Perl 6 - the future is here, just unevenly distributed

IRC log for #sdl, 2014-08-11

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

All times shown according to UTC.

Time Nick Message
00:10 ldris joined #sdl
01:28 FROGGS_ joined #sdl
03:32 kaare joined #sdl
04:52 kaare joined #sdl
07:24 basiliscos joined #sdl
08:20 FROGGS_ Getty: not as far as I know
08:20 FROGGS_ but libSDL should be buildable, but there are no compiler flags for that kind of arch in Alien::SDL
08:21 FROGGS_ so it treats it as a regular linux when it probably should not
12:58 Getty well, it must be something specific, its just one test that fails (even tho it exists, so i cant tell if there is not like more behind, but its one test file only)
12:59 Getty raspberry is not that different, it just has like..... the lowest hardware ;)
13:13 mrhorner joined #sdl
13:28 FROGGS_ ahh, can you tell which test fails?
13:31 Getty <scratchhead>.... i hope so, else i would kill myself (it took 6<<< hours)
13:31 kaare joined #sdl
13:31 FROGGS_ :/
13:32 Getty https://gist.github.com/Getty/e49939c67c4e8f8e4fe5
13:32 Getty so.... IF i interpret the output right only the not-fine-exit of sdlx_contoller_interface is blocking the show
13:33 Getty but i am not that good at test output parsing ;) so correct me if i am wrong
13:33 FROGGS_ t/sdlx_controller_interface.t ... All 800700 subtests passed
13:33 FROGGS_ I don't think we run that many tests :o)
13:34 Getty AHA?
13:34 Getty so the 6 hours were a mistake?
13:34 Getty i thought it just takes a long time on Raspberry
13:34 Getty its not uncommon
13:34 Getty regexp module installation can take an hour
13:34 Getty Regexp::Common i think is like 30 minutes
13:35 Getty so i wouldnt be surprised about 800.000 tests who run in 1sec on real computers, and like 2 hours on raspberry
13:35 Getty thats why i wasnt confused by that ;)
13:35 FROGGS_ compilin libSDL will take the most time, that'd be Alien::SDL then
13:35 Getty yeah its all about Alien::SDL
13:36 Getty or?
13:36 FROGGS_ that test in SDL only runs 36 tests
13:36 Getty <scratchhead>
13:36 Getty what i installed
13:36 FROGGS_ you install SDL which has a dependency on Alien::SDL
13:37 Getty yeah so those tests are not Alien::SDL?
13:37 FROGGS_ no
13:37 FROGGS_ Alien::SDL has only about five test files
13:37 Getty i waited 6 hours on those tests
13:37 Getty those last ones
13:37 Getty or more
13:37 Getty i was sleeping most of the time
13:37 Getty ;)
13:37 FROGGS_ :o)
13:37 FROGGS_ that's a good way to spent 6 hours :o)
13:38 Getty hahaha i was like laying in bed and was realizing that SDL is the solution for my problem
13:38 Getty and then i said lets start installing the  module on the raspberry
13:38 Getty and when i woke up in the middle it was still going on that 800.000 tests
13:38 Getty and then later i saw it broke
13:38 FROGGS_ you might want to force-install/no-test it since I dunno where to start looking for the bug
13:39 Mithaldu that test taking so long probably indicates a semantic bug in the control flow handling
13:39 Getty yeah sure, in genearl, if you want i can give you access to a raspberry
13:39 Getty if you wanna experiment
13:39 Mithaldu in other news: the whole sdlx stuff is kind of shite
13:39 FROGGS_ Mithaldu: I know :o(
13:39 Getty Mithaldu: if he says that they dont do 800.000 tests, then this is also a clear sign ;)
13:39 Getty hehe
13:39 FROGGS_ but I have no time and actually no interest in improving it much
13:39 Mithaldu Getty: what it does is this:
13:40 Getty Mithaldu: but seriously: raspberry pi is REALLY slow
13:40 Mithaldu it sets up an event loop
13:40 Mithaldu sets up 3 events
13:40 Mithaldu then tells the loop to run
13:40 Mithaldu and only stops when all three events were run once
13:40 Getty ok sounds reasonable
13:40 Mithaldu but because the raspi is so slow the loop is suffocated and one of the events is never run
13:40 Getty hahaha
13:40 Getty which SHOULDNT happen
13:41 Mithaldu you have the high count because two events do a test everytime they're run
13:41 Getty raspberry pi found even bugs in xslate cause of its slowness ;) never underestimate the rpi testing ;) hehe
13:41 Mithaldu https://metacpan.org/source/FROGGS/SDL-2.544/t/sdlx_controller_interface.t#L74
13:41 Getty what about a timeout around?
13:41 Mithaldu i suspect you only got to the end because there's SOME kind of timeout
13:41 Getty that would at least kill the infinity
13:41 Mithaldu maybe even just an integer overflow
13:41 Getty no!
13:41 Getty --verbose
13:41 Getty yeah ok
13:41 Getty that more likely
13:42 Getty but i was on verbose mode
13:42 Getty i must do that
13:42 Getty ese i cant install that stuff at all
13:42 Getty many libs dont install on raspberry pi under.... i dont know that CPAN timeout, i think its 10 minutes
13:42 Getty so i have to activate --verbose which deactivates the timeouts
13:42 Mithaldu cute
13:42 Getty yeah
13:42 Getty raspberry pi and perl is fun
13:43 Getty the only moment where oyu hate the toolchain for being so perfect
13:43 Getty cause the best solution is always "deactivate half of it"
13:43 Getty this situation literally brought me to learn how to pack debian packages
13:43 Getty as thats the only way to deliver perl to a new machine
13:43 Getty setting up perl catalyst and co. from scratch => 1 day
13:44 Getty (or it feels like 1 day, i think can be done in 8-10 hours if it would be one script)
16:51 sivoais joined #sdl
21:24 basiliscos joined #sdl

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