Perl 6 - the future is here, just unevenly distributed

IRC log for #webwork, 2012-06-15

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

All times shown according to UTC.

Time Nick Message
01:30 mgage_ joined #webwork
01:30 mgage_ hi
11:44 whytheplatypus joined #webwork
12:08 whytheplatypus_ joined #webwork
15:10 aubreyja joined #webwork
15:10 aubreyja joined #webwork
15:12 mgage_ hi
15:34 aubreyja Hi
15:35 aubreyja So, I figured out why AskSage was giving me the 'can't find new()' error
15:35 aubreyja I had listed it in global.conf as qw(Prereqs…AskSage)
15:36 aubreyja but it should have been qw(AskSage…Prereqs)
15:36 aubreyja but then I started getting errors from Safe.pm
15:38 aubreyja The problem seems to be that even though I do all the unsafe stuff I know of in a .pm file, some of it still happens at run time and is called from a macro file.
15:40 aubreyja joined #webwork
15:40 aubreyja joined #webwork
15:59 djun joined #webwork
16:17 aubreyja Hi Djun - thanks for the link
16:17 djun NP, hope it might help.  I'm looking forward to following in your footsteps with R once you have solved all of the hard problems :)
16:34 mgage_ aubreyja:  more  modern coding often does just in time "use" of prerequisites and this causes exactly the problem you are finding with Safe.  Sam and I ran in to this a lot.  What is the Prereqs module doing? loading in modules that you use in AskSage?
16:35 whytheplatypus_ joined #webwork
16:35 djun aubreyja: I'm just looking through documentation trying to get an understanding of how things work and to clean up documentation in the code.  I grepped for dangerousMacros.pl and am confused at what I see.  First of all, it looks like the file is still being loaded in PGloadfiles.pm (but the file is in macros/old_scripts which indicates it is no longer being used).  Second, there is the following from Translator.pm:
16:35 djun $pt -> unrestricted_load("${courseScriptsDirectory}dangerousMacros.pl");
16:35 djun I thought this feature was removed because it was considered dangerous?
16:35 mgage_ I'm just back from Swarthmore (they seem really excited about webwork and I think their technology specialist Andrew Ruether helped by Swarthmore student is interested in eventually helping out with development.
16:36 djun mike: great!
16:36 mgage_ I haven't had a chance to look closely at the code you are working with yet.
16:37 aubreyja be back - we're on now for giving our presentation
16:37 whytheplatypus joined #webwork
16:38 mgage_ djun:   dangerousMacros.pl is left over from the CGI.pm days of WebWorK1 and contained all of the calls which accessed the disk -- that way all of the subroutines that we thought might pose a security risk to the machine were in one place. :-)
16:41 mgage_ the items in dangerousMacros are mostly replaced now by items in IO.pm, PGcore and PGalias() -- PGalias() is only partly finished -- if you like you can help me finish it off -- it's the biggest chunk left from non-object oriented code from WeBWorK1 to the object oriented code that I'm headed for with PGcore.  (PGresponse, etc. etc. are also in support of PGalias).  If you look at the PG alias code you'll see that the method for handli
16:41 mgage_ .html files is new but the other file types are just copied from the old alias() macro in dangerous macros.  My procedure is to move code one step at a time so that I can run continuous checks that it is still behaving the way I expect.
16:42 mgage_ For the most part this works except that I badly need automated unit tests instead of the by hand,  seat of the pants, checks that i do now.
16:43 mgage_ the call to dangerousMacros in Translator is one that you should be able to comment out without affect (and I believe that it is an oversight that it is still there)
16:44 mgage_ I need to go run an errand but I'll be back on in about 50 minutes. I'm planning to write up reports and perhaps look at some code this afternoon.
16:46 mgage_ djun:  is it ok if I add your blog to the blogs we're collecting using the "planet" aggregator at http://webwork.maa.org/planet/
16:49 djun mike: sure - I'm not sure which one would be best to add.  Let me set things up so that there is a permalink for webwork related posts and I'll send you the URL
18:33 djun joined #webwork
18:36 mgage_ djun: you might also be interested in posting some of your blogs to the teachingopensource group (POSSE operates under their rubric)  http://teachingopensource.org/index.php/Main_Page  I think many of your postings would fit in.    their planet is at http://planet.teachingopensource.org/  and their sign up page for the planet is at:   http://teachingopensource.org/index.php/Planet_Feed_List    we don't have an automatic way to sign up
18:36 mgage_ our planet
19:07 mgage__ joined #webwork
19:26 djun Hi Mike - thanks!
19:26 djun I've set things up for my blog here: http://www.cielosystems.com/category/tags/webwork
19:26 mgage__ np -- I think your blog posts will be more interesting than mine.
19:28 mgage__ ok -- give me a fiew minutes and I'll log in to the maa server and try to add your url to jason's and mine.  Encourage others to join in -- I think this may become another useful channel for building WeBWorK community
19:32 mgage__ aubreyja: can you login to MAA -- I can't get through the VPN right at the moment
19:32 mgage__ djun: nudge jason if he isn't watching
19:43 djun OK, thanks.  BTW I think the RSS feed for that tag is feed://www.cielosystems.com/taxonomy/term/23/0/feed
19:43 mgage_ ok
19:43 djun which is probably more that the aggregator will be looking for
19:44 mgage_ I'm still trying to get through the maa vpn wall -- not sure why I can't   -- yes it will be looking an rss feed but I figured that I would be able to deduce it from what you gave me.  Nice post by the way -- I need to learn to write like that.
19:51 mgage___ joined #webwork
20:10 aubreyja Hi
20:11 mgage__ hi -- can you get in to MAA -- I'm being blocked by the VPN.  I want to add Djun's blog to the  planet
20:11 djun thanks, mike! :)  I'm dashing out the door for a couple of hours...
20:18 aubreyja yep, I'm able to get in - I'l add it
20:18 aubreyja actually Mike - for webwork.maa.org you don't need to be behind the firewall
20:18 mgage__ ok -- thanks.  I've asked john wyatt to unlock me -- do you need djun's url?  I can repost it
20:18 aubreyja I've got it
20:19 mgage__ ok -- I'd forgotten that.  ty   -- sounds like you've done good work out there.
20:19 aubreyja yeah, it's been good. I've learned a lot about Safe and Translator.pm :)
20:20 aubreyja I've got the error messages mostly resolved, now stuck at
20:20 aubreyja '' trapped by operation mask
20:20 aubreyja yeah: ''
20:20 mgage__ Send me copies of the files that you put together towards teh end and I I'll try ot put them up on hosted2 (or you can post them to your github) -- we can do this as a relay race while you recover for a bit.
20:20 aubreyja sure
20:21 mgage__ the Swarthmore folks seemed excited. -- we need to figure out how to grab this enthousiasm on a regular basis as it occurs and channel it
20:21 aubreyja great!  I think the main thing is to regularly check in with them and invite them to things
20:22 aubreyja /help
20:22 mgage__ there is a cpan module called something like Safe::Hole that can be used to tunnel through some of this stuff -- Sam and I looked at it a while back -- we've mostly tried to avoid using it if there was some other approach -- but in this case it might be essential.
20:24 aubreyja Djun pointed me to it.  I eventually created a LWP::Request object in the initialization part of PGcore.  Then wrote a getUA() method that I could use in macros
20:24 aubreyja sorry LWP::UserAgent object (hence getUA)
20:25 aubreyja still had to allow all sorts of things into the safe compartment in Translator though
20:25 aubreyja it's an interesting exercise for me to figure out how to do this
20:26 mgage__ it's an interesting exercise for everyone -- I spent dozens of hours or more working on this in the earlier days.  I'd actually like to get back to it.. It's pretty challenging.
20:27 aubreyja yeah, I like the security model a lot actually, but it's not easy to get around (which is good)
20:28 mgage__ Another thing we should consider is getting more in touch with the perl community,  perl monks, and so forth -- I've been too shy about this, -- you've done better.  But I think there is a possibility that WeBWorK is in fact one of the cutting edge projects using perl that is not completely inwardly focused (e.g. Moose -- while a really interesting project is about improving perl for its own sake, not about accomplishing something outs
20:28 mgage__ perl)
20:29 mgage__ I'll have to look at the code -- but it's not impossible that a second pair of eyes will see an alternative way that doesn't involve jumping through so many security hoops.  That has certainly happened before.
20:30 aubreyja yeah, I've tried this a bit.  One approach might be to go to something like YAPC::NA or something similar and give a talk
20:30 aubreyja we could also try to invite some perl luminary to a code camp (although this could be expensive)
20:31 mgage__ let's think about it -- the first approach might not be that hard and would give us some feedback on our likely reception.  We don't have to participate with the perl community but it might be that we've been missing a golden opportunity.
20:32 mgage__ we can also ask Djun his opinion
20:36 aubreyja Djun's post is up - great stuff!
20:39 mgage__ ty -- yes -- he is way ahead of the rest of us on this aspect.  I wrote a newsletter -- which I'm also going to put up on my blog.  (It's an expanded version of what I sent to participants.)  Vicki's been busy so I haven't gotten any feedback yet. -- ok if I sent it to you and see what you think before I post?  I'd like someone to give a quick check to see if I've left something important out.
20:41 aubreyja sure - I can have a look this afternoon - we're just working until dinner
20:42 aubreyja btw, do you want me to make the svn read only?
20:43 aubreyja On the one hand,we need to get instructions up for moving over, but on the other hand we don't want too many updates going to svn….
20:43 aubreyja can updates to svn be pushed to github? I thought that was possible
20:44 mgage__ yes please -- make all of the svn read only -- except first tag the current branch of webwork2 and pg   2.5.0
20:44 mgage__ I'll send the piece but no hurry on looking at it -- or you can skip it entirely if you're worn out.  :-)
20:47 mgage__ updates to svn can be pushed to github (or pulled in to github, not quite sure how it works)  but it's likely to be delicate.  Since I'm positive we want to switch to github exclusively for development I'd just as soon upload a few changes manually and avoid learning how to set up the pull from svn.  It might be easy, but I don't know.  They have made it so that the download from github using an svn client is completely trivial so we c
20:47 mgage__ distribute directly from github in the future and the schools just using svn for updates can transition simply by changing where their svn client points to.
20:48 mgage__ I'll work on getting directions for switching over.
21:03 aubreyja Sounds good - should I also make the NPL
21:03 aubreyja read only?
21:04 mgage__ yes  -- djun is done uploading it
21:04 aubreyja If so, we should start specifically emailing people who have recently contributed in case they are doing more work
21:04 aubreyja so they can start moving to git
21:04 mgage__ and thanks for handling this.  (we also seem to have email as another back channel irc :-) )
21:05 aubreyja :) - np
21:05 mgage__ what you might do is to make the trunk read only but allow contributions on the branches.
21:06 mgage__ but yes I'll try to get messages to active problem developers so that we can start the switch over and see what problems arise.
21:06 mgage__ is it possible to just make trunk readonly? -- seems like it should be
21:06 aubreyja I think I just did that
21:07 aubreyja and good idea about the branches - I did that as you suggested
21:08 aubreyja it might be a good idea to put together a script to convert NPL installations to the new repository…not sure yet how complicated it will be to make sure symlinks keep working
21:08 aubreyja and we need to fix NPL-update and anything in bin/ that uses it, and the macro search paths
21:11 aubreyja 2 commits to CSUOhio just came in - they're working on getting ready for their mathfest short course...
21:13 mgage__ hmm -- ok I'll talk to them, and you might want to talk to Djun as well.  --- just switching to using github should not be too disruptive -- but making some of the other changes Djun is thinking of should be done in a -dev branch (or repository) until probably after MathFest so as not to distract Barbara and Filipe (any others working on that deadline).
21:14 mgage__ can you keep the CSUOhio part of trunk open for now so that they are not disturbed.
21:52 whytheplatypus_ joined #webwork
21:53 whytheplatypus_ joined #webwork
21:54 whytheplatypus_ joined #webwork
22:41 mgage__ afk
23:51 aubreyja joined #webwork
23:51 aubreyja joined #webwork

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