Perl 6 - the future is here, just unevenly distributed

IRC log for #perl11, 2016-08-27

| Channels | #perl11 index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:31 travis-ci perl11/cperl#1478 (smoke/maint-5.24c - be24e8a : Reini Urban): The build failed. https://travis-ci.org/perl11/cperl/builds/155485722
00:40 mako left #perl11
07:32 anton joined #perl11
08:21 travis-ci perl11/cperl#1480 (smoke/morebugs - 33a800d : Reini Urban): The build was fixed. https://travis-ci.org/perl11/cperl/builds/155537056
08:44 basiliscos joined #perl11
08:53 travis-ci perl11/cperl#1481 (smoke/maint-5.24c - 2138bb7 : Reini Urban): The build is still failing. https://travis-ci.org/perl11/cperl/builds/155537071
09:26 mako joined #perl11
09:34 travis-ci perl11/cperl#1482 (maint-5.24c - 2138bb7 : Reini Urban): The build has errored. https://travis-ci.org/perl11/cperl/builds/155545456
10:09 travis-ci perl11/cperl#1483 (master - 33a800d : Reini Urban): The build passed. https://travis-ci.org/perl11/cperl/builds/155545467
10:30 travis-ci perl11/cperl#1482 (maint-5.24c - 2138bb7 : Reini Urban): The build passed. https://travis-ci.org/perl11/cperl/builds/155545456
10:51 mako joined #perl11
11:01 travis-ci perl11/cperl#1484 (maint-5.24c - a225d66 : Reini Urban): The build passed. https://travis-ci.org/perl11/cperl/builds/155553856
11:37 travis-ci perl11/cperl#1485 (master - 0ec7aaf : Reini Urban): The build has errored. https://travis-ci.org/perl11/cperl/builds/155553877
12:08 travis-ci perl11/cperl#1486 (smoke/various - 66bf271 : Reini Urban): The build passed. https://travis-ci.org/perl11/cperl/builds/155560698
12:11 sten joined #perl11
12:19 mako joined #perl11
13:52 mako joined #perl11
13:58 travis-ci perl11/cperl#1487 (maint-5.24c - c5cfc31 : Reini Urban): The build passed. https://travis-ci.org/perl11/cperl/builds/155576131
14:35 travis-ci perl11/cperl#1488 (master - 66bf271 : Reini Urban): The build passed. https://travis-ci.org/perl11/cperl/builds/155576284
14:47 mako joined #perl11
15:17 willthechill mako: unanswered question from yesterday...
15:17 willthechill "what I'm asking, is if you have used your own windows machine to remotely log into the appveyor VM windows desktop, so that you can see all the processes which are running in the appveyor VM instance?"
15:17 willthechill "there is a username and password to do so"
15:18 willthechill this is NOT the same thing as running the code on your local machine.  it is also NOT the same as simply viewing the output of the appveyor instance.
15:18 willthechill it requires you to use your own windows machine to remotely log into the appveyor windows desktop
15:18 willthechill so we can figure out what process is actually frozen or hung
15:46 mako joined #perl11
16:03 mako joined #perl11
16:13 mako joined #perl11
16:15 mako willthechill: When it comes to Appveyor I'd like to make you an offer: Please give me 14 days (yes, this is much) to come up with a working test. Would you mind?
16:16 mako willthechill: I'm going to try a radical way, but this will take some time.
16:17 willthechill so did you ever log into appveyor remotely?
16:18 mako no, I didn't.
16:18 willthechill okay then that is the first step
16:18 willthechill let me know whenever you are ready, and I will click "rebuild"
16:18 willthechill then the VM will go live
16:19 willthechill again, we just need to know which process is hanging
16:19 mako excellent, I hope 14 days are not too much, are they?
16:20 willthechill I'll be here for 1400 days, waiting for you to tell me that you are ready for me to hit the "rebuild" button
16:20 willthechill no new code, just "rebuild" on the current code
16:21 willthechill this takes me 14 seconds to achieve
16:21 willthechill if you are unable to use windows to login to appveyor remotely, then please tell me now so that I don't keep asking you over and over and over again
16:22 willthechill I can find somebody else to log into windows in far less than 14 days
16:23 willthechill we could already be done with it by now
16:23 willthechill like, in less than 5 minutes total effort
16:23 willthechill that does not take 14 days to achieve, my friend
16:24 willthechill I don't mind finding somebody else, I just assumed it would be trivially easy to achieve
16:24 willthechill I saw bulk88 do it in person, it took him about 10 seconds
16:25 mako Wait, wait, we are talking about 2 different issues. The 14 days have nothing to do with Appveyor remote desktop.
16:26 willthechill okay well you have come up with your own private, second issue
16:26 willthechill the only issue at hand for RPerl is to log into appveyor
16:26 willthechill can you do it, or not?
16:26 willthechill I need to know the answer
16:29 mako Yes, I can.
16:30 mako Trying...
16:30 willthechill no the VM is not live
16:30 willthechill that's why I have said that I need to click "rebuild"
16:31 willthechill okay I clicked rebuild, now you can try it, same URL as always
16:31 willthechill https://ci.appveyor.com/project/wbraswell/rperl
16:36 mako can see it
16:37 willthechill okay great!
16:37 willthechill it is frozen now
16:37 willthechill so we have about 50 mins until the VM shuts down
16:37 willthechill so please tell me, which process is hanging or frozen, etc?
16:37 willthechill bulk88: are you here?
16:39 mako # Test File #1 of 205: blib/lib/RPerl/Test/Conditional/program_05_good.pl ### hanging
16:39 mako Had a look in taskmanager.
16:39 mako Looks like cc1plus.exe is busy
16:39 willthechill okay great now we are finally getting some answers!!!
16:40 willthechill what args to cc1plus?
16:40 willthechill file names, etc?
17:07 mako Damn I lost it somehow. Have a look: https://raw.githubusercontent.com/93r/rperl/test/t/appveyor.png
17:07 mako https://raw.githubusercontent.com/93r/rperl/test/t/appveyor.png
17:07 willthechill checking
17:07 mako Wait where is my link?
17:07 willthechill https://raw.githubusercontent.com/93r/rperl/test/t/appveyor.png
17:08 mako Ah, yeah, my irc-client was reacting slowly.
17:09 willthechill I see no cc1plus in your image?
17:10 mako Yes, I lost it.
17:10 mako But it was there.
17:10 willthechill okay so what is frozen now?
17:12 mako 'lib/RPerl/Test/Conditional/program_05_good.pl' seems to hang.
17:12 mako AFK. Back in 30 min.
17:31 willthechill joined #perl11
18:09 basiliscos joined #perl11
18:12 mako joined #perl11
18:37 mako joined #perl11
18:42 mako willthechill: Do you have any 'appveyor.yml'? I need to use the same Appveyor config you are using.
18:44 willthechill sorry, no
18:44 willthechill it comes from github automatically I think
18:44 willthechill I have no such file
18:46 mako Hm, weird. Because when I do an Appveyor test with my fork of RPerl it gives me: "Specify a project or solution file. The directory does not contain a project or solution file."
18:46 mako So therefore something is missing.
18:48 willthechill yes, like I said, it comes from github automatically
18:48 willthechill the only solution I know is to hook up your github account to your appveyor account, the same way I have done so
18:49 mako In appveyor settings?
18:51 willthechill https://www.appveyor.com/docs/
18:53 willthechill that is the appveyor "getting started" URL
18:53 willthechill it tells you how to hook up your github to appveyor
19:06 mako joined #perl11
19:10 mako willthechill: Thx. I'll study this stuff. But for now I close any testing or development or whatever for today. I'm absent-minded.
19:11 willthechill mako: okay goodnight my friend!
19:11 willthechill :-)
19:11 mako willthechill: BTW: Sorry for my '14 days'-stuff. It wasn't my intention to piss you off. :-(
19:11 willthechill nope not mad
19:12 willthechill I will just keep repeating myself until I get through, haha!
19:12 mako That's the reason you are the captain, captain.
19:13 willthechill LOL good one  ;-)
19:13 willthechill oh wait, one more thing!
19:13 willthechill cool new cover pic for Learning RPerl just went live
19:13 willthechill http://rperl.org/learning_rperl.html
19:14 willthechill it isn't "Roadie the Roadrunner" himself...  but I think it might be his great-great-great-grandfather!
19:14 willthechill hahaha
19:16 mako Nice, like it. The day will come we will have an RPerl release having a name that contains the term "runner". :-)
19:17 willthechill maybe so, maybe so!  :-)
19:17 mako But anyway, would it really be such bad idea to develop a custom Appveyor test involving build perl from scratch, using M$'s C++ compiler, linker, and maker?
19:17 willthechill bad idea?  I don't guess so
19:17 willthechill time to do it now?  not at all
19:17 willthechill gotta get normal appveyor working first
19:17 willthechill must do things in order
19:17 willthechill not out-of-order
19:18 mako But you wouldn't mind if I would come up with something like that? Someday.
19:19 willthechill bulk88 already does MSVC on his system
19:19 willthechill it works
19:19 willthechill but it is a pain
19:19 willthechill I will never complain about somebody getting RPerl to automatically test on a new platform or OS
19:20 willthechill as long as it doesn't break something else
19:21 mako Very good. Did bulk88 make his efforts available, trackable, or traceable?
19:21 willthechill negative
19:21 willthechill that is why we MUST get NORMAL appveyor working now
19:22 willthechill and by "now" I mean ASAP
19:22 willthechill if we can't get normal appveyor working, then everything else is silly
19:22 willthechill normal appveyor is the "easiest" starting point
19:22 willthechill I put "easiest" in scare quotes because it is still quite difficult to achieve
19:22 willthechill apparently
19:26 mako When I was remotely connected to Appveyor, I noticed they've got MS C++ build tools already installed. You should ask bulk88 how this can be used to our advantage.
19:27 mako Using M$ compiler, linker, and maker mainly that is.
19:27 willthechill yes, I assume it can somehow be used to create a MSVC automated build
19:27 willthechill after we have non-MSVC working
19:28 willthechill bulk88 has been missing lately
19:28 willthechill :-(
19:28 mako Maybe building with Microsoft compiler should avoid the hanging of the tests. (Remember test are hanging because of missing SEH support in g++).
19:29 willthechill no way man
19:29 mako No way?
19:29 willthechill if the only way to support RPerl in Windows is MSVC then I will actually call it off
19:29 willthechill we can't expect people to have MSVC
19:29 willthechill it's bad enough they are using Windows itself already
19:29 willthechill requiring MSVC is not acceptable to me
19:30 willthechill ALLOWING it is fine
19:30 willthechill REQUIRING it is not
19:30 willthechill and I insist on free tools working first
19:30 willthechill if MSVC never ever works, I will not cry
19:30 willthechill (although we already know it can be made to work, albeit with quite a lot of crazy effort, because bulk88 has done it)
19:31 mako Ok, I'm respecting this. You've got good points.
19:31 mako Prefer to look at the situation the way you do.
19:32 willthechill it was crazy difficult for bulk88 to get it working in MSVC
19:32 willthechill requiring thinks like binary kernel hacks or the like
19:32 willthechill IIRC
19:32 willthechill way way way beyond mortal humans
19:32 willthechill so MSVC is like the one-thousandth thing on our to-do list at this point
19:33 willthechill give or take a few hundred
19:33 willthechill haha!
19:35 mako Got similar issue years ago with Active-Perl. Had to use a hexeditor and myself getting well acquainted with assembler to get stuff working. Beyond mortal too. Something you don't wish even to your enemys.
19:37 willthechill exactly!
19:37 willthechill even gmake is too much for us right now, and that could probably be solved with a 1-line patch to an open source software product
19:37 willthechill MSVC is waaaaay too much for us right now
19:38 willthechill we can't even run in normal appveyor mode
19:38 mako But the question is still remaining: How can we bypass this exception handling issue (SEH) with g++ on Windows? The apparent reason our tests are hanging.
19:38 willthechill if we can't beat the game on "easy" mode, then we sure won't beat it on "hard" or "expert" mode
19:38 mako Rightly spoken.
19:38 willthechill we have no idea what is causing appveyor to hang right now
19:38 willthechill all you did was log in once
19:39 willthechill we don't know why that particular file is hanging after cc1plus finishes
19:39 willthechill yes there are exception handling issues
19:39 willthechill no, we don't know if that is what is causing appveyor to hang
19:39 willthechill need more info
19:41 mako Ok, I should avoid saying exception handling is the culprit without rock solid proof. Sorry, my bad.
19:44 willthechill yes we really need more info from inside appveyor to know why lib/RPerl/Test/Conditional/program_05_good.pl is hanging
19:44 willthechill obviously it does not hang in Linux!
19:44 travis-ci RPerl build passed. Will Braswell says 'Learning RPerl, Add Roadrunner Engraving Cover Image'
19:44 travis-ci https://travis-ci.org/wbraswell/rperl/builds/155622870 https://github.com/wbraswell/rperl/compare/37c4ba84754b...01ce0c4d792c
19:45 willthechill I guess it doesn't hang on your Windows box either, if you are passing all RPerl tests
19:45 willthechill so there is something specific about appveyor which is causing it to hang?
19:49 mako Many issues. I noticed Appveyor builds are running under PowerShell. That's different compared to my system. Doing 'make test' under PowerShell instead of usual cmd.exe...
19:53 willthechill okay
19:55 mako Hm, 04_type_scalar.t, hanging at 5/310 :-(
19:57 willthechill is that the same file hanging as appveyor, or a different one?
19:59 mako A different one. Trying 'make test' (really 'dmake test') under cmd.exe...
19:59 willthechill hmmm
19:59 willthechill the mystery grows deeper
19:59 willthechill :-P
20:05 mako Test, 04_type_scalar.t, hanging at 113/310 under cmd.exe
20:09 willthechill how did we get from hanging on 5/310 to 113/310?
20:10 mako Good question. I just changed the command interpreter.
20:11 willthechill wow
20:12 mako I should consider 'cygwin'.
20:12 willthechill I don't know what to tell you about that one
20:13 willthechill I'm sure bulk88 could tell us if he were here
20:14 mako Ah bulk88. What is he doing anyway? Looking for him, but he's not available.
20:15 willthechill nobody knows!
20:15 willthechill I see him in person once a year if I'm lucky
20:15 willthechill otherwise we just hope and pray to see him online
20:16 willthechill rurban is even more rare
20:16 willthechill and ingy is, like, dead or in limbo or something
20:18 mako The availability of your team, captain. :-D Do they have something better to do? You don't know, do you?
20:21 willthechill oh those aren't my team members
20:21 willthechill rurban and ingy are the co-founders of Perl 11, along with myself
20:21 willthechill bulk88 is the only fully-authorized RPerl contributor on that short list
20:22 willthechill he lives a life of purposeful anonymity and obscurity
20:22 willthechill this is quite common within the Perl community
20:26 mako Yeah, I have already seen that bulk88 is very anonymised. Ok, his decision. Why not. But would be nice to reach him more easily when he's needed.
20:26 willthechill I know him in real life, I can text him if there is an emergency
20:27 mako Ah, ok.
20:28 willthechill but texting doesn't solve our complex coding problems, haha!
20:28 willthechill okay I've got to get going, will be away from the computer for several hours
20:28 willthechill good job, will talk to you soon!
20:31 mako ok, could you please keep in mind I want you to ask something about '--CXX' of 'script/rperl'? Not now, just do your stuff. We'll talk later. No deal.
20:32 willthechill okay
20:32 willthechill *away from keyboard*
22:05 mako joined #perl11
22:43 mako joined #perl11

| Channels | #perl11 index | Today | | Search | Google Search | Plain-Text | summary