Parrot 4.0.0 "Hyperstasis" | parrot.org | Log: irclog.perlgeek.de/parrot | #parrotsketch meeting Tuesday 19:30 UTC
Set by moderator on 18 January 2012.
00:06 whiteknight joined
whiteknight good evening, #parrot 00:07
00:56 kurahaupo joined 01:16 hernan joined 01:17 hernan joined 01:18 hernan joined 01:19 hernan joined 01:20 hernan joined
whiteknight hmmm. Something is weird with parrothist.pod. When I try to read it in I get all sorts of encoding errors and stuff 01:20
01:21 hernan joined 01:22 hernan joined
bacek_at_work whiteknight, all good on my box. 01:32
whiteknight must be my copy is messed up
ah, it was iso-8859 and I was trying to read it as utf-8 01:43
dalek sella: 3925a59 | Whiteknight++ | src/unstable/reflect/ (4 files):
auto-cache reflectors on thier target objects with properties. Misc other cleanups
02:01
sella: 46e13c3 | Whiteknight++ | src/unstable/assert/ (4 files):
Several cleanups and updates for Assert
sella: 22c8af8 | Whiteknight++ | src/date/Date.winxed:
fixes for Date. The day coming out of the decodetime opcode is off by one from the rest of my algorithms, so fix that. Make sure to update the weekday when we add days
sella: db1d2c6 | Whiteknight++ | src/date/DateFormatter.winxed:
Add MMMM and MMM specifiers for DateFormatter
sella: 5b63ae4 | Whiteknight++ | / (6 files):
Add in a new parrot_release utility that will help to automate the release process. Still needs work/debugging
sella: 84de03f | Whiteknight++ | src/utilities/parrot_release.winxed:
Fix a few issues with parrot_release utility
sella: 0fe13a0 | Whiteknight++ | s (2 files):
fix Assert compilation error, and add it back into the build
02:04
rrot: be03965 | Whiteknight++ | / (2 files):
Convert parrothist.pod from iso-8859-1 to utf-8, for ease
02:05
rrot: e141fad | Whiteknight++ | tools/release/release.json:
Fix a typo in json that I shouldn't have committed
rrot: 86d1c01 | Whiteknight++ | / (37 files):
Merge branch 'master' of github.com:parrot/parrot
03:05 davidfetter joined 03:38 Psyche^ joined 07:01 jsut joined 07:20 woosley joined
alvis mortiz++ # for getting to submissions.pod before I could! 08:27
but i'll want to fixup some bits *after* i figure out how to get 'parrotbug' to create new issues without jeopardizing our credentials. 08:28
08:50 mj41 joined 08:54 seme joined
moritz I think there's really just one good solution: a proxy with a separate account 08:59
alvis moritz: you're probably right, but cotto mentioned taking a look at using git and the local machine, so i'm checking on that as an option. 09:19
make that possible (but not probable) alternative. 09:20
cotto alvis, don't spend too much effort 09:24
(on using the local machine's git credentials)
moritz one can check github.user and github.token in ~/.gitconfig 09:26
09:27 lucian joined
alvis cotto: ok. i'll give myself through the weekend to see if i can come up with something. 09:28
cotto alvis, wfm 09:29
alvis moritz: yes, but ... that won't really work for someone without a github account. 09:33
i *think* the idea of parrotbug is to allow just about anyone to submit a bug report ... 09:34
even if he or she doesn't use github.
cotto yes 09:35
I was going to raise the same objection
one (probably minor) issue is that github's tos say that one human may only have a single account 09:36
alvis, we now have a parrotbug repo and a parrotbug team, which has push/pull access to the repo, and a parrotbug user who's on the parrotbug team 09:43
alvis cotto: oh, wow. ok, then. 09:44
cotto they don't do/contain anything yet 09:45
alvis cotto: cool. that'll help with testing. :) 09:47
cotto alvis, how solid an idea do you have of how the proxy account would work?
alvis cotto: hmm, well, i'm fairly satisfied with the api, but beyond that ... i'm speculating. ... 09:49
cotto: all of the pieces are there; i've just never tried to put 'em together to say what the docs say will work.
moritz likes Net::GitHub, fwiw 09:50
cotto alvis, ok. Can you spell out what the workflow will look like, as far as you understand it?
moritz it's what I used for scraping the gh issues for the track URLs
s/track/trac/
alvis cotto: give me a few minutes, and i will. back in a few 09:52
cotto alvis, no rush 09:53
I'm starting to see how this could work. 09:59
I'm out for the night. 'night 10:11
alvis cotto: 'k, 'night. 10:19
msg cotto please see gist.github.com/1639268 10:20
aloha OK. I'll deliver the message.
moritz that sounds... dubious
what about 10:21
1) have a mail box for bugs@parrot.org
2) have a cron job that fetches new mails for it
3) if a new mail looks like ticket, the job turns the email into a github issue 10:22
*looks like a new ticket
that way you only need to store the parrotbug token on the server where the cron job runs
alvis mortiz: sorry, i was away. i thought of something like that. ... 10:30
moritz: i asked staff about a project email account, but they said no go. i didn't think of a using a parrot.org email account. <chagrin> 10:31
moritz: btw, which part sounds particularly dubious? the how to catch and process the actual issue in the parrot repo? 10:34
moritz "catch and process"? 10:35
you just open the ticket in the parrot repo
and the rest is all done by us poor humans
alvis well, what i mean is, the proxy throws an open issue event to the parrot repo ...
but then the parrot repo has "catch" and open the issue. that's the part i'm uncertaion about. 10:36
moritz it doesn't throw an event. It simply does a JSON API call that opens the ticket; it includes the subject, body and so on
search.cpan.org/~fayland/Net-GitHub...eate_issue 10:38
alvis moritz: take a look at github.com/blog/964-all-of-the-hooks.
moritz alvis: hooks are not everything; you can also use the (synchronous) API, which is easier if you are the one that starts the action
once you've created a Net::GitHub object with the proper username, repo name and token, the ticket creation is a single method call away. 10:39
alvis moritz: yes, i agree iff the person who launches parrotbug has a github account, unless i misunderstand you. ... 10:40
moritz alvis: parrotbug can simply send the email
alvis: and a cron job receives that email 10:41
alvis: and it's the cron job that has the parrotbug credentials
alvis ah, you mean using the approach you outlined ... yes, i agree, using your approach.
moritz: ok, ... makes sense to me. 10:45
moritz to reduce spam, we can do some simple checksumming on both the client and server side, and if they agree, we're pretty sure it's not spam 10:46
alvis moritz: sounds good. (i'll gist up something later for cotto and let him tell me which way to go.) 10:48
seme hey guys when I'm editing parrot source am I supposed to use spacing or tabs? 10:50
moritz spaces 10:51
seme also what would you guys consider to be a good tutorial on parrot... I've been reading the one on docs.parrot.org
alvis seme: spacing. you can check 'editor' if you're using emacs or vim
seme: it also has some formatting for kate and maybe others; i'm not sure. 10:52
seme I did the make vim-install from the editor directory but for some reason when I edit the Grammer.pm file it is all blue very odd
alvis seme: hmm, that's a tough one. give me a sec for some links ... 10:53
10:53 jsut_ joined
alvis there's parrot.github.com/html/pct_tutorial.html and there's benabik's github.com/benabik/cish 10:56
seme: there's also -- though not a tutorial -- whiteknight.github.com/Rosella/winx...ction.html 10:57
seme: also, as a rule, read just about everything whiteknight writes. 10:59
seme: fwiw, we're working -- though, admittedly, slowly -- to improve our docs. 11:00
seme thanks I appreciate the pointers 11:03
I'm very interested in the project but I'm just learning. I'll try to help where I can :) 11:04
alvis seme: good luck, and if you have questions, just ask. folks are very good about helping someone along. 11:07
11:51 benabik joined 12:04 bluescreen joined
alvis msg cotto please see gist.github.com/1639784 as well. 12:24
aloha OK. I'll deliver the message.
13:02 mj41 joined 13:23 benabik joined 14:10 mtk joined 14:20 jsut joined 14:31 PacoAir joined 14:44 whiteknight joined
whiteknight good morning, #parrot 14:46
14:48 slavorg joined
aloha (parrot/parrot) Issues opened : 700 (Expand note about rebaseall in README_cygwin.pod) by gerdr : github.com/parrot/parrot/issues/700 15:22
dalek sella: 7c1520c | Whiteknight++ | / (6 files):
Modify Template.get_standard_template routines to automatically end the '.template' extension to passed in file names.
15:36
sella: e086158 | Whiteknight++ | / (4 files):
add in three new templates for creating parrot release announcements (eventual replacement for parrot's crow.pir) in txt, html and markdown. Now added to the release tool yet
sella: 1736218 | Whiteknight++ | src/utilities/parrot_release.winxed:
Stub out some logic into parrot_release to automatically create release announcements. Add TODO notes. A few refactors to support this.
sella: c8fece3 | Whiteknight++ | src/date/Date.winxed:
Use string arrays to hold month and day name literals to avoid boxing
sella: e60d386 | Whiteknight++ | src/date/DateFormatter.winxed:
Add in three new format sequences to DateFormatter to work with 12-hour clocks
sella: 05b95d7 | Whiteknight++ | src/date/ (2 files):
the decodetime op is utc, so Date.now() should use decodelocaltime op instead. This fixes some weirdness I was seeing yesterday.
15:56 benabik joined 16:03 particle joined 16:15 PacoAir joined
dukeleto ~~ 16:17
whiteknight good morning, dukeleto 16:19
dukeleto whiteknight: a jolly good morning to you as well 16:21
whiteknight: what are you hacking on today?
whiteknight dukeleto: work stuff
frustrating, infuriating work stuff
dukeleto: you? 16:32
cotto msg alvis I like where you're going with that, but we may not even need to set up an email address and cron job if we get the parrot.org Drupal install involved. 16:37
aloha OK. I'll deliver the message.
cotto msg alvis I could hack out a custom github issues api for Drupal that'd act as the proxy between anonymous users and github. It'd probably take about 3 hours to get something deployable from the Drupal side.
aloha OK. I'll deliver the message.
cotto msg alvis Then all we'd need is for parrotbug.pl to hit that api on parrot.org . 16:38
aloha OK. I'll deliver the message.
whiteknight seen nine 16:41
aloha nine was last seen in #parrot 36 days 20 hours ago saying "I'll see what I can do tomorrow :) Time for bed now". 16:42
whiteknight that's a little troubling
16:44 PacoLinux joined
dukeleto whiteknight: roughly the same :) 16:50
whiteknight oh, how's the new gig going for you?
dukeleto whiteknight: good. So much to learn in the first few months
cotto looks like a snow day
dukeleto cotto: i heard seattle got some serious snow 16:51
cotto serious for seattle, yes 16:52
3-5 inches, and now it's icy
whiteknight as opposed to all the jocular, light-hearted snow they get some times
wise-cracking, flippantly disrespectful snow
cotto it takes about 1/2" before Seattle starts freaking out
dukeleto cotto: yeah, in PDX it is 1/4" :) 16:56
whiteknight philly starts freaking out around the 6-12 inch mark, but only because traffic gets even slower and we hate when traffic gets slower 16:59
last time it snowed that much, the streets were plowed but the sidewalks weren't shovled, so I had to walk to work in the middle of the road 17:00
cotto different world 17:01
17:06 fperrad joined
atrodo whiteknight> Around here, people freaked out last week when there was a lot of snow but nothing stuck to the road. Highways were jammed. But 12 inches of snow isn't exactly rare. 17:15
moritz it's funny and sad at the same time how some countries and institutions are surprised by snow every year again 17:17
whiteknight every year we have at least one snow storm where people rush out and buy shovels and snowblowers and stuff. I always wonder where the shovels are from last year 17:47
benabik o/ #parrot 17:48
whiteknight: These things do break.
17:56 dmalcolm joined
cotto ~~ 18:22
18:57 contingencyplan joined
cotto interesting post on memory usage efficiency and allocation request size: blog.mozilla.com/nnethercote/2011/0...01-and-up/ 18:59
dalek kudo/nom: 04b34ec | moritz++ | src/core/ (5 files):
revert the last few commits to deal with test failures

The following commits are reverted here: 0ac25d036db5630d9794135915174dd1ca0d1d08 First cut implementation of <prior> aa1d8b645107becb8fa6c7b3d87968a9ff1da296 optimize Range.roll and Range.pick 0deac24dfe5fda0ce8d46cdae66eba2d928f0c2d try to fix sleep() with non-Num argument 2fd18f4b66a69f2b9c925f1fbe720bfcf277faad fix Range.{roll,pick} 42068aafbf3abf0b581f81ad8871b20823ee53ca remove debugging statement ed943e669a8ea11e28f1e32771f8c0ca2f7ad361 stringify stuff before passing it to the regex engine
the previous state of the art is in the branch nom-with-failing-tests
19:10
19:13 ehiggs joined 19:31 benabik joined 19:38 not_gerd joined
not_gerd hello, #parrot 19:38
whiteknight hello not_gerd 19:39
not_gerd I finally got Rakudo to build on Cygwin 19:40
whiteknight nice
moritz \\o/
sorear opbots trust kthakore 19:44
slavorg Ok
sorear opbots trust kshannon
slavorg Ok
sorear opbots trust not_gerd
slavorg Ok
moritz opbots trust cosimo
slavorg Ok
19:51 bluescreen joined 20:06 jsut_ joined
whiteknight cotto: Oddly relevant: theoatmeal.com/blog/seattle_snow 20:10
cotto whiteknight: that's almost exactly how it works 20:16
dalek kudo/nom: f56c3ec | moritz++ | src/core/Range.pm:
optimize Range.roll, and Range.pick if the number of range elems is much larger than the number of elements to pick

this is a squashed version of these three commits:
   aa1d8b645107becb8fa6c7b3d87968a9ff1da296
   2fd18f4b66a69f2b9c925f1fbe720bfcf277faad
   42068aafbf3abf0b581f81ad8871b20823ee53ca
plus a small fix to .pick(*)
20:29
not_gerd I never noticed that before: is it normal that the Parrot DLL is ~64MB 20:30
cotto not_gerd: stripped? 20:33
It's big, but I've never seen it that big
not_gerd strip -S gets rid of ~30MB 20:36
cotto that's better, but still crazy
not_gerd it's comparable in size to my statically-linked 64bit clang-3.0.exe 20:41
so perhaps not *totally* insane...
(and lua52.dll is ~200kB ;)) 20:42
whiteknight when we rip out IMCC, expect the size to go down significantly
moritz but you'll have to wire up a replacement of sorts :-)
whiteknight moritz: sure, we'll cheat and offer it as a separate dll 20:43
moritz: eventually, I want us all to ween off the beast, and not use PIR as a code intermediary
Rakudo will benefit from such a change significantly, I think
....assuming we can get our act together and create a bytecode generating API that's worth using
benabik We can probably speed things slightly by repeating my GSoC project but with Winxed instead of NQP. 20:48
It'll even be slightly easier since Winxed doesn't use PCT, so there's no bootstrapping stage. 20:49
not_gerd just for the record, a stripped 64bit Windows binary cross-compiled from Cygwin is only ~3.5MB - don't know what's up with my Cygwin build... 20:58
20:59 mj41 joined
dalek kudo/nom: 7006f70 | moritz++ | src/core/control.pm:
try to fix sleep() with non-Num argument
21:01
whiteknight benabik: I'm going to get back on the PACT train, sooner than later 21:03
21:09 not_gerd left
whiteknight I'm running out of Rosella-related distractions. 21:16
dukeleto ~~ 23:02
dalek rrot: 91106b7 | (Gerhard R)++ | MANIFEST.generated:
add Cygwin-specific files to MANIFEST.generated
23:16
rrot: d654278 | (Gerhard R)++ | config/init/hints/cygwin.pm:
change inst_libparrot_* flags on Cygwin

NOT against the DLL cygparrot*.dll located in bindir
rrot: b9da642 | dukeleto++ | / (4 files):
Merge pull request #699 from gerdr/gerdr/cygwin

cleanup cygwin build
aloha (parrot/parrot) Issues closed : 699 (cleanup cygwin build) by gerdr : github.com/parrot/parrot/issues/699