Perl 6 - the future is here, just unevenly distributed

IRC log for #webwork, 2013-11-11

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

All times shown according to UTC.

Time Nick Message
01:13 mgage_ joined #webwork
03:10 mgage_ joined #webwork
04:36 mgage_ joined #webwork
05:02 mgage_ joined #webwork
14:41 mgage_ joined #webwork
16:25 mgage_ joined #webwork
16:30 mgage_ left #webwork
17:00 goehle joined #webwork
17:01 goehle hey mgage
17:08 mgage hi goehle
17:08 goehle hey
17:08 goehle catching up
17:08 goehle looks like you have been busy
17:08 goehle i'll take a look at the pull requests when I get through some of this weeks classwork
17:08 goehle any updates on the slow gateways
17:08 mgage yeah -- I got a little time to devote to webwork after I finished my midterm
17:08 goehle is it slow because of generation of gateways, or because of the past answer database
17:09 mgage my first suspicion is that it is generation of gateways and in general the tuning of the apache server (more or less what Danny Glin reported)
17:09 goehle ok
17:10 mgage In his case it was definitely not pastAnswers since that wasn't around when he solved his problem.
17:10 mgage -- the first thing you might look at is this PGProblemEditor3 interaction with math4 theme -- it isn't handling the overlay properly
17:11 goehle right
17:11 goehle I got your email
17:11 mgage I didn't see any error messages but I didn't try to trouble shoot it.
17:11 goehle this is on vanilla 2.8
17:11 goehle ?
17:11 mgage I can if I get free time before you do, but if you can reproduce the issue than you'll be quicker.
17:11 mgage Yes -- I pulled a fresh copy from openwebwork   release/2.8
17:13 mgage I have submitted and pulled one   1 line bug fix - JJ had left a print message in the Tags.pm file so if you had the tagging mechanisms turned on the print message would get included in an xmlrpc call and everything would collapse.  I changed the print to a warn so I could figure out the remaining problems.
17:13 goehle no worries.  Some of these look like they can be merged pretty quickly
17:13 goehle for example
17:13 goehle pull 230
17:14 goehle i'm just going to merge it
17:14 goehle make sure it compiles and loads a page correctly
17:14 goehle and then merge
17:14 mgage I think so -- and there is some overlap but I don't think they will cause  conflicts
17:15 mgage yep -- I'm sure pretty sure it's safe
17:15 goehle do you have a good problem for me to test the math4jsfix
17:16 mgage the bigger past answer fix was important for me -- it could cause a lot of space wasting -- on the other hand it's easily reversed.   -- you might want to think about it
17:16 mgage yes
17:17 mgage are you a prof in my linear algebra course?
17:17 goehle I was
17:17 goehle but I wanted to test it on my local machine
17:17 mgage I didn't take you out
17:17 goehle to make sure the pull works
17:17 goehle as for the past answer thing, that wont' actually waste that much space, since its a varchar
17:17 mgage kk -- that is what I hoped.
17:18 mgage you can download my problems from fall13mth208 to your machine
17:19 mgage I think you can use FileManager to tar up and download the entire l"local" directory to get all the problems -- or you can download the sets individually
17:20 mgage you can also download the set def file to get the examples I pulled from the library.
17:20 mgage I think something like WBWK3 had some issues with matrices.
17:21 mgage I probably should have written more detail into the code comments.
17:28 goehle ok
17:28 goehle I'll take a look
17:28 goehle i was able to repro the modal issue
17:28 goehle its awfullly strange
17:29 goehle ah nm
17:33 goehle it was because the DoM wasn't ready
17:33 goehle the browser might have changed
17:33 goehle so that its processing the js even faster
17:33 goehle so the page wasn't ready when the js was loading
17:34 goehle i'm going to move all the js down to the bottom of the page
17:34 goehle so that this shouldn't happen
17:35 mgage sorry -- back -- catching up
17:35 mgage kk -- that seems an easy fix.
17:36 mgage I'm going to have seminars and meetings in the afternoon but I'll try to pull some of the remaining requests tonight.  I'm going to start getting conservative about anything but bug fixes however.
17:38 mgage isn't the current practice to put these functions inside an "unload" wrapper so that they only fire after the DOM is ready?  seems to me I see this in jquery scripts all the time.
17:38 goehle maybe
17:38 mgage "onload" wrapper
17:39 goehle I'll take a look at that
17:43 mgage afk
18:25 goehle ok
18:25 goehle so I submitted some things
18:25 goehle I'm also doing some testin
18:29 goehle as far as I can tell
18:29 goehle scratch that
18:29 goehle I'm certain the upgrade tables process doesn't change the size of the past answer table varchar
18:43 rbeezer joined #webwork
19:07 mgage ok -- that needs to be fixed.
19:07 mgage I'll look into it
19:10 goehle does it?
19:10 mgage goehle: I'll also look into the error for hardcopy in hint_solution_tweaks
19:10 goehle hey
19:11 goehle I'm not sure the past answers thing is such a big deal
19:11 goehle any new courses have the bigger field
19:11 goehle however, it shouldln't be that bad to fix
19:11 mgage that's a point -- perhaps it's better to leave the old courses untouched.
19:11 goehle the hardcopy problem was caused by a CAPA file
19:11 mgage kk -- I was wondering it it was an actual tex error.
19:11 goehle no
19:11 goehle so I wouldn't worry about that one
19:12 goehle if that din't break something else would have
19:12 mgage kk
19:12 goehle as for the past answers, table upgrades should upgrade the table
19:12 goehle so thats an arguement for
19:12 goehle however changing column types in general is tricky
19:12 goehle I'm not sure what the current behavior is there
19:12 mgage yes -- what happens to the contents.
19:12 goehle right
19:13 mgage I don't think we've had any upgrades that changed table type.
19:13 goehle this is one of the cases where you can change type without issues
19:13 goehle but its not a huge deal I think
19:13 goehle another question
19:13 goehle is compundProblem4.pl supposed to be executable
19:13 mgage I'm inclined to let this sit for now because of the risk of changing data -- we can decide to make old course tables upgradeable in the next release if we think it's important.
19:14 mgage I suspect that few will care.
19:14 goehle and do you have a problem which uses it
19:14 mgage no it doesn't need to be executable.
19:14 goehle thats a minor issue then
19:17 mgage I just sent you some example problems.
19:17 goehle great
19:17 mgage I gave it a lot of permissions because I was constantly editing it -- I'll turn them off
19:21 mgage I just updated -- I think the permissions will be 644 now
19:23 mgage I did quite a bit of work on the design decisions part of the wiki.  I think the hint/solutions fixes line up with what I have put there.
19:24 mgage go ahead and continue to modify things on those pages (or raise issues on the discussion page) -- I'll put Dick Lane on it this evening -- and we'll get a bunch more feedback. :-)
19:24 goehle :)
19:24 goehle I noticed you changed some stuff with the design decisions
19:24 goehle I'm not fully done making them
19:24 goehle but its a start
19:25 mgage and feel free to change them back -- or raise a discussion.
19:25 mgage I mostly added categories -- and I fleshed out the solution/hint section a bit.
19:26 mgage we'll need a whole section on compoundProblem behavior -- compoundProblem4 just handles the underpinnings -- there is quite a ways to go to make it convenient to write these kinds of scaffolded problems
19:26 goehle yeah
19:27 goehle would that all go under Problem_Page_Design
19:27 goehle or should it get its own page?
19:27 goehle also I think I merged before your file permissions thing came through
19:27 mgage probably on it's own page because it is complicated (linked from the Problem page design
19:28 mgage I'll send an update soon -- I'll have more fixes
19:28 goehle ok
19:28 goehle I also didnt ever merge librarybrowser3tweaks
19:28 goehle it closed automatically for some reason
19:28 goehle I think all of its changes were contained in the other pull requests
19:29 mgage one thing I  thought about is whether to have separate pages or sections for hardcopy behavior and for screen behavior -- rather than trying to explain them both at once.   When the behavior is complicated I think that kind of division makes sense.
19:29 mgage I think that is what happened.  -- those pull requests were not completely orthogonal.
19:30 goehle I think seperate is good
19:31 mgage kk -- there might be a few cases where the behavior is simple and everything can be explained in the same spot.
19:32 goehle yeah
19:38 goehle I've been trying to seperate this out more along the programming divisions than the conceptual ones
19:38 goehle so Hardcopy_Design contains (mostly) decisiosn that affect Hardcopy.pm
19:38 goehle (although its not strictly true all of the time)
19:39 mgage I think that works -- but actually I think we want to aim at a division that makes sense to users -- those giving suggestions now but even more importantly those who use it later --- i.e.   hardcopy print out is doing something they don't expect -- we want to be able to point them to a page that tells them what they should expect
19:40 mgage then they can argue about whether it's the right thing or not :-)
19:40 mgage And I think these questions come up more often in terms of just one mode -- so it's easier to have the pages split that way.
19:41 mgage Users in general however will probably not know what the coding divisions are -- so we should divide the pages along the use case divisions
19:42 mgage ditto with "show past answers for professors" -- a page that tells how to turn that on or off
19:42 mgage I'll need to add more stuff to the solutions/hints page for those who really don't want to use knowls.
19:57 goehle this is sounding like those should get their own pages
19:57 goehle a page for show past answers
19:57 goehle and a page for hints/solutions
19:57 goehle which are referenced from the design decisions page
20:02 mgage afk
21:24 aleahy joined #webwork

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