Perl 6 - the future is here, just unevenly distributed

IRC log for #webwork, 2013-09-20

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

All times shown according to UTC.

Time Nick Message
02:07 rbeezer joined #webwork
02:19 mgage_ joined #webwork
02:20 mgage_ rbeezer: hi
02:21 rbeezer hey Mike
02:21 mgage_ could you check something for me quickly  -- see if you can reach https://openwebwork.org
02:21 mgage_ or http://openwebwork.org
02:21 rbeezer Not found :-(
02:22 mgage_ ok -- the site seems to go down sometimes -- but at least once I couldn't reach it and others could.
02:22 rbeezer ping: unknown host openwebwork.org
02:22 mgage_ I was able to reach it about 10 minutes ago
02:23 mgage_ maybe a problem with looking up the ip number
02:23 rbeezer Ping me again alter and I can ping again  ;-)  I'll be ahere for a couple hours probably
02:23 mgage_ kk
03:49 mgage_ joined #webwork
03:49 mgage_ rbeezer: http://openwebwork.org/the-webwork-software-process came back up -- so I guess it was just intermittent -- I'll look further if it keeps happening. thanks for the help.
03:49 rbeezer no problem, that was easy  ;-)  anytime.  Consider me your WestCoast late-night resource
03:50 mgage_ will do.
12:04 mgage_ joined #webwork
12:52 mgage_ joined #webwork
13:22 goehle joined #webwork
13:23 goehle \me
15:38 aubreyja joined #webwork
15:38 aubreyja joined #webwork
15:53 mgage joined #webwork
16:46 goehle left #webwork
16:46 goehle joined #webwork
16:50 goehle joined #webwork
16:51 goehle test
17:06 goehle hey mgage
17:34 mgage goehle: hi
17:36 goehle not that I need more to do
17:36 goehle but since I administer our labs at WCU
17:36 mgage yes
17:36 goehle I could use them (30 or so computers) to set up some bots to load test webwork
17:37 goehle It would be an interesting problem because I've never written a bot before
17:37 goehle but it would distract me from other things
17:37 mgage things you want to be distracted from ?  :-)
17:37 mgage you are already doing plenty
17:38 goehle sure.  And it would slow down my bug fixing etc
17:38 goehle like redoing mathview for the file manager
17:38 goehle or getting the reset-able homework stuff worked out
17:38 goehle but I feel like it would be really useful
17:38 mgage it's a good think to keep in mind in general -- and the other activities you mention are not that time sensitive (perhaps math view is)
17:39 goehle if I could get a bot going that could be used to stress test webwork.
17:39 goehle and it would save us from finding out things the hard way
17:39 mgage I'm interested in figuring out ways to do automatic unit testing in general -- as well as time measurements.    load testing is in there too.
17:40 mgage sure go ahead and see how hard it is to set up a bot.  -- some of the tools in webwork2/clients mightt be relevant -- and some of the work on the web service we are planning for October are relevant as well
17:40 goehle I'll  poke around
17:40 goehle what kind of stuff should it measure
17:40 goehle average time between submit and response
17:41 mgage It may be easier to test some components such as database retrieval and so forth separately -- not a perfect predictor of behavior in the wild but it would help.
17:41 goehle true
17:41 goehle I was thinking I would use something which pretended to be a web browser
17:41 goehle so everything gets tested
17:41 goehle although that makes it harder to tell whats failing sometimes
17:41 mgage yes that kind of thing there is a module called DPROF   and an improved version that was written at the NYT
17:42 mgage Fake on macs is one thing that can pretend to be a web browser but it's a bit clunky and there are probably better choices on unix.
17:42 goehle http://en.wikipedia.org/wiki/Selenium_(software)
17:43 mgage that looks very good actually
17:45 goehle I'll give it a try
17:45 mgage kk ty
17:46 goehle np.  I'm just trying to avoid being responsible for breaking peoples stuff ;)
17:47 mgage I actually think that this particular instance can be handled without a lot of load testing -- but I'm also trying to get a balance established between getting new stuff tested and not trying anything new for fear something might go wrong.
17:48 goehle yeah.  I agree with Gavin's comment.  It will probably be fine
17:48 mgage Having easy ways to back out of an experiment is a good start -- and we have made a lot of progress in that direction in the last year -- but there is still more to do.
17:48 goehle I like that balance.  It will be easier though if testing can be done quickly and with minimal risk.
17:49 mgage I'll get back to comment more on that this evening -- it's an excellent case to consider when trying to write the release procedure document.
17:49 mgage yes -- a suite of unit tests or load tests would be fantastic.
17:50 mgage I need to run in a minute but if you wish we could talk this weekend about a bot that runs through the library and flags errors in problems -- I have had one working (off and on) for several years now --- but it needs more work to become really robust.  perhaps that would be a good place to start using selenium.
17:51 rbeezer joined #webwork
17:51 goehle sounds good, although my weekends can be busy with Howie.  Sunday evening would be when I could be more available
17:51 mgage kk
17:57 mgage afk
21:45 mgage_ joined #webwork
23:53 mgage_ joined #webwork

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