the  github web interface...
20:14 aubreyja There's Git::Repository, for example, on cpan
20:14 Paul_Pearson1 What kind of course info would be publicly available then?
20:14 goehle my only concern would be that we would be pushing a lot of this stuff to a server that nobody owns
20:14 goehle but we could run our own git server for course specific stuff
20:14 goehle if necessary
20:15 aubreyja Well, as background, I set up git access for somebody to their course directory
20:15 aubreyja so it's a local git server served over http from our webwork server
20:15 aubreyja so they can git clone their webwork course, work on problems, etc., locally, and git push them back to their live course
20:16 goehle were you thinking about trying to integrate that stuff into the web interface
20:16 mgage_ the nice thing about having things in a git repos is that it provides a good start for figuring out how to upload improvements from the instructors at various institutions back into the central libraries.
20:16 goehle so course creation would involve cloneing modelCoursel.git or something?
20:16 aubreyja I wasn't thinking of putting course up on github - yes, goehle - also integrating it into the web interface
20:16 mgage_ to get it used by more than a few it would have to be integrated into the web interface
20:16 aubreyja essentially
20:17 Paul_Pearson1 definitely integrate it into the web interface
20:17 goehle sounds fancy, it would make pushing courses and problems from university servers to the main git repo easier
20:17 goehle like mike was saying
20:18 aubreyja yeah, we should try it
20:18 mgage_ I hadn't really gotten started on figuring a out a good method for doing this -- if it's not too difficult to have been keeping things in git all a long then that makes it easier.
20:19 Paul_Pearson1 we should get Jason Grout (or others with git experience and insight) to offer suggestions.
20:19 goehle so would even pg editing be done via git stuff?
20:19 mgage_ so could we add an "undo" to PGproblemEditor by having all the saves go through git?
20:19 aubreyja I'll look into if there is any information about how github itself handles the integration into the web interface
20:19 aubreyja yes to both I think
20:19 goehle could definitely be interesting
20:19 goehle so saving as local would just push the problem to your course git
20:20 goehle then maybe saving on the server pushes to server git
20:20 mgage_ and not too heavy duty -- most unix servers would have git either installed or easily available
20:20 pendos joined #webwork
20:20 goehle and then somehow "save to OPL" saves to the OPL git
20:20 goehle although having save to OPL might be dangerous...
20:20 aubreyja well, for those with push access we could allow that
20:20 mgage_ right to the last one -- I think that as long as you are using the PGproblemeditor it would be saved to the server git.
20:21 mgage_ you could use external clients and a git and editor on your desktop if you wanted.
20:21 aubreyja yeah, that's what I just set up for yoav and got me thinking about this
20:21 goehle right, but then we could integrate some version stuff so that you could push some of your changes from your course.git to the whole server
20:22 goehle that way if you fix a typo, you could push to the server and suddenly everyone using the problem gets the fixed version
20:22 mgage_ the ideal would be to allow push access from any webwork installation to github that editors draw from to add to the OPL
20:22 goehle instead of waiting for it to get fixed on the OPL and coming back down
20:23 aubreyja I wonder if github has something like 'push notifications' that we could use to ping servers that sign up for them…?
20:23 mgage_ I see -- that would probably work -- should probably require admin privileges of some kind -- most sites wouldn't want every instructor to have that kind of control -- too likely to have accidents
20:23 goehle right
20:23 goehle and we wouldn't want every server to be able to push to opl
20:24 mgage_ basically nothing would push to opl -- but you could push to a "submitted for consideration" version of the opl
20:24 goehle yeah, thats better
20:25 aubreyja http://developer.github.com/v3/
20:26 mgage_ I forwarded a note to you from Robin Cruz -- there might be an uninitialized variable somewhere in the Achievements code -- could you take a quick look and see if something is obvious to you?
20:26 goehle line 1729 of which file?
20:27 mgage_ no clue unfortunately -- Achievements isn't that long?
20:28 mgage_ these perl warning messages are not that helpful for undefined variables.
20:28 goehle AchievementEvaluator.pm only goes out to 242
20:28 goehle um, it kind of sounds liek what was happening to me
20:28 goehle when numattemts was getting eaten and was blank
20:28 goehle or max_attempts
20:28 goehle or whatever that variable is called
20:31 goehle I would be very curious if it happens when achievements are enabled but not assigned, and curious about where the red screens of doom appear
20:31 goehle if its after problem submission then the achievmentEvaluator might be to blame
20:31 goehle maybe something isn't getting set correctly when it goes to eval the achievements
20:32 mgage_ any place where it checks due dates?
20:32 goehle some of the achievements check the due dates
20:32 goehle which would explain why it doesn't have a helpful error message
20:33 goehle because the error is not coming from achievement code, but the achievement evaluator code
20:33 Paul_Pearson1 We want the OPL to be high quality and curated well.  As I understand the new paradigm (and I'm not sure I do), fixes to pg problems would flow from the OPL to a university server initially at server setup, and then corrections would flow from the university back to the OPL.  Each university is making different corrections to the OPL.  How do the corrections to the OPL, made by several universities, then flow back to all of the universities?
20:33 mgage_ no -- that's just the standard perl warning message when an undefined quantify gets used in a comparison -- usually perl also does the right thing even though there is a warning
20:34 mgage_ they update  again -- but the flow upstream to the OPL needs to be controlled by an editorial board I think.
20:34 mgage_ unless that makes updates way to slow
20:37 Paul_Pearson1 OK.
20:37 Paul_Pearson1 Right now the editorial board is pretty small...
20:38 Paul_Pearson1 So, we ditch bugzilla for the OPL in favor of git?
20:39 mgage_ not sure yet -- it might be possible to use github's issues like bugzilla.  some people have also suggested trac
20:40 mgage_ it would be nice to be able to fix a bug and mark it closed from one place -- that might be easy with github and github-issues --- it might be possible with github and bugzilla as well
20:40 goehle I can't get any errors by changing due dates on my test course
20:40 goehle and I've got achievements enabled and whatnot
20:40 goehle I dont know what version they are working with though
20:40 mgage_ ww2.5.1.1
20:41 mgage_ no -- I take that back ww2.5.2  because they are using essayQuestions
20:41 goehle 2.5.2 should have pretty new achievements code
20:41 goehle also, you can't submit answers when you act as
20:41 goehle (at least not normally)
20:41 goehle and the achievement code doesn't run unless you are submitting
20:42 mgage_ yes -- you can reset this in the config files if you wish
20:42 goehle sure, but do they have that enabled
20:42 goehle they said that you can see the errors by acting as
20:42 mgage_ probably not
20:42 Paul_Pearson1 If github and github-issues provide enough features to make bug reporting (not to be confused with bug fixing) feasible, then doing away with bugzilla would probably be OK.
20:43 mgage_ yes -- but it's too soon to decide until someone gets more experience with the capabilities
20:44 mgage_ goehle: you're right -- I hadn't noticed that he refers to acting as a student -- it's possible that no students are seeing those errors.
20:44 goehle I'm not saying its not the achivement code
20:44 goehle but its not something I recognize
20:45 mgage_ k -- I'll continue to keep an eye on it -- and I'll ask robin to check the error on a fake student login
20:45 goehle I have seen that error before when Max attempts was blank for some reason
20:46 goehle but I can't seem to force that to happen using the method he describes
20:46 goehle and it could be a completely different uninitialized variable
20:46 mgage_ have you tried acting as a student after extending the due date?
20:46 goehle yeah
20:46 mgage_ k
20:46 goehle hmm
20:46 goehle now I dont know that I have acted as a student
20:46 goehle after extending a due date
20:47 goehle for something that was aready due
20:47 goehle no wait
20:47 goehle its after 9/14
20:47 goehle yeah
20:47 goehle so I've got a test user, and a hw that is normall due on 9/14
20:47 goehle i've extended the due date for them
20:47 goehle and am acting as them and see no errors
20:48 goehle (w/ achievemetns enabled)
20:48 goehle so there must be something else going on
20:49 goehle you coudl try disabling achievemetns and see if the error is still there
20:49 goehle to narrow it down a bit
20:49 mgage_ what about line 1728 in PGanswermacros?
20:50 goehle thats a very good candidate
20:50 Paul_Pearson1 To speed up improvements to the OPL, we need to find a way to turn more professors into bug reporters and more of our bug reporters into bug fixers, and this paradigm shift sounds promising.  Making this all seamless via the web interface would be crucial for its success.
20:51 goehle Paul, that would hellp
20:51 goehle I mean, I submit bugs, but I also have to fix the bugs to use the problem in my course
20:51 goehle so there are fixed versions of teh problems already in existance by the time the bugs are submitted
20:51 Paul_Pearson1 Right, we have people making local bug fixes that never become global bug fixes.
20:51 goehle so maybe submitting a fixed version coudl be attached to the report
20:52 mgage_ yeah -- we need both a few software tools and an updated social structure
20:52 goehle so si this the due date being changed
20:53 goehle or the reduced scoring period
20:53 goehle because I'm not testing with reduced scoring
20:53 mgage_ I've asked paul this already -- we're having another code camp on Oct 12-14 in Fitchburg MA and we're still getting organized as to personel and projects -- any suggestions are welcome
20:54 mgage_ the scoring period is a possibility, or it could be some other random bug that leaves the score undefined in some instances.  -- probably the best fix is an explicit check near 1728
20:54 goehle assuming it is PGanswermacros.pl
20:54 goehle but thats a very good possibility
20:55 mgage_ that variable has been undefined in lots of other places -- so I agree it's a good candidate.
20:55 goehle oh wait nm
20:55 goehle that part of the code is when you are not in reduced scoring
20:56 goehle ah got it
20:56 mgage_ yes?
20:56 goehle so if you are editing the due date for a student
20:57 goehle and you delete their score on a problem and save
20:57 goehle the score will get saved as ""
20:57 goehle and then you get errors if you view the problem
20:57 goehle (even while acting)
20:57 mgage_ yep -- nice catch
20:57 goehle Its possible someone is deleting those manually for some reason
20:57 goehle or maybe there is another mystery
20:58 goehle but maybe a good fix would be to check and see fi that field is "" when saving stuff
20:58 mgage_ that's a likely explanation.  I guess we can put in two fail safes -- one that always converts blank scores to 0 in ProblemSetDetails.pm and one that checks for definition of the scores in PGanswermacros.pl
20:58 goehle good idea
20:59 goehle (*whew* not my fault ;) )
20:59 mgage_ you willing to do this -- and send me a pull request?
20:59 goehle sure
20:59 goehle for pg-dev and webwork2-dev?
20:59 mgage_ thanks -- yes -- and ww2.5.1.1 -- I'll take care of propagating the fix foward
21:00 goehle hmm
21:00 goehle ok so if I want to make a change to 2.5.1.1
21:00 goehle and I'm on my dev server
21:00 mgage_ if you can update your repo from ww2.5.1.1 before submitting so that I can get a clean commit --- it's dangerous if I have to reconcile things.
21:00 mgage_ you have git set up?
21:00 goehle I do
21:01 mgage_ do you have remotes
21:01 goehle I do
21:01 goehle my upstream is openwebwork/webwork2-dev
21:01 goehle so I could do
21:01 goehle git pull upstream ww2.5.1.1
21:02 mgage_ so if you are linked to openwebwork/webwork2-dev  by some remote  you can also set a local branch to track ww2.5.1.1  then you can simply do
21:02 goehle and that should put my system in ww2.5.1.1 mode?
21:02 mgage_ exactly
21:02 mgage_ do checkout  ww2.5.1.1 first
21:02 goehle where does my existing system go?
21:02 goehle I suppose I would
21:02 goehle ah yeah
21:02 goehle right
21:02 goehle so I dont mess with my master
21:02 mgage_ yes
21:02 mgage_ actually I would do the following
21:02 mgage_ git checkout ww2.5.1.1
21:03 mgage_ git checkout -b fix_undefined_variable
21:03 mgage_ that puts you on a new branch extending ww2.5.1.1 (so you can backout if all hell breaks loose)
21:03 goehle any ideas on how to check I'm actually in 2.5.1.1
21:03 mgage_ git branch
21:03 goehle right
21:03 goehle but a sanity check
21:03 mgage_ there will be a * next to the branch you are on
21:03 goehle on the files themselves
21:04 goehle ok
21:04 goehle VERSION
21:04 goehle says 2.5.1.1
21:04 mgage_ then while on fix_undefined_variable pull from ww2.5.1.1
21:04 goehle so I *should* be ok
21:04 mgage_ yep
21:04 mgage_ now fix things, test    then
21:04 mgage_ git checkout ww2.5.1.1
21:04 goehle and merge
21:05 mgage_ git merge fix_undefined_variables
21:05 goehle then push to *my* git
21:05 goehle then issue a pull request to you or openwebwork?
21:05 mgage_ to openwebwork for this one
21:05 goehle (on for branch 2.5.1.1)
21:05 goehle ok
21:05 mgage_ yep
21:05 goehle lets give it a try
21:05 goehle using git properly
21:05 goehle how exciting!
21:06 mgage_ ideally this will come in and I'll get a message saying this can be merged without conflicts and I'll just accept it.
21:06 mgage_ when there are merge conflicts it is both more work, and I might make a mistake in deciding which version to accept.
21:07 aubreyja on my way out - ttyl - should be around sat/sun though too - we really need to think about fitchburg!
21:07 Paul_Pearson1 I just got an idea for Webwork::Fitchburg.  How to use git properly.
21:07 mgage_ yes -- I'll be working Saturday morning
21:07 Paul_Pearson1 And make a video of it!
21:07 aubreyja ok, I'll talk to you then. Good idea Paul
21:08 mgage_ or more to the point writing a how to manual or a video
21:08 mgage_ yes
21:08 mgage_ ok -- I've got to head home in a few minutes
21:09 goehle sec
21:10 goehle pg-dev doesnt have versions?
21:10 goehle (on git hub at least)
21:11 goehle bleh, the checkout borked my installation hmm
21:11 mgage_ you can  git checkout master   to get it back I think
21:12 goehle sure
21:12 goehle but I need to figure out how to get it running
21:12 goehle without changing things
21:12 mgage_ ok what version is your master on?
21:12 goehle 2.5.2
21:12 mgage_ I have the new versions 2.5.1 and 2.5.2 switching pretty easily
21:13 mgage_ ok it's the database.conf that is causing troubles -- just copy it over from database.conf.dist each time you switch -- haven't quite fixed that yet
21:13 mgage_ it's because the table structure changes between the two versions
21:13 goehle actually apache is comlaining about stuff in Stats.pm
21:13 goehle is that the same thign?
21:13 mgage_ no
21:14 mgage_ but that is a new upload from Paul in Stats.pm -- and we had conflicts so I might have made some wrong choices.
21:14 goehle syntax error at /opt/webwork/webwork2/lib/WeBWorK/ContentGenerator/Instructor/Stats.pm line 575, near "HEAD\nprint"\n  (Might be a runaway multi-line << string starting on line 572)
21:14 goehle I can try to git it fixed
21:15 goehle and upload that fix with the undefined variable fix
21:15 mgage_ I see that too -- somehow I missed a conflict
21:15 goehle I just wnated to make sure it was actually broken before I started changing files
21:15 goehle ah that would explain the <<
21:15 mgage_ yes
21:15 goehle which should it be?
21:16 goehle HEAD
21:16 goehle or the other one
21:16 mgage_ one sec
21:16 mgage_ that's the issue I kept having --it should be HEAD in this case
21:17 mgage_ you can just fix that as well and I'll get all the changes when you upload.
21:18 goehle sure
21:18 goehle hmm
21:18 goehle now it wont read database.conf
21:18 goehle I can probably figure this one out
21:18 goehle though if you need to go
21:19 mgage_ is it there?
21:20 mgage_ you'll need to copy it from database.conf.dist -- there are not usually any local modifications -- we might stop distributing database.conf.dist and just distribute database.conf
21:21 goehle it was there, and a refresh fixed everything
21:21 goehle ok, I'm ediing UserDetail.pm now
21:21 goehle so looks like I'm g2g
21:22 mgage_ Perhaps ProblemSetsDetail.pm needs to be looked at also?
21:22 goehle yeah
21:22 goehle its ProblemSetDetail
21:22 goehle just figured it out myself
21:22 mgage_ kk -- I'm headed home
21:22 mgage_ be back on in 45 minutes or so
21:58 goehle bah
21:58 goehle someone was too clever by half
22:03 Paul_Pearson joined #webwork
22:25 Paul_Pearson goehle?
22:25 goehle yeah
22:25 Paul_Pearson I can send you a version of Stats.pm that works and should be ww2.5.1.1 compliant.
22:26 goehle I got my version working
22:26 goehle mgage just forgot to take out a merge conflic
22:26 Paul_Pearson I'm terrible with git right now...sorry for messing things up for you.
22:26 goehle no, its Mikes fault :)
23:02 goehle_ joined #webwork
23:02 goehle_ hey
23:02 mgage_ joined #webwork
23:02 mgage hi -- the pull request worked fine.  I left a note for you on the github site -- does that get sent to you by email?
23:05 goehle_ hey
23:05 goehle_ I did
23:05 goehle_ and I responded via email
23:05 goehle_ I wonder if that oges back ot gitub
23:07 mgage yes it does -- https://github.com/openwebwork/webwork2-dev/pull/49
23:08 goehle_ yeah, so $variable = 0 || "" kind of blew my mind
23:09 goehle_ I also changed things in PGAnswermacros so that it sets recorded_score to 0 if its ""
23:09 goehle_ so even if thre are current problems they should go away after a submit cycle
23:09 mgage there's a new command \\ in the latest perl that is supposed to help with these initialization statements.  -- it actually has been a kind of common idiom -- but as you notice it doesn't work that well in practice
23:09 mgage yep -- thanks -- that was nice work
23:10 goehle_ well, intitialization wasnt the issue here I guess.  it needed to be $variable = defined $default ? $default : ""
23:10 goehle_ good times in any case
23:10 mgage sorry it's // as an operator  http://www.thegeekstuff.com/2012/01/perl-5-10-new-features/
23:11 mgage the defined OR operator is much cleaner -- and we are getting to the pint where nearly everyone has perl 5.10 I think.
23:12 goehle_ oh sure
23:12 goehle_ thtas kind of neat
23:12 goehle_ hey, do you know why they dont just interpret "" as 0
23:12 goehle_ for numerical comparisons
23:12 mgage smart match is another useful idea --
23:12 goehle_ they switch betwen strings and numbers seamlessly otherwise
23:13 mgage they do -- but they also issue the warning if you have the -w flag on (I think or perhaps it's use strict) -- while I do like perl's flexibility I think we have to balance it by running it with as much lint protection as possible
23:13 goehle_ ah ok
23:13 goehle_ man too bad I didn't know aobut //
23:13 goehle_ that would have made fixing that much easier
23:13 mgage it's also why I have the page turn red on warnings -- too many people were just ignoring the warnings
23:14 mgage it's annoying but I've held my ground against getting rid of the red pages -- this way things do get fixed.
23:14 goehle_ no I think its a good idea
23:15 goehle_ I used to use PerlReload to automatically reload perl modules after I changed them
23:15 goehle_ but I stopped becaues it was giving me lots and lots of warnings about things being redefined
23:15 goehle_ and that was making me ignore red pages
23:16 mgage I'd like to have a clean up version of webwork -- maybe next summer or sometime when there is lots of time.  I'd like to change things like using //,   also rename some of the file names so they correspond more closely to what you see on the screen (I can never remember that ProblemSetsList is the homeworksets editor ).  -- things that will undoubtedly cause lots of small breaks, but when we are done it will be a lot saner for new peopl
23:16 mgage come in and help out
23:17 goehle_ its a good idea
23:19 goehle_ doing some consistent commenting would be a good thing to do at the same time
23:19 goehle_ for example, adding a comment before every function
23:19 goehle_ listing what the inputs to the function are
23:19 goehle_ I've gotten used to looking at the shifts to see what variables are coming in
23:19 goehle_ but it was veyr confusing at first
23:20 mgage yes -- we've had some attempts on that.  it's a good thing for a code camp because people can egg each other on.  It's easy to get bored with cleanup like this
23:20 goehle_ thats a problem with open sourc estuff
23:20 goehle_ nobody wants to work on boring things ;)
23:20 mgage actually that is a perl idiom (the shifts) that seems unusual at first but it ends up working pretty well as long as you are consistent about shifting in the needed variables at the top of the block.
23:21 mgage yep
23:21 goehle_ I like the shifts
23:21 goehle_ but a comment with a short description of the variables coming in would go a long way towards explaining how to call the functions properly
23:21 mgage me too -- I find it quite elegant -- and it gives you automatic tail arguments
23:22 mgage ($a, $b, $c) = @_;  works well too, but for just a few variables
23:22 mgage I like shift
23:22 mgage I agree about the comments
23:22 goehle_ well I would certainly be willing tow ork on it, I think its important
23:22 mgage I'll take you up on it at some point. :-)
23:23 goehle_ sure, I guess the quesito nis prioities
23:23 mgage but only when you have the time to spare
23:23 goehle_ yeah, not much right now, got reappointment stuff to do
23:23 goehle_ and there are some things I should probably do with essay answers
23:23 mgage that's what I figured -- good luck with that.  I don't owe you any letters do I?
23:23 goehle_ nope, you already snet me one
23:25 mgage good -- just making sure.  If you feel like taking a little time out you could write a blog post on your achievements addition, and another on essay answers.  -- not a huge publication, but things are changing  a little and a web presence is not actually detrimental to reappointment these days :-)  might even help a little
23:28 goehle_ I was actually doing that when you emailed me abou tthis bug
23:28 mgage_ joined #webwork
23:29 goehle_ then I got sidetracked obviously :)
23:29 mgage :-) -- I'll get back to putting your changes into the other branches.
23:30 goehle_ ah, I mean emailed me earlier this evening
23:31 goehle_ so I did hav ea questoin
23:31 goehle_ is there an easy way to get the pg interpreter to not try and evaluate perl code and leave \{ as a \lbrace
23:31 goehle_ or should I just do a regexp and turn \{ into \lbrace and \} into \rbrace
23:32 goehle_ (this is so that students can use \{ and \} in their latex in essay answers)
23:32 goehle_ I know that you can make EVE3 (or whatever its called) not evaluate perl, but I dont know if you can get the translater to ignore it
23:37 goehle_ oh, and about the blogs, do you want them to be instructional?  or just descriptions of the features?
23:39 mgage I think newsy and descriptions -- perhaps enough instruction to get people interested.  -- if we really want instructions to refer to we could put them on the wiki where they are easier to update (and you could refer to a page in the wiki from the blog).  That's my gut reaction -- I don't have any firm beliefs at the moment and i don't think we need to enforce blog uniformity -- at least not yet

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