Parrot 2.3.0 Released | parrot.org/ | Channel log: irclog.perlgeek.de/parrot/today | Priority: apply deprecations, merge branches, test exceptions_refactor branch | welcome our GSoC students
Set by moderator on 27 April 2010.
darbelo s/updating/slowly updating/ 00:02
cotto_work epic want: www.gizmag.com/first-commercially-a...ack/14423/ 00:03
darbelo ZOMG! We have finally reached the future! 00:04
... and svn still hasnt finished merging. 00:06
kid51_at_dinner purl, svn appreciators is kid51 00:14
purl OK, kid51_at_dinner.
dalek izkost: 9701482 | sorear++ | examples/cgi.pl:
Add CGI.pm example
kid51 Has runcore_purge branch been merged into trunk? 00:17
darbelo I think so, cotto would know better.
sorear it has 00:18
darbelo cotto_work should know as well.
kid51 clears up disk space
cotto_work yup 00:19
I thought I nuked it.
kid51 I'm just looking at my local drive's space.
cotto_work definitely no value in keeping that around 00:22
darbelo Oh great, and now I have to resolve all of the non-conflicts by hand. 00:24
dalek rrot: r46124 | Util++ | trunk/docs/book/pir/ch04_variables.pod:
[book] Fixed typos introduced by r46119
 00:25
darbelo > local delete, incoming delete upon merge 00:27
svn --
cotto_work you spelled svn-- incorrectly. 00:28
It should be svn--.
darbelo svn-- ? really?
Okay. I'll do it right this time. 00:29
cotto_work svn -- is an ineffective substitute for svn--
darbelo svn--
cotto_work much better
purl much better is to design each table in isolation with the table name providing context
cotto_work ;)
cotto_work goes svn--
er... home
the keys are right next to eachother 00:30
plobsing_ alias svn--="svn"
kid51 r 46124: make test; make buildtools_tests; make codetest all PASS: linux/i386 00:31
cotto_work coke, I'd appreciate it if you could get me a dump of our trac instance. I'm going to take a shot at a demo trac+git+svn site. 00:32
00:36 tcurtis joined 00:39 Mokurai1 joined
dalek rrot: r46125 | darbelo++ | branches/include_dynpmc_makefile:
Branch has merged into trunk and is no longer needed.
 00:41
darbelo Branch merged. I'm going to get some food. 00:43
msg me if anythig breaks after r46126. 00:44
kid51 Getting this error: 00:45
make[1]: Entering directory `/home/jimk/work/parrot/src/dynpmc'
make[1]: *** No rule to make target `../../config/gen/makefiles/dynpmc.in', needed by `Makefile'. Stop.
make[1]: Leaving directory `/home/jimk/work/parrot/src/dynpmc'
make: *** [dynpmc-clean] Error 2
That's on a 'make realclean' at r46126
darbelo Does it happen on perl Configure.pl ; make realclean ? 00:46
kid51 I had just done a 'make clean' and an 'svn up' 00:47
Then I did: make realclean && perl Configure.pl --test
... and I didn't get as far as Configure.pl
darbelo Yeah, the old makefile is trying to use a file that's been removed.
run Configure.pl without realcleaning and you'll be fine afterwards. 00:48
nopaste "kid51" at 192.168.1.3 pasted "Where 'dynpmc' occurs under config/" (11 lines) at nopaste.snit.ch/20402 00:49
darbelo You might also want to change your workflow to realcleaning *before* svn upping.
Otherwise tree changes in the svn up might prevent a complete cleaning. 00:50
kid51 Actually, that's what I almost always do. But as I was expecting upcoming commits to not touch config/, I for once only did a make clean after my last make test
darbelo: You might want to post to list about that. 00:51
sorear Why can't we just have a svn subcommand that blows away everything not under version control? 00:52
darbelo purl msg darbelo Post to list about the makefile changes when you get back.
purl Message for darbelo stored.
kid51 thx
darbelo waves
kid51 Either 'make test' is running a lot faster these days ... or Linode has given me more RAM ... or both. 00:54
... but now, the longest part of running a smoke test is the transfer to Smolder ;-)
108 wallclock seconds for make test 00:55
not bad
sorear kid51: well, we have a lot fewer cores to test. 00:56
kid51 Yes ... but weren't those only being tested on 'make fulltest'?
make test PASS smolder.plusthree.com/app/projects/...ails/33535 r46126 00:57
dalek rrot: r46126 | darbelo++ | trunk (5 files):
Merge include_dynpmc_makefile back to trunk, despite svn's efforts to prevent it.
kid51 afk 00:58
ttbot Parrot trunk/ r46124 i386-freebsd-64int make error tt.taptinder.org/file/cmdout/286862.txt ( tt.taptinder.org//buildstatus/pr-Pa.../rp-trunk/ ) 01:12
cotto How nice. When I synced my branch with trunk, svn added the files that had been created since the last sync but didn't bother committing them. 01:14
That explains why my build worked here.
dalek kudo: 3966c62 | (Solomon Foster)++ | src/core/ (2 files):
Straighten out type issues with Real.exp. In the process, add Real * Real, Complex ** Real, and Real ** Complex.
 01:17
kudo: 0857098 | (Solomon Foster)++ | src/core/ (2 files):
Remove Complex ** Any and Any ** Complex, as they should no longer be needed. Change Any * Any to convert its arguments to Numeric and call multiply on them then.
cotto If I had a desk in front of me, it'd have a big head-shaped indentation in it.
sorear Not a svn-shaped indentation? 01:21
cotto I hate to lose history but it's getting seriously annoying to salvage this branch. 01:28
sorear let
cotto ? 01:29
sorear let's ditch this svn thing and go back to manual patch handling.
Whiteknight and we could send the patches via carrier pigeon
dukeleto i hear that git thingy is pretty useful 01:35
cotto++ # demo trac+git site 01:36
dukeleto is backlogging
01:44 dalek left 01:45 dalek joined
cotto Don't give me karma yet. I haven't even done anything. 01:52
01:53 sorear joined
dukeleto cotto++ # take that karma and like it 01:58
01:59 Psyche^ joined
cotto Fine. Be that way. 02:01
02:07 kurahaupo joined 02:10 ash_ joined
dalek kudo: ed321b4 | (Solomon Foster)++ | src/core/Cool-str.pm:
Fix comb version of the match with take bug fixed earlier today in split.
 02:15
02:33 kurahaupo joined 02:48 darbelo joined
dalek kudo: e16cf45 | duff++ | tools/contributors.pl:
[tools] fix typo and output sorted keys
 02:55
02:58 JimmyZ joined 03:00 janus joined
darbelo Okay, mail sent to the list warning about svn upping without a realclean. 03:03
purl: msg Coke include_dynpmc_makefile merged. I'll be busy with GSoC work so the rest of the recursive make calls are up to you. 03:06
purl Message for coke stored.
dukeleto darbelo: do you need anything from me? I have am travelling for work and I am a bit behind with GSoC stuff 03:14
s/have// 03:15
darbelo dukeleto: Not that I can think of, why? 03:16
dukeleto darbelo: just making sure. 03:20
03:20 ingy joined
darbelo Oh, okay. I don't know about other students but I'm pretty much set up. Merging that last branch to trunk was the last step to "Get my affairs in order" fo GSoC. 03:20
dukeleto darbelo++ 03:24
darbelo: we will be doing the "blog once a week" thing this year, just to let you know. You already attend the weekly #ps too, so that isn't news to you :) 03:25
darbelo Yeah, I've got the blog set up already. I'll star with the writing tomorrow^Win the morning. 03:27
I talked with allison and she recommende I try to get as much of the design on 'paper' as I can before starting on the coding. 03:28
sorear I used to do that a lot 03:29
but ... design gets out of hand quickly for me
dukeleto darbelo: that sounds like a great idea. The more coding I do, the more I see the benefit to "coding in your head" and writing notes on design, before you ever write a line
sorear don't overdo it, I warn you 03:30
dukeleto sorear++ # good advice
cotto at the same time, you don't find some holes in your thinking until you start trying to hack them out
dukeleto darbelo: finding the middle path between design and writing actual code is an art 03:31
darbelo Well, 'design' might be an overstatement. What I plan to do has already been specified in the strings PDD.
dukeleto cotto: yes, it is definitely a cyclical process. design, code, design, code, etc... The important thing is that the cycle *starts* with design
cotto +1 03:32
purl 1
sorear plan to throw one away
actually, expect to throw away more like 4
brooks was such an optimist
darbelo We've been throwing parts fo the vm away for ten years now. 03:33
We've become quite good at it.
dukeleto Deattachment is a virtue. 03:34
sorear The amount of code you need to throw away grows as n log(n) where n is the size of the project
dukeleto makes up words
cotto You can get a head start on throwing stuff away. There was someone who started working on NFG a while ago. (Simon Cozens?)
You could talk to him to see if he got far enough to run into any difficulties. 03:35
darbelo Yeah, the branch is still there. But it's all in perl.
sorear ...perl?
darbelo Yep.
sorear parrot core code can be written in perl now? 03:36
cotto Yeah, as a first approximation.
you wish
darbelo It was a prototype. It came pre-thrown away, so to speak.
sorear doesn't understand how you can prototype code in the wrong language
JimmyZ perl 6 is easy to write. 03:37
darbelo trac.parrot.org/parrot/browser/bran...pseudocode
sorear JimmyZ: I'm not complaining about perl.
JimmyZ heh 03:38
sorear However, I was under the impression prototypes had to work to be considered prototypes
darbelo He was prototyping NFG, not NFG-for-parrot.
sorear I see, this was long before NFR 03:39
cotto NFR? 03:40
darbelo But I wasn't here when he was doing it, so I can't really speak about what motivated him to do it in perl
sorear cotto: normalization form rational; an arithmetic version of NFG proposed by TimToady 03:43
each character in a NFG string is a rational number <main character>.<combiners, in order, 21 bits each>
s/NFG/NFR/
Coke cotto: do you know what such a dump would involve? 03:44
(if you can step me through the commands, I can probably do it.)
sorear (there would need to be a slightly clever implementation here, mixing numbers and pointers, to not be terribly inefficient)
advantage: order isomorphic to NFC
cotto Coke, I don't.
sorear advantage: simple semantic presentation "it's just an array"
advantage: no coupling between characters in the same or different strings 03:45
cotto Cloning the svn repo is easy enough.
trac.edgewall.org/wiki/TracBackup perhaps 03:46
What db backend are we using? 03:47
sorear, oic.
Coke cotto: no clue.
03:47 ingy joined
Coke I'll open a ticket and cc you. 03:47
cotto thanks 03:48
Does all the interesting admin stuff have to go through osuosl or is that just the best arrangement for the time being? 03:50
mikehh, ping 03:51
Coke cotto: I have root on the box. 03:52
but now is not a good time for me to sink time into this.
depending on how much they come back with (they may have already done this for another org, for all I know), we'll see what happens.
it's not best, just expedient for me.
cotto np 03:57
04:00 jsut joined 04:26 mikehh_ joined 04:39 Andy joined 04:40 snarkyboojum joined, ingy joined 04:53 ash_ joined 04:54 jsut_ joined 05:03 rurban_ joined 05:29 particle1 joined
cotto The setup for trac isn't that bad other than the fact that it slurps the entire svn repo into its db the first time you try to access the trac instance. 05:33
05:58 viklund joined
JimmyZ ttbot? 06:00
purl ttbot is probably TapTinder build bot owned by mj41 and reporting tt.taptinder.org/buildstatus/pr-Parrot/rp-trunk build errors. or see taptinder
JimmyZ taptinder?
purl taptinder is continues integration build tool - taptinder.org . For Parrot project running on tt.perl6.cz/ and reporting build failures to #parrot channel as ttbot.
cotto Woohoo. I'm admin on my local Trac install. 06:05
parrot git? 06:08
purl i guess parrot git is repo.or.cz/w/parrot.git or github.com/leto/parrot
cotto dukeleto++ 06:10
mj41 purl forfet taptinder 06:16
purl mj41: sorry...
06:17 uniejo joined
mj41 taptinder? 06:23
purl taptinder is continues integration tool - taptinder.org . For Parrot project running on tt.taptinder.org/ and reporting build failures to #parrot channel as ttbot.
mj41 ttbot?
purl ttbot is TapTinder build bot owned by mj41 and reporting tt.taptinder.org/buildstatus/pr-Parrot/rp-trunk build errors. See also TapTinder.
06:24 szabgab joined 06:32 iblechbot joined
cotto I think trac with svn and git may be workable. 06:48
Actually, I think I'm going to do whatever it takes to make it workable because I'm tired of svn, but I don't think it'll take too much.
06:54 fperrad joined
sorear trac with svn *and* git? 06:57
I find dealing with svn's annoyances enough, I don't think I could handle simultaneously coping with git's
07:07 fperrad_ joined 07:21 eternaleye joined
Tene sorear: I can deal with svn primarily by using git-svn as an intermediary. 07:22
When dealing with svn directly, I all too often screw something up.
07:41 snarkyboojum joined 08:15 jsut joined
dalek rrot: r46127 | fperrad++ | trunk (7 files):
[library] add Archive/TAR
 08:22
bacek ~~ 08:35
08:50 jsut_ joined 08:57 mikehh joined
fperrad msg darbelo I don't understand your commit r46043 in src/dynpmc/gziphandle.pmc 09:00
purl Message for darbelo stored.
dalek rrot: r46128 | gerd++ | trunk/ports/mandriva (4 files):
update the packing information for mandriva
 09:11
ttbot Parrot trunk/ r46128 i386-freebsd-64int make error tt.taptinder.org/file/cmdout/287065.txt ( tt.taptinder.org//buildstatus/pr-Pa.../rp-trunk/ ) 09:16
dalek rrot: r46129 | fperrad++ | trunk/examples/pir/befunge/io.pir:
fix r46105
 09:28
ttbot Parrot trunk/ r46129 i386-freebsd-64int make error tt.taptinder.org/file/cmdout/287119.txt ( tt.taptinder.org//buildstatus/pr-Pa.../rp-trunk/ ) 09:35
09:43 riffraff joined 10:45 khairul joined 10:57 khairul joined 11:37 gaz joined 11:50 snarkyboojum joined 11:54 plobsing joined 11:57 whiteknight joined
whiteknight good morning, #parrot 11:58
12:02 tetragon joined
bacek whiteknight, 2 mistakes in 3 words. It's not "good". And it's not "morning". 12:02
:) 12:03
moritz it's always morning somewhere :-)
whiteknight bacek: every morning is a good morning 12:04
I could have been struck by a meteor while I slept
bacek whiteknight, but it's not morning anymore :)
whiteknight then I must be really late to work! 12:05
bacek whiteknight, you are! 12:07
bacek spent whole day at meeting at company which don't pretend "don't be evil" 12:08
Night, humans. I'm going to dream about electric sheeps for next few hours 12:10
moritz have the appropriate amount of fun
dalek rrot: r46130 | fperrad++ | trunk/runtime/parrot/library/Archive/TAR.pir:
[TAR] now, compute the checksum
 12:27
12:34 ruoso joined
ttbot Parrot trunk/ r46130 i386-freebsd-64int make error tt.taptinder.org/file/cmdout/287205.txt ( tt.taptinder.org//buildstatus/pr-Pa.../rp-trunk/ ) 12:36
Parrot trunk/ r46132 i386-linux-thread-multi make error tt.taptinder.org/file/cmdout/287236.txt ( tt.taptinder.org//buildstatus/pr-Pa.../rp-trunk/ ) 12:41
dalek rrot: r46131 | fperrad++ | trunk/runtime/parrot/library/Archive/TAR.pir:
[TAR] fix prefix
 12:43
rrot: r46132 | fperrad++ | trunk/runtime/parrot/library/TAP/Harness.pir:
[TAP] refactor TAP;Harness;Archive with Archive/TAR library
darbelo purl: msg fperrad It was an accidental commit, I was testing something unrelated and forgot to undo a change. I undid it in r46133. 12:47
purl Message for fperrad stored.
12:49 alexn_org joined 12:52 iblechbot joined
dalek rrot: r46133 | darbelo++ | trunk/src/dynpmc/gziphandle.pmc:
Undo accidental commit.
 13:00
rrot: r46134 | fperrad++ | trunk/runtime/parrot/library/distutils.pir:
[distutils] refactor target 'sdist_gztar' with library Archive/TAR & GzipHandle PMC
rrot: r46135 | darbelo++ | trunk/src/dynpmc/Rules.in:
Make dynpmcs build after pmcs are built to avoid parallel build breakage.
Coke sorear: easy enough to treat git like svn. (use whatever github is using.) 13:03
13:03 rurban_ joined
Coke darbelo++ #fixing those deps. 13:04
darbelo: i think Andy D. might be mistaken - I think because those dynpmcs are not autogen'd, they miss depending on .dump files that they should. yes? 13:05
darbelo: ah, you cheated and gave it an overbroad dep? 13:07
darbelo Coke: Yes. I am a cheat.
Coke that's fine for now.
darbelo I'm replying to the list now, outlining a proper fix. 13:10
Coke allison, chromatic, et al.: I've been marking new stuff that I see going in as experimental to help us avoid support issues if we decide to rip it out; we should probably review these experimental tickets near supported release boundaries to see if we wish to make them supported. 13:12
allison Coke: good idea
Coke (e.g. new pmcs, runtime/ support, etc.)
(easier to add the marking as we go than it is to remember to add it before the supported release if we want it.) 13:13
darbelo Coke++ 13:14
13:16 particle joined
dalek rrot: r46136 | fperrad++ | trunk/runtime/parrot/library/distutils.pir:
[distutils] fix r46134
 13:16
rrot: r46137 | coke++ | trunk/DEPRECATED.pod:
Track experimental status of new module.
TT #1598 created by coke++: Archive::Tar in parrot library 13:23
TT #1598: trac.parrot.org/parrot/ticket/1598
ttbot Parrot trunk/ r46137 i386-freebsd-64int make error tt.taptinder.org/file/cmdout/287414.txt ( tt.taptinder.org//buildstatus/pr-Pa.../rp-trunk/ ) 13:25
dalek tracwiki: v5 | fperrad++ | OSInterfaceAdditions 13:27
tracwiki: trac.parrot.org/parrot/wiki/OSInter...ction=diff
dukeleto 'ello 13:29
how do people feel about making parrot GSoC students write a blog post for the parrot.org blog each week? The alternative is for them to have to figure out where they will put their blog post, and that has lots of headaches 13:32
moritz dukeleto: +1 to "forcing" them to parrot.org 13:34
darbelo parrot.org worked fine for me last year. +1
whiteknight dukeleto: I'm happy either way. 13:35
particle dukeleto: parrot.org, with gsoc2010 tag on each post
whiteknight so long as the blog posts they write are included in the parrot aggregator, I don't care where the "original" is located, really
parrot.org is fine
particle so they can be aggregated to a tpf feed of all gsoc2010 student blogs
dukeleto: did you see my pafo press release about gsoc2010?
Coke parrot.org is fine, having them tag their posts with GSOC2010 or something. 13:36
there was a TPF one just recently, too
particle ah, good. that's what i was getting at. i haven't seen it yet ,will check my feed reader.
hrmm, a little plain, but it gets the job done. 13:40
dukeleto particle: where is it?
purl it's it!
particle parrot.org
purl hmmm... parrot.org is drupal? or aggressively cached.
particle news.parrot.org to be precise, but it's on the home page
dukeleto particle: we should mention that Parrot+RTEMS also got accepted to the RTEMS GSoC 13:41
particle good point....
dalek tracwiki: v2 | dukeleto++ | GSoCersStartHere 13:43
tracwiki: trac.parrot.org/parrot/wiki/GSoCers...ction=diff
dukeleto particle: i will try to write something up soon, and maybe the RTEMS guys will write something
Coke GSoCersStartHere is an anagram for Grocers Hate Rests. 13:44
dukeleto Coke++ 13:46
dalek rrot: r46138 | NotFound++ | trunk/t/pmc (2 files):
improve packfile PMCs tests
 13:48
13:49 ruoso joined
ttbot Parrot trunk/ r46138 i386-linux-thread-multi make error tt.taptinder.org/file/cmdout/287462.txt ( tt.taptinder.org//buildstatus/pr-Pa.../rp-trunk/ ) 13:51
dalek rrot: r46139 | NotFound++ | trunk/src/pmc (3 files):
[cage] replace several string_from_literal with CONST_STRING
 14:05
14:08 Andy joined
NotFound Someone in building on windows right now? 14:09
There is a problem in os.pmc:517, the argument of string_from_literal must ba a literal. 14:10
dukeleto test 13 in t/library/pg.t is failing for me on ubuntu 9.10/x86 14:13
14:19 JimmyZ joined
JimmyZ NotFound: What can I help you? 14:20
NotFound JimmyZ: fixing that ;) 14:21
JimmyZ good :)
NotFound I don't have a windows machine right now and I don't like to commit an untested fix. 14:22
JimmyZ I have a windows machine 14:24
Coke JimmyZ++
14:26 bubaflub joined
JimmyZ NotFound++ 14:26
Andy Don't use string_from_literal, then. 14:27
It's jsut a wrapper around Parrot_str_new. 14:28
and in fact, if yo udon't pass a length to Parrot_str_new, it will take the length for you.
ttbot Parrot trunk/ r46139 cygwin-thread-multi-64int make error tt.taptinder.org/file/cmdout/287544.txt ( tt.taptinder.org//buildstatus/pr-Pa.../rp-trunk/ ) 14:31
NotFound Andy: is not a wrapper, is a ugly macro unable to check its correct usage.
Andy tomato, tomatoh 14:32
So don't use it.
14:34 particle1 joined
NotFound I don't. 14:34
14:39 tetragon joined
tewk NotFound: So i did the PBC_COMPAT magic and got newly generated _4s not _1s, I assume _?s correspond to platforms, and _4 is x86_64, what is _1? 14:47
NotFound tewk: 32 bit low endian, I think. 14:48
14:52 tetragon_ joined
dalek rrot: r46140 | NotFound++ | trunk/t/pmc (7 files):
more improvement of packfile pmc tests
 14:53
14:58 theory joined
dalek izkost: 49be3c0 | (Ævar Arnfjörð Bjarmason)++ | Configure.pl:
This takes a path to parrot_config, not parrot
 15:05
plobsing tewk: check out the doc in t/native_pbc/integer.t for a description of what the _\\d means 15:06
15:09 khairul joined
Coke who is Devin Austin? 15:29
purl Devin Austin is no longer Software Developer/Code Monkey at Midstream Media
Coke who is Devin Austin not? 15:30
15:35 davidfetter joined
dalek rrot: r46141 | NotFound++ | trunk (3 files):
more packfile error checking
 15:42
Coke there are still one or 2 slots left at yapc.
dalek kudo: fe59fa8 | moritz++ | src/Perl6/Grammar.pm:
enable rx{...} quoting form. At some point we need a more general solution for these quoting mechanismsn...
 15:44
tewk allison: I'd like to check-in some changes to ParrotThreads, right now Parrot_pmc_reuse is used which causes problems when ParrotThreadPMC is extended in PIR land. 15:56
allison Tewk: sounds reasonable
15:57 darbelo joined
dalek rrot: r46142 | bubaflub++ | trunk/runtime/parrot/library/Archive/TAR.pir:
[codingstd] removed extra =back to fix POD syntax
 15:59
rrot: r46143 | fperrad++ | trunk/runtime/parrot/library/Archive/TAR.pir:
[TAR] more options and fix the logic
Coke any parrot folks want to share a dorm room at yapc and save a few bucks? 16:00
I will probably not share so as to avoid inflicting snorage. :| 16:01
darbelo snorage--
particle1 sno makes you angry? 16:02
Coke COKE SMASH
particle1 have a sno-cone. that makes everybody happy! 16:03
cotto NOM
sorear seen pmichaud 16:04
purl pmichaud was last seen on #parrot 6 days, 14 hours, 56 minutes and 27 seconds ago, saying: cotto_work: (troll refutation) Thanks. :-) [Apr 23 01:07:58 2010]
ttbot Parrot trunk/ r46145 i386-linux-thread-multi make error tt.taptinder.org/file/cmdout/287795.txt ( tt.taptinder.org//buildstatus/pr-Pa.../rp-trunk/ ) 16:09
cotto htop++ (Why didn't someone tell me about this sooner.) 16:14
dalek rrot: r46144 | NotFound++ | trunk/t/native_pbc (4 files):
update native_pbc fox x64
 16:15
rrot: r46145 | petdance++ | trunk/src/string/api.c:
collapse redundant calls to Parrot_str_new_init
whiteknight htop? 16:20
purl htop is cool
cotto awesomer version of top 16:21
whiteknight thanks, purl
top?
purl i think top is rather lame. Glance is better or top is usually at the top of the cpu list because it's reading /proc and killing the kernel at the same time. or but not if you do 'top -n 1'
cotto no, htop is an awesomer version of top 16:22
purl okay, cotto.
Coke cotto++ # never seen that before
macports++ # installed it in 5s.
Coke infects the server guys here. 16:24
16:24 ash_ joined
darbelo chuckles at the amount of mail that every 'please stop sending so much mail' thread generates on the students list. 16:27
ash_ lol 16:28
there are a bunch of GSoC emails 16:29
bubaflub haha. same thing happened last year
when i was added to the list i didn't have a chance to change my email to "digest" so i initially got flooded...
Coke htop++
darbelo Last year it was mostly confined to two or three threads too, making it easier to have your mailer ignore it. 16:30
Andy Is it OK to change diagnostics? 16:31
darbelo Only if it is for the better ;) 16:33
Andy In string_ord, I'm adding a diagnostic that specifically checks for STRINGNULL, and complains about a "NULL string" instead of an "empty string" 16:35
ash_ I just got a new chunk of nq-nqp working, it now knows about infix ops and method calls, and generates llvm-ir code for them, although i don't have runtime support for all of the method calls yet
bubaflub ash_++
Coke Andy: that seems reasonable. 16:36
Andy whoo, I just changed my first .t file 16:37
ash_ gist.github.com/382765 has an example of something that works
ttbot Parrot trunk/ r46142 i386-freebsd-64int make error tt.taptinder.org/file/cmdout/287881.txt ( tt.taptinder.org//buildstatus/pr-Pa.../rp-trunk/ ) 16:38
ash_ dang, my nq-nqp executable is 28M... because I am statically linking to the llvm, but still 16:41
particle1 how big is llvm? 16:42
darbelo Laid down on the floor, about a meter. 16:44
;)
dukeleto particle: what kind of metrics are you looking for? 16:45
there is going to be a talk at Open Source Bridge this year about buildings DSLs with LLVM. I hope I can make it to that talk
the talk is by Jeremy Voorhis, github.com/jvoorhis 16:46
i think parrot can learn a lot from his stuff 16:47
ash_ particle: I have all of the static libraries on my computer, its 410.3 MB, but that is all of the code gen librarires for all of the architectures the llvm supports, you can make it build only your host architecture and it would be signifcantly smaller 16:48
dalek rrot: r46146 | petdance++ | trunk (5 files):
Disallow NULLs in string_ord, and fix const and ARGMOD on Parrot_sprintf_format
ash_ you can also build it as a shared library, but on the llvm site they say they don't recommend that for some reason
Andy why would 46145 have failed only once? tt.taptinder.org//buildstatus/pr-Pa.../rp-trunk/ I can't even see why it failed at all. 16:49
ash_ dukeleto: interesting, do you know anything else about the talk on DSL's with the llvm? 16:50
16:51 lucian joined 16:53 lucian joined
Coke it's been failing on and off for a day. 16:55
tickling a segfault. 16:56
darbelo A backtrace might be helpful.
dukeleto ash_: it was just accepted to OSBridge, if you look at Jeremy Voorhis' git repos, he has ruby-llvm bindings and some other related stuff 16:58
ttbot Parrot trunk/ r46146 i386-freebsd-64int make error tt.taptinder.org/file/cmdout/287939.txt ( tt.taptinder.org//buildstatus/pr-Pa.../rp-trunk/ ) 16:59
17:00 spinclad joined
cotto_work good morning 17:14
purl Good Morning Mr Rogers
cotto_work forget good morning
purl cotto_work: I forgot good morning
cotto_work trac with git+svn is looking promising. 17:46
Coke talk submissions for yapc now closed. 17:47
particle coke: after they mentioned open slots? 17:48
Coke yes.
this just in from #yapc. 17:49
bubaflub cotto_work: are you playing with a trac git+svn instance? 17:50
cotto_work Yeah. I have an install on my laptop. 17:51
bubaflub very cool.
particle why git+svn?
cotto_work so all the old links to the svn repo in trac don't break 17:52
if we switch to git 17:53
afaict that's the biggest technical blocker
particle would that mean we have a read-only svn, with old links still operational, but new links resolving to git?
cotto_work yup 17:54
particle and you can make that happen... have two link resolvers in trac?
cotto_work That's what I'm finding out.
particle i'd be stunned if it works. 17:55
darbelo It'd also solve rakudo's problem of needing a 'svn rev' to build parrot.
bubaflub can just do a git submodule
cotto_work There's a small patch that allows trac sub-projects. It may be possible to use that plus some mod_rewrite (or similar) trickery.
bubaflub (instead)
particle bubaflub: don't say 'just' when it comes to trac integration 17:56
bubaflub particle: i was referring to the rakudo build process
i mean instead of having to provide a svn revision number the project can use git submodules
so building rakudo wouldn't require both git and svn checkouts 17:57
that's all
yeah, trac is a tempestuous wench
moritz purl: yeah, trac? 18:01
purl hmmm... yeah, trac is more project tracking
darbelo purl: forget yeah, trac 18:03
purl darbelo: I forgot yeah, trac
18:11 ash_ joined 18:31 particle joined 18:55 lucian_ joined 18:57 lucian__ joined
dalek kudo: c4857ab | (Solomon Foster)++ | src/core/metaops.pm:
Turn on zero-argument versions of infix:<~> and infix:<//>.
 19:05
kudo: 4aeb5a2 | (Solomon Foster)++ | src/builtins/assign.pir:
Have $a op= $b call op() to get a value for $a if $a is not defined.
kudo: 9d955f2 | (Solomon Foster)++ | src/core/ (2 files):
Be more specific about types for Complex mixed-type math. Change Any, Any versions of the basic math functions to coerce to Numeric (instead of called PIR functions which coerce to Num).
19:33 jan joined
NotFound realclean required for r46147 19:33
19:41 joeri joined
dalek TT #1599 created by coke++: experimental packfiledebug PMC 19:43
TT #1599: trac.parrot.org/parrot/ticket/1599
rrot: r46147 | NotFound++ | trunk (7 files):
new PMC packfiledebug, update PBC_COMPAT, realclean recommended
 19:48
rrot: r46148 | jonathan++ | trunk/runtime/parrot/library/P6object.pir:
[p6object] Update to make sure we go finding methods through introspection rather than in the namespace, which is an especially bad idea now :method (happily) doesn't put methods there any more.
rrot: r46149 | coke++ | trunk/DEPRECATED.pod:
track new packfiledebug pmc...
NotFound Coke: I think there was already a generic experimental note for all packfile PMCs 19:49
Coke NotFound: not in DEPRECATED.pod
one for packfile structure... 19:50
ash_ are nqp object attributes write-once only? 19:51
nqp: say(4); # does this channel have an eval bot? 19:52
guess not
cotto_work we used to have polyglotbot 19:54
particle this is #parrot. we eat pir and spit out machine code. 19:55
Coke ... not exactly. 19:56
particle segfaults
purl No whammies!
dalek TT #652 closed by doughera++: non-constant initializations in src/pmc/*.c 20:01
TT #652: trac.parrot.org/parrot/ticket/652
cotto_work mikehh: ping 20:13
dalek kudo: cdd3430 | jonathan++ | src/ (6 files):
A bunch of changes required to make us run under latest Parrot, mostly because we relied wrongly on :method's old semantics of putting things into the namespace, and the 4-argument substr op has gone.
 20:20
kudo: 9d927ac | jonathan++ | build/PARROT_REVISION:
Bump PARROT_REVISION to latest Parrot.
kudo: 6e91b9b | jonathan++ | (10 files):
Merge branch 'master' of git@github.com:rakudo/rakudo
20:23 tcurtis joined 20:24 ash_ joined
NotFound I have a todo passing on i386 but failing on x86 20:29
20:30 tcurtis left, tcurtis joined
cotto_work ? 20:30
darbelo i386 vs x86 ?
I'm guessing he missed a "-64"
NotFound Sorry, amd64
t/pmc/packfile.t - 34 20:32
20:36 spinclad joined
mikehh cotto_work: pong 20:39
cotto_work can you verify that trac.parrot.org/parrot/ticket/1592 is no longer an issue? 20:40
Coke cotto: any clues on your random segfaults on taptinder? 20:42
cotto_work ? 20:43
Coke taptinder has segfaults sporadically in your nqp profile stuff.
mikehh cotto_work: the test now works, dukeleto++ modified it - I am not sure about the issues NotFound++ raised 20:44
NotFound mikehh: I think is more a dessign issue. Without a clear design in mind is hard to know how to improve the test 20:45
cotto_work ok. Those issues are out of scope for the existing ticket. Further discussion should happen in a new ticket. I'll close TT #1592.
NotFound Ok
dalek TT #1592 closed by cotto++: t/op/io.t fails in non optimized build 20:50
TT #1592: trac.parrot.org/parrot/ticket/1592
20:53 theory joined
ttbot Parrot trunk/ r46148 i386-freebsd-64int make error tt.taptinder.org/file/cmdout/288112.txt ( tt.taptinder.org//buildstatus/pr-Pa.../rp-trunk/ ) 20:54
20:56 kjeldahl_ joined
Coke dislikes GETATTR_String_str_val 20:59
21:01 particle joined 21:03 rurban_ joined
Coke String use GETATTR_String_str_val instead of SELF.get_string() which kind of blows up for subclasses that are using their own string storage. 21:06
Any disadvantages to changing all those to SELF.get_string() ?
cotto_work Are they in hot code paths? 21:08
ATTR access is much faster than a VTABLE function call.
Coke don't care.
it's wrong, so... =-)
cotto_work oh
you mean for pir subclasses?
Coke I subclass String. I don't store my string in that ATTR... boom. 21:09
this is a PMC subclass.
cotto_work yeah. That'd explode.
Coke something like a StringBuilder.
cotto_work The attr code assumes that if the attr exists, it's used in the same way in children.
dalek rrot: r46150 | mikehh++ | trunk/MANIFEST:
re-generate MANIFEST
 21:10
rrot: r46151 | mikehh++ | trunk/src (3 files):
add svn properties
Coke 2 choices - reimplement all the vtables with the .get_string() and .set_string_native(), or update string to not break its own encapsulation.
Coke I suppose I can go with the "don't change String" path for now.
21:20 lucian joined 21:25 wagle joined
dalek rrot: r46152 | mikehh++ | trunk/src/pmc/packfiledebug.pmc:
fix codetest failure - there should be at least one space between a C
 21:26
rrot: r46153 | mikehh++ | trunk/src/pmc/packfiledebug.pmc:
fix codetest failure - line length
rrot: r46154 | chromatic++ | trunk/src (2 files):
[PMC] Made CodeString's emit() method use Parrot_str_join() rather than

benchmark by 0.454% by removing code.
rrot: r46155 | chromatic++ | trunk/src/pmc.c:
[pmc] Fixed a documentation typo.
21:27 particle joined 21:31 particle joined 21:40 jsut joined
dalek rrot: r46156 | mikehh++ | trunk/src/pmc/packfiledebug.pmc:
fix codetest failure - tabs in leading whitespace
 21:42
rrot: r46157 | mikehh++ | trunk/src/packfile.c:
fix codetest failure - tabs in leading whitespace
21:45 cotto_work joined
cotto_work particle: ping 21:45
particle pong 21:46
cotto_work how well do you know trac guts?
particle not very 21:47
cotto_work so I'm just about as well-off trying to figure them out myself as asking you?
particle aye 21:50
dalek rrot: r46158 | mikehh++ | trunk/src/dynpmc (2 files):
add copyright, $Id
 21:59
22:06 particle joined
cotto_work I might need a small custom plugin to make svn links dtrt but this is still looking quite doable. 22:08
mikehh t/pmc/resizablestringarray.t - Non-zero wait status: 11 - Parse errors: Bad plan. You planned 261 tests but ran 211 - in make corvm/make coretest (looks like a segfault) - it passes make test 22:09
cotto_work Hacking out some Python code is a small price to pay. Even writing PHP code wouldn't be out of the question if it were necessary. ;) 22:10
mikehh thats on Ubuntu 10.04 amd64 (g++)
I ran it a couple of times - same result 22:11
cotto_work Who's g and what'd he do to earn the karma?
.)
mikehh that satrange character who compiles slightly differently to his bro gcc 22:12
strange maybe
darbelo cotto_work: Writing PHP is an option? You must seriously hate svn... 22:30
I mean, keep that up and you'll be suggesting amputations with no anesthesia.
cotto_work PHP isn't an option but I'd take it if it were in preference to continuing to fight svn. 22:31
Coke msg chromatic : arg. you beat me to it, doing it a completely different way. 22:39
purl Message for chromatic stored.
Coke msg chromatic You're still doing a lot of concats there that you don't need to be doing. I can improve your score, I think. 22:40
purl Message for chromatic stored.
Coke I'll take it to a branch, I guess, to make it more obvious. 22:41
22:41 kid51 joined
dalek rrot: r46159 | coke++ | branches/codestring:
Avoid even more string concat in CodeString
 22:48
22:57 Whiteknight joined 23:00 kurahaupo joined
dalek rrot: r46160 | coke++ | branches/codestring/src/pmc/codestring.pmc:
First pass at converting codestring to swap out its storage for

as possible. Build fails - need to fix String vtables (or Codestring) to avoid trying to access the string value via ATTR instead of the vtables.		 23:04
ttbot Parrot trunk/ r46157 i386-freebsd-64int make error tt.taptinder.org/file/cmdout/288417.txt ( tt.taptinder.org//buildstatus/pr-Pa.../rp-trunk/ ) 23:05
tewk cotto_work++ 23:06
cotto_work Coke: is parrot.org on its own VM or otherwise completely independent environment? 23:13
i.e. if we want to patch Trac, is it reasonable to do so?
GeJ I have a foo_group.h that is created during the `make` phase but it isn't cleaned up with `make realclean` does that ring any bell to anyone? 23:29
from the instances of 'foo_group' in Parrot source, il looks like it has something to do with dynpmc 23:31
mikehh Gej: I also notices that when I ran svn status -u 23:32
noticed
Gej: I haven't tracked it down yet but it probably needs adding to svn:ignore and make clean 23:33
dalek TT #1600 created by mikehh++: t/pmc/resizablestringarray.t FAILS in make corevm/make coretest 23:34
TT #1600: trac.parrot.org/parrot/ticket/1600
mikehh I am going to check TT #1600 on my i386 platform - need to reboot - bbl 23:38
cotto_work fulltest is happy for me at r46160 23:43
Coke cotto_work: yes, we have our own VM that is separate from other OSU hosted sites. 23:47
cotto_work ok. That's good. 23:48
Coke cotto_work: is there a cheap way to check to see if we're invoking the string vtable from String vs. something else? 23:49
... hurm. I guess I could fix this a different way. 23:50
cotto_work from where?
Coke I'm trying to fix the String attr issue.
cotto_work you could pretty easily check pmc->vtable->base_type
against enum_class_String 23:51
Coke danke, will keep that in mind. 23:55
Coke wonders if it's better to resize an RSA to 0 or create a new one. 23:57
Coke guesses resize to 0.
cotto_work You won't need a new PMC header that way. 23:59

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!