Camelia, the Perl 6 bug

IRC log for #mojo, 2013-09-08

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

All times shown according to UTC.

Time Nick Message
00:02 mst joined #mojo
00:05 dsteinbrunner joined #mojo
00:08 asarch joined #mojo
00:44 jberger omfg! ack.vim!!!
00:47 jamesw a rake project, heh
00:48 zacts ack.vim?
00:48 zacts the grep tool?
00:48 zacts what does it do for vim?
00:49 btyler there's a plugin which lets you ack from the vim command bar and (I think?) sticks the results into a buffer
00:49 jberger jamesw, as I installed it through VAM I didn't even notice, that is a little funny
00:49 zacts (history note: there was also an ancient C compiler for minix named the Amsterdam Compiler Kit)
00:50 btyler personally I stick to vim-in-terminal + tmux, so my pattern tends to be ctrl-z/switch tmux window-->ack for stuff->fg/switch tmux window
00:50 jberger zacts, you know when you need to make that little tweak to everything that will be affected by that little change :-)
00:50 zacts I use vim + tmux also
00:50 zacts :-)
00:51 jamesw tmux + nvi + map heh
00:51 jberger I can see how vim+tmux would be useful, but as I'm only just learning vim, I think I will stick to this for now :-)
00:51 btyler jberger: for sure, learning both at once would be a headache :)
00:51 mst lots of separate copies of $vi plus job control is a wonderful thing
00:52 jberger indeed, but first things first
00:53 * jberger recovering gedit user
00:54 jberger my terminal is a tabbed terminal anyway, so I can get some of those niceties that way
00:54 zacts jberger: taking the time to learn vi/vim is well worth it
00:54 jberger IIRC mst uses some crazy editor of his own design
00:55 jberger zacts, I'm starting to believe that
00:55 zacts darkburn is my favorite vim theme
00:56 jberger zacts, you like it better than solarized dark?
00:56 * sri can't stand solarized
00:56 zacts jberger: zenburn is supossedly designed for long coding sessions
00:57 zacts it's too bright for me
00:57 zacts so I use darkburn
00:57 zacts and it works well for me
00:57 * sri likes twilight and monokai
00:57 jberger sri: I was just about to ask
00:57 jamesw i thought it was just traditional vi
00:57 sri didn't i link a screenshot earlier?
00:58 btyler oh, haven't seen that one, going to check it out. I love inkpot's terminal version: http://i.imgur.com/oTrT0pr.png
00:58 btyler (darkburn, that is)
00:58 jberger yeah, but you didn't say what it was called
00:58 dsteinbrunner joined #mojo
00:59 sri http://twitpic.com/8dzse4/full # have not used macvim in ages though
01:00 sri http://twitpic.com/b6xjq7/full # my textmate looks almost the same though... i so love the search in project feature!
01:01 jberger ack.vim is probably somewhat like a "search in project"
01:01 sri i never liked it much
01:01 sri guess i'm too spoiled
01:02 jberger hehe
01:02 sri same reason i can't use sublime text
01:02 mst jberger: jamesw is correct, I use ex-vi (albeit a somewhat patched version)
01:02 jberger I feel pretty spoiled right now, then again, gedit isn't quite what you would call a full featured editor
01:04 mst basically, my key reason is: if I typo a command in vim, it almost always turns out that the typo is also a valid command and vim does something completely unexpected
01:04 mst if I typo a command in ex-vi, it beeps.
01:05 sri http://i.imgur.com/ShllRtC.png # this is what my textmate setup looks like now
01:05 sri twilight theme and consolas font
01:05 btyler mst: heh, yeah, runaway vim happens probably once every few weeks to me. still, I've grown dependent on vim's undo history being a proper tree that goes back months...couldn't imagine only having one undo level (or even linear undo)
01:06 mst btyler: right. I use :w and 'git commit' to provide that
01:07 mst since an in-editor undo tree becomes much less helpful the first time you rebase or merge the branch you're working on
01:07 zacts mst: based on nvi?
01:07 btyler certainly true, and it has bitten me once or twice
01:07 mst zacts: no, based on ex-vi
01:08 zacts oh ok
01:08 mst zacts: the thing that nvi, elvis and vim are all 'improved clones' of
01:08 zacts so the original vi from 4.3 BSD?
01:08 zacts kind of thing?
01:08 mst yes.
01:08 zacts oh cool! \o/
01:09 jamesw i like my multiple undo and windows
01:09 zacts to each his/her own
01:09 jamesw but i like having both multiple undo and undo-redo
01:10 mst there's a comment in it from Bill Joy apologising for using slightly more core than ed.
01:10 mst and the fact it's 'visual interface for ex' is very clear
01:10 zacts (although, it would be cool to be able to say "I have my own port of the original ex-vi that I use for my editor")
01:10 mst the vi part is basically implemented as
01:11 mst BEGIN { if ($0 eq 'vi') { run_command('vi'); exit; } }
01:11 zacts btw where can you get the source for 4.3BSD utilities like ex-vi these days?
01:11 mst http://ex-vi.sf.net/ is where mine started from
01:11 zacts oh cool
01:12 jamesw probably tuhs
01:14 zacts found this: http://www.mckusick.com/csrg/
01:14 jamesw http://www.tuhs.org/Archive​/4BSD/Distributions/4.3BSD/
01:15 zacts neat
01:16 mst right, the sf.net version has been updated somewhat
01:16 mst slightly less archeological value, lots more 'actually builds on modern systems'
01:17 btyler re the Bill Joy comment...that is really dang cool
01:19 mst he basically takes the address of the base of the heap on startup
01:20 mst and uses it directly as an array of pointers into the swap file
01:20 mst I am not a particularly good C programmer, so I was really quite excited when I finally worked out wtf he was doing
01:20 * jberger is impressed
01:23 basic6_ joined #mojo
01:24 jberger sri: does mojo_highlight_data work for you?
01:25 sri last ime i tried it did
01:28 btyler jberger: I gave mojo.vim another shot today and didn't find it doing anything, but assumed something in my setup was conflicting somehow
01:28 jberger btyler it works on my template files, just not on my embedded templates
01:29 btyler yeah, I was only trying it with lite apps, didn't think to check a full one
01:29 sri wasn't there a config setting for DATA templates or so
01:30 mst oh. yeah. another useful feature of ex-vi for me: no syntax hilighting
01:30 sri jberger: ":let mojo_highlight_data = 1"
01:30 mst I think I have at least three lines in my .vimrc now dedicated to making sure that stays off :)
01:30 jberger sri: correct, but it isn't working for me
01:54 bluescreen joined #mojo
01:55 bluescreen_ joined #mojo
03:24 preflex_ joined #mojo
03:42 d4rkie joined #mojo
03:42 d4rkie joined #mojo
06:39 mire joined #mojo
06:42 dotandimet joined #mojo
06:51 d4rkie joined #mojo
07:11 mire joined #mojo
07:16 dotandimet joined #mojo
07:20 connor_goodwolf joined #mojo
07:31 punter joined #mojo
07:33 jamesw joined #mojo
07:37 ryozi joined #mojo
08:05 sh4 joined #mojo
08:08 arpadszasz joined #mojo
08:14 dod joined #mojo
08:18 dod joined #mojo
08:36 Vandal joined #mojo
08:38 denis_boyun joined #mojo
08:49 batman sri: i did think about it, and i think it's the correct behavior (wait() dieing)
08:49 batman uhm.. "dieing" ? :P
08:52 ladnaV joined #mojo
09:04 Vandal joined #mojo
09:05 dotan joined #mojo
09:21 basiliscos joined #mojo
09:24 trone joined #mojo
09:35 vividsnow joined #mojo
10:08 navi joined #mojo
10:09 dqw joined #mojo
10:28 dotan joined #mojo
10:31 sh4 joined #mojo
10:31 jberger joined #mojo
10:35 shmuel joined #mojo
10:47 bowtie joined #mojo
10:49 mire_ joined #mojo
11:01 preflex joined #mojo
11:14 jberger joined #mojo
11:55 shmuel joined #mojo
11:56 dqw joined #mojo
12:05 dod joined #mojo
12:19 denis_boyun joined #mojo
12:21 espent joined #mojo
12:47 sri bad batman! that behavior is wrong
13:47 gryphon joined #mojo
14:01 sri https://github.com/kraih/mojo/commit/64c758c643dd9​869910bd356b6c4ad0ec8e230d0#commitcomment-4040081
14:01 sri wonder what that comment means
14:01 sri old kernel + too new libc header files?
14:07 asarch joined #mojo
14:24 basiliscos joined #mojo
14:39 shmuel joined #mojo
14:59 d4rkie joined #mojo
15:31 asarch joined #mojo
15:39 cstamas sri: i did some testing with mojolicious on debian testing and the test suite seems to be failing
15:40 cstamas sri: http://paste.debian.net/36962/
15:40 cstamas sri: it can be that i overlooked something
15:40 cstamas sri: it would be nice if you can take a look
15:40 cstamas thx
15:40 sri is it the test mentioned in the comment above?
15:40 cstamas sri: forgot to mention that this is Perl 5.18
15:41 cstamas t/mojo/app.t
15:41 cstamas line 467
15:41 sri looks like kernel/libc headers version mismatch
15:42 sri umm, that test has no line 467
15:42 cstamas sri: here http://paste.debian.net/36962/ line 467
15:43 sri kernel version?
15:43 cstamas 3.2.0-3-amd64
15:44 sri yea, version mismatch with the libc headers
15:44 sri headers define SO_REUSEPORT, but the kernel doesn't support it
15:44 cstamas aha
15:44 sri debian screwup i suppose, not sure what we could do about it
15:45 cstamas sri: this is my local environment so not debian in general
15:46 sri so you installed a custom kernel/libc headers?
15:46 cstamas i am building/testing inside a chroot so yeah what you say can be a problem on my side
15:46 sri meh
15:48 cstamas thx for the info, i will try again later, after cleaning up my env.
15:50 sri cstamas: can you install IO::Socket::IP?
15:51 sri it appears to use the exact same way to probe for SO_REUSEPORT support
15:51 sri cstamas: also, was that you? https://github.com/kraih/mojo/commit/64c758c643dd9​869910bd356b6c4ad0ec8e230d0#commitcomment-4040081
15:52 cstamas sri: no, not me
15:52 sri ouch
15:52 cstamas sri: but seems to be the same issue
15:58 cstamas sri: installing IO::Socket::IP does not help
15:58 sri but do the test pass?
15:58 cstamas no
15:58 sri good
15:59 arpadszasz joined #mojo
15:59 cstamas maybe i misunderstood it
15:59 cstamas i installed IO::Socket::IP and tried to do the mojo test with that installed
16:00 sri doesn't matter
16:00 cstamas i guess you ment to run the IO::Socket::IP test suite, is that right?
16:00 sri i just wanted to know if IO::Socket::IP passes its tests
16:00 cstamas i will run the tests... stay tuned
16:00 sri since it uses the same code to probe for SO_REUSEPORT support
16:01 sri there is no way this test could pass http://api.metacpan.org/source/PEVA​NS/IO-Socket-IP-0.22/t/11sockopts.t
16:02 cstamas okay, it failed the same way
16:02 sri your box is broken, so much is certain
16:02 cstamas "Cannot socket() - Protocol not available at t/11sockopts.t line 32."
16:03 sri not sure if we should disable the test by default
16:03 cstamas thx, i will try to fix it...
16:09 sri i've also poked LeoNerd, maybe he can come up with a more reliable way to probe for support
16:10 sri version mismatch is annoying, we had the opposite before
16:10 cstamas ok, please ping me if you have an update on this
16:10 cstamas i will try to do some work on my side too and report back
16:10 sri doubt we can do much though
16:11 sri only solution i see is disabling the test by default if the problem is widespread enough
16:44 tpeiffer joined #mojo
16:47 mst sri: or make the detection a bit more aggressive
16:47 mst sri: i.e. have it trap the 'kernel says no' error and bail out the test as skip_all
16:55 tpeiffer I am interestted in a little discussion on the plugin ttrenderer and getting variables seen within an action space exported to a named template
16:56 tpeiffer I have used template toolkit once before, a few years ago and I honestly can't seem to figure out the namespace.
17:10 dod joined #mojo
17:13 cstamas sri: kernel 3.9 is needed
17:13 cstamas sri: after the kernel upgrade the tests pass
17:13 mst sri: #debian-perl is still discussing how to do the probe
17:22 mst sri: I'll get leonerd to figure out a better probe and then you can steal that - sane?
17:23 sri mst: i guess more aggressive probing would work for me
17:23 sri not sure it would for leonerd
17:24 mst I'm already discussing it with him; one of us will get back to you once we have a solution
17:33 good_news_everyone joined #mojo
17:33 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/dYXtkw
17:33 good_news_everyone mojo/master ced064d Sebastian Riedel: fixed portability bug in SO_REUSEPORT tests
17:33 good_news_everyone left #mojo
17:33 sri cstamas: can you test that?
17:34 sri mst: that one seems to work on os x back to 5.10.1
17:35 sri lets wait for travis
17:35 sri https://travis-ci.org/kraih/mojo/builds/11126550
17:35 sri also pass
17:35 mst sri: the 'correct' answer seems to be to setsockopt and trap the ENOPROTOOPT error
17:36 sri that does sound more correct
17:38 mst sri: which I'm trying to get leonerd to trial in IO::Socket::IP
17:38 mst then we can percolate it back up
17:38 sri sure, don't think i need something more correct though
17:40 sri do i hate this libc header version mismatch mess
17:40 sri last time it was new kernel and old headers
17:43 arpadszasz joined #mojo
17:55 good_news_everyone joined #mojo
17:55 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/pnVBDA
17:55 good_news_everyone mojo/master 27e0949 Sebastian Riedel: small optimizations
17:55 good_news_everyone left #mojo
17:55 wk joined #mojo
18:01 btyler joined #mojo
18:02 espent joined #mojo
18:10 cstamas sri: seems that it works for me \o/
18:10 sri \o/
18:11 cstamas my battery is almost empty, bbl
18:22 trone mojo::ioloop + useragent in mojo app are AMAZING!
18:22 trone mojo++
18:52 dotandimet joined #mojo
18:58 punter joined #mojo
19:28 denis_boyun joined #mojo
19:29 marcus trone: \o/
20:10 jberger d==
20:18 jberger_ joined #mojo
20:47 rem_lex|pivo joined #mojo
20:56 good_news_everyone joined #mojo
20:56 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/tH-j0A
20:56 good_news_everyone mojo/master f3ad746 Sebastian Riedel: fixed Mojo::IOLoop::Server formatting
20:56 good_news_everyone left #mojo
20:57 good_news_everyone joined #mojo
20:57 good_news_everyone [mojo] kraih tagged v4.34 at e57a04d: http://git.io/G0VGDQ
20:57 good_news_everyone left #mojo
21:00 sri ember has an interesting release process http://emberjs.com/blog/2013/09/0​6/new-ember-release-process.html
21:00 sri every day an unstable release, every week a beta release and every 6 weeks a stable release
21:08 bd so basically they call the last commit each day "unstable" *scnr*
21:28 moltar joined #mojo
21:39 mire joined #mojo
21:46 sri \o\
21:46 sri /o/
21:51 punter Do you think the Mojolicious part of this file is programmed completely wrongly? https://github.com/akarelas/KJabberd/blob/​master/lib/KJabberd/Listener/WebSocket.pm
21:52 punter In particular, I'm worried that the bareword "app" will always refer to the same app, regardless of how many Moo objects I have created (and that will create a problem)
21:53 punter (in line 48)
21:54 punter I think I should do it properly
21:54 trone I working on a sub runs parallel ten $ua->get($url) (as http://mojolicio.us/perldoc/Mojolicious/G​uides/Cookbook#Parallel_blocking_requests), but at runtime when perl get $delay->wait it says "Mojo::IOLoop already running". What I can do to return my results from sub?
21:55 trone (I would the ua runs parallel, but I have to manage all them when they finished to return the merged results.
21:56 punter (properly = OO-style & without DSL)
21:56 sri punter: if you're worried about it, don't do it
21:56 punter what do you mean?
21:57 sri my $app = Mojolicious->new; $app->routes->websocket('/' => sub {...});
21:57 punter excellent, thanks!
21:57 punter That's what I should do, right?
21:57 punter thxz
22:01 sri i have no idea what happens when you embed a lite app like that
22:14 punter fixed: https://github.com/akarelas/KJabberd/blob/​master/lib/KJabberd/Listener/WebSocket.pm
22:22 good_news_everyone joined #mojo
22:22 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/2FmwiQ
22:22 good_news_everyone mojo/master 7f8eaf1 Sebastian Riedel: better examples for parallel blocking requests
22:22 good_news_everyone left #mojo
22:24 dotandimet joined #mojo
22:30 jberger joined #mojo
22:35 good_news_everyone joined #mojo
22:35 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/R1rgqg
22:35 good_news_everyone mojo/master 94dc1cb Sebastian Riedel: added section for spin-off projects
22:35 good_news_everyone left #mojo
22:47 zacts for small to medium sized sites, do most people usually keep <html> templates in the __DATA__ section, or in external files?
23:01 good_news_everyone joined #mojo
23:01 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/S2814A
23:01 good_news_everyone mojo/master 3588dd3 Sebastian Riedel: documentation tweaks
23:01 good_news_everyone left #mojo
23:15 dvinciguerra joined #mojo
23:59 Adura joined #mojo

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