Camelia, the Perl 6 bug

IRC log for #padre, 2010-08-01

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

All times shown according to UTC.

Time Nick Message
00:17 teejay joined #padre
01:32 CSJewell2 joined #padre
03:52 |Sno| joined #padre
03:59 [Sno] joined #padre
04:10 kanishka joined #padre
05:09 garu szabgab: created, and added to TPF's wiki => http://perl.org.br/Artigos/RakudoStar
05:09 szabgab garu: nice
05:10 szabgab now I expect all the .br perl hackers to start showing their capabilities to the international perl community by writing nice  things in perl 6 :)
05:11 garu szabgab: well, there already are two awesome brazilian perl hackers (ruoso and fglock) that have been doing active perl 6 development and promotion for several years now ;)
05:12 garu fglock was even responsible for v6.pm if I'm not mistaken
05:13 garu I remember watching a talk about PASM back in 2002 in a free software conference here in Rio
05:13 garu and Linux Maganize had a 6 page article about Parrot last month
05:14 garu maybe we can get them to talk about rakudo as well
05:20 szbalint szabgab: hey, are you already in Pisa?
05:39 szabgab szbalint: not yet
05:39 szabgab my flight is tomorrow afternoon
05:40 szbalint alright. See you there then :)
06:15 Alias__ joined #padre
06:31 azawawi joined #padre
06:45 szabgab garu: still here?
06:46 szabgab see this: http://www.pm.org/groups/south_america.html
06:46 kaare joined #padre
06:58 Hyppolit svn: r12119 | adamk++ | http://padre.perlide.org/trac/changeset/12119
06:58 Hyppolit Adding leading "Do not edit" warning, and trailing copywrite statement.
06:58 Hyppolit trunk/Padre-Plugin-FormBuilder/ trunk/Padre-Plugin-FormBuilder​/lib/Padre/Plugin/FormBuilder/
07:12 Hyppolit #979: Reload/Refresh Files dialog only ever gives you the option to close the selected file. (new defect) [ http://padre.perlide.org/trac/ticket/979 ]
07:12 Hyppolit #979: Reload/Refresh Files dialog only ever gives you the option to close the selected file. (assigned defect) [ http://padre.perlide.org/trac/ticket/979#comment:1 ]
07:14 azawawi joined #padre
07:14 azawawi CSJewell: ping
07:19 CSJewell pong. What's up, azawawi?
07:19 azawawi finally figured out the problem... :)
07:19 CSJewell Great. What was it?
07:19 azawawi PL_origalen = 1 .... // not being extern-ed...
07:20 CSJewell On the perl5 end, or the perl6 end?
07:20 azawawi perl5...
07:20 azawawi adding 'extern "C" PL_origalen;' solves it
07:21 CSJewell Ouch. So we're going to have to wait until 5.12.2 at a minimum, if not 5.14.0 before fixing the issue...
07:21 azawawi but still it (blizkost) does not evaluate right.
07:21 * azawawi debugs
07:22 azawawi but it is a step forward :)
07:23 danlucraft joined #padre
07:23 CSJewell Keep working on it, and if there's anything I can help out with, let me know.  And send a patch to perl5's 'blead' (the 5.13.4-to-be), please... I'll let them know that it fixes a problem, and ask if it can get into 5.12.2.
07:24 azawawi maybe parrot and perl overwrite each other's definitions
07:24 CSJewell It's possible.
07:24 azawawi since i see a lot of undef-s before including perl...
07:25 azawawi jnthn++ :)
07:25 CSJewell Also, you do know that Strawberry builds perl5 with the 'MULTIPLICITY' define.  That changes a LOT of definitions in perl5.
07:26 CSJewell (Without MULTIPLICITY, a lot of variables are global. With it, they're per-interpreter.)
07:26 CSJewell If Parrot's overwriting definitions, the problem may be on the p6 end, then...
07:27 szabgab azawawi: any chance for an announcement about Rakudo in Arabic?
07:30 azawawi sure
07:30 azawawi once i finish with this bug... :)
07:34 CSJewell azawawi: http://gist.github.com/503062 is where I'm referring to for Strawberry's build options... they come from perl/lib/CORE/config.h .
07:34 CSJewell That's 64-bit, but 32-bit I think is similar.
07:36 CSJewell azawawi: I have to ask: Has parrot been checked on other systems where the archname =~ m/-multi-thread\z/?
07:36 azawawi CSJewell: another thing http://perldoc.perl.org/perlembed.html, say PL_origalen = 1
07:37 CSJewell (-multi for MULTIPLICITY, -thread for USE_ITHREADS?)
07:37 azawawi CSJewell: looking at the latest, i think it should have the I prefix...
07:39 garu szabgab: I see it
07:39 garu hi azawawi! hi CSJewell :)
07:39 CSJewell The I prefix (as far as I understand) is a way of telling the C preprocessor that this needs to be a per-interpreter variable, and is really #defined to an access to the interpreter's copy of the variable in MULTIPLICITY situations.
07:40 CSJewell hello, garu.
07:40 azawawi garu: hi there
07:40 CSJewell So maybe you're accessing the global version of the variable, and not the per-interpreter one?
07:41 garu szabgab: what about it? :)
07:42 CSJewell (I'm throwing out anchors based on what limited knowledge I have/can read.)
07:44 azawawi CSJewell: im reading the code to understand it better... im doing the same as you... been a while since i did some C code :)
07:51 CSJewell The things Perl has to do in order to compile on half the universe.
07:51 azawawi yeah :)
08:08 waxhead_ joined #padre
08:12 [1]Casan joined #padre
08:42 Steffen joined #padre
08:43 szabgab garu: it is now a real interactive map and not the static one we have on the main page of pm.org
09:00 garu szabgab: yeah, very nice! But... this is so a while now, isn't it? I mean, after you click on the map element, it becomes interactive
09:00 garu unless I'm missing something.. I really thought this feature was already implemented at least a couple of months ago :)
09:01 garu wow, 6am... I really gotta get some sleep
09:02 garu g'night!
09:03 Sewi gn8
09:18 danlucraft joined #padre
09:39 waxhead_ szabgab++ Perl 6 Hashes screencast
09:43 waxhead szabgab, wow.. that's some cool stuff in perl 6...
09:43 waxhead so the next part for the screen cast will be operations on the hash, ie sorting, sorting on values etc.. ? :)
09:59 [Sno] joined #padre
10:03 LeK joined #padre
11:34 azawawi joined #padre
11:43 azawawi left #padre
11:48 kanishka joined #padre
11:49 danlucraft joined #padre
11:53 danlucraft1 joined #padre
12:17 kanishka joined #padre
13:01 chorny joined #padre
13:05 danlucraft joined #padre
13:42 kanishka joined #padre
13:52 Hyppolit svn: r12120 | adamk++ | http://padre.perlide.org/trac/changeset/12120
13:52 Hyppolit Bumping dependency
13:52 Hyppolit trunk/Padre-Plugin-FormBuilder/
14:21 Hyppolit svn: r12121 | adamk++ | http://padre.perlide.org/trac/changeset/12121
14:21 Hyppolit Part two of the ConfigSync refactoring.
14:21 Hyppolit Applied a variety of changes of FBP and FBP::Perl to get better working FBP dialog code.
14:21 Hyppolit The tools.sync action has been switched over to use Padre::Wx::Dialog::Sync2.
14:21 Hyppolit You can click the login button and have it fail without crashing now :)
14:22 Hyppolit trunk/Padre/ trunk/Padre/lib/Padre/Wx/ trunk/Padre/lib/Padre/Wx/Dialog/ trunk/Padre/lib/Padre/Wx/FBP/
14:22 Alias__ Shazam!
14:24 Steffen One of these days, I'll have to get back to Padre and prototype some crazy shit for your to refactor. :P
14:26 Alias__ More the opposite I suspect
14:27 Alias__ Most of your nice features in the threading code are gone
14:27 Alias__ Throttling, dynamic worker spawning
14:27 Alias__ And so on
14:27 Alias__ I had to cut deeply to maintain momentum
14:30 Steffen The throttling was simple enough.
14:31 Steffen I always wanted to play with the swarm stuff.
14:31 Alias__ I had to cut that too :(
14:31 Alias__ I'm working back towards it
14:46 Alias__ Steffen: Once some of the server-mode stuff starts working better, I'm sure I can tempt you back :)
14:49 Hyppolit svn: r12122 | adamk++ | http://padre.perlide.org/trac/changeset/12122
14:49 Hyppolit Added ->status to ::Main to allow rapid transient messages to be presented in the status bar, as fast as 100/sec (ADAMK)
14:49 Hyppolit trunk/Padre/ trunk/Padre/lib/Padre/Wx/
14:50 Steffen In a world without other responsibilities, I'd like to tackle the server.
14:51 Alias__ The swarm server, or the rest?
14:51 Alias__ Madre-Sync was built for us by the university project mattp did
14:51 Alias__ Although it's being tidied up a lot now we're merging it
14:54 Steffen "the rest" being sync?
14:54 Steffen And yeah: Swarm.
14:55 Alias__ Sync, Telemetry and Popularity
14:55 Alias__ The bits and pieces
14:56 Steffen The popularity stuff is still annoying.
14:56 Alias__ Annoying how so?
14:56 Alias__ It's a plugin you don't need to turn on
14:56 Steffen Sure it's off by default?
14:56 Steffen Wait.
14:56 Alias__ Absolutely
14:56 Steffen Different thing.
14:56 Alias__ You mean the survey?
14:56 Steffen I meant the "how did you learn about... " exactly.
14:56 Alias__ I don't much like that either
14:57 Alias__ I just added the "Nth" subsystem
14:57 Alias__ To centralise the logic about shit we annoy people with at startup
14:57 Alias__ The survey now only appears on the third startup
14:57 Steffen Annoying nonetheless.
14:57 Alias__ And I'll be layering in the other stuff after maybe the 10th and 100th startup
14:58 Alias__ Well, if you turn on Sync it shares your startup count
14:58 Alias__ So they'll never appear
14:58 Alias__ State aggregates across multiple computers
14:58 Alias__ So, in theory, you start Padre on a new machine the first time
14:58 Alias__ It doesn't say anything at all
14:58 Alias__ Then you go to Preferences Sync and login
14:59 Alias__ And it knows you've used Padre 500 times, and knows what shit not to bug you about ever again
14:59 Steffen Have you thought about the implications of using Sync with multiple versions of Padre?
14:59 Alias__ A little
14:59 Alias__ Not properly
14:59 Steffen *nod*
14:59 Steffen It's hard.
14:59 Alias__ Not necesarily
15:00 Steffen In the end, it's going to need an API version thing.
15:00 Alias__ Yup
15:00 Steffen And some kind of schema migration.
15:00 Alias__ The hooks are in place already
15:00 Alias__ I have the solution sitting at work
15:00 Alias__ Our database migration system kicks the ass of anything else I've ever seen
15:00 Alias__ ORLite::Migrate is a super-light crippled copy of it
15:00 Alias__ That only goes in one direction
15:01 Alias__ Our work one can migrate forwards or backwards for over 2 years
15:01 Alias__ And the design supports branching
15:01 Alias__ So if you migrate from checkout Foo to checkout Bar, it will roll the schema back 3 months to when you first branched, then roll forwards down the branch
15:01 Alias__ And so on
15:02 Alias__ All the same concepts will work for configuration
15:03 Alias__ The question is really moot until Sync lives up to it's promise
15:03 Alias__ And then I'll probably be forced to implement it :)
15:03 Alias__ Or just let it run and see what happens and if/when it explodes :)
15:04 * Alias__ bed
15:07 kanishka joined #padre
15:10 Hyppolit svn: r12123 | adamk++ | http://padre.perlide.org/trac/changeset/12123
15:10 Hyppolit Add magic "STATUS" message that lets any task say something in the main thread's statusbar with trivial effort.
15:10 Hyppolit trunk/Padre/lib/Padre/
15:10 Steffen Cheap solution: Support forward migration only and let the user have a separate configuration for each version.
15:11 Steffen following the idea that people occasionally upgrade but don't downgrade.
15:11 Steffen No, maybe not.
15:11 Alias__ "You have an existing configuration for a new version of Padre. Would you like to fork a version-specific config?"
15:11 Steffen Are you going to ask that on every startup?
15:12 Steffen If not, you don't get configuration changes propagated.
15:12 Alias__ Of course now, once they say they are going to fork, we update their settings and then it won't be a problem
15:12 Alias__ Of course not...
15:12 Alias__ We just need the server to support more than one config per user
15:12 Alias__ Hell, you could probably have ranges
15:12 Steffen Yes, for sure.
15:12 Alias__ The current config system is round-trip safe
15:13 Alias__ It supports unknown keys without breaking them, I think
15:13 Steffen *nod*
15:13 Steffen Anyway. I have to get back to preparing my YAPC talk.
15:13 Alias__ Which would allow a kind of $BACKCOMPATIBLE mechanism like we're using for plugins now
15:13 * Alias__ bed for real this time
15:13 Steffen It's still in a bad shape.
15:27 CSJewell2 joined #padre
16:07 kanishka joined #padre
16:42 kanishka1 joined #padre
17:54 CSJewell joined #padre
18:30 kanishka joined #padre
19:09 shadowpaste joined #padre
20:42 GabrielVieira joined #padre
22:47 rbaumer joined #padre
23:43 LeK left #padre

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