Parrot 2.5.0 Released! | parrot.org | Log: irclog.perlgeek.de/parrot/today | Nopaste: nopaste.snit.ch:8001 | Priorities: review experimental features for promotion or removal, fix 'make html' (talk to Coke), update tutorial (talk to tcurtis), pre-release testing.
Set by moderator on 13 July 2010.
Coke sloooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooow 00:00
whiteknight oh 00:01
Coke 00:10
we really did not have a lot going on this month in trunk. 00:12
presumably it's going to get hammered tomorrow.
darbelo kind of doubts it. 00:17
cotto_work I'll merge some random branches. 00:18
cotto_work runs a rng
Looks like I'll be mergeing gsoc_threads, pmc_i_ops and avl_string_cache. 00:19
darbelo Hmm, svn ls ^/branches makes me think some pruning is in order. 00:20
Coke darbelo: ayup. on my list. 00:22
if you could regen the branch listing for the wiki and email the list, that'd be most helpful. 00:23
darbelo There was a script for that, right?
Coke seen chromatic?
purl chromatic was last seen on #parrot 2 hours, 13 minutes and 35 seconds ago, saying: TT #1639, dukeleto.
cotto_work running it now
darbelo cotto++ 00:24
Coke darbelo: yes. see: trac.parrot.org/parrot/wiki/BranchDescriptions and tools/dev/branch_status.pl
cotto_work: a patch to update the wiki output to reference the script woudl probably be a good idea postrelease. 00:25
msg bacek - when merging changes from nqp-rx into parrot, please document /the reason/ for the update, not just "latest version". Danke. 00:26
purl Message for bacek stored.
Coke picks an actual type of parrot for the release.
Coke hates release names, but will follow the inertia. 00:27
s/hates/dislikes/
cotto_work copy/pasted
dalek rrot: r48141 | coke++ | trunk/NEWS:
improve the state of the NEWS
 00:30
darbelo msg kid51 After gerd's patch in r48115, is there still a need for the tt509_install_files branch? 00:32
purl Message for kid51 stored.
kthakore um ... 00:36
hi
I am getting a segfault in t/pmc/stringhandle/t
how do I run it with parrot
so I can do it in gdb
cotto_work figure out the number of the test and run that file. 00:37
Tene kthakore: that test is actually perl 5
kthakore: so look for the generated pir files in t/pmc/stringhandle_*.pir
cotto_work ./parrot t/pmc/stringhandle_17.pir
Tene I think 00:38
gdb ./parrot
run t/pmc/stringhandle_17.pir
kthakore ah ok
cotto_work you can run all the tests via prove
-v
dalek tracwiki: v38 | cotto++ | BranchDescriptions 00:40
tracwiki: clearly more branches are in order
tracwiki: trac.parrot.org/parrot/wiki/BranchD...ction=diff
darbelo whiteknight: ping 00:42
whiteknight pong 00:43
Coke step f--
darbelo Are there any branches of yours that can be safely removed from the repo?
Coke surely there is a better way to update the bootstrapped files? 00:44
nopaste "kthakore" at 192.168.1.3 pasted "Test failes" (105 lines) at nopaste.snit.ch/22182 00:45
darbelo tt389_fix, ns_func_cleanup, hash_allocator, gc_massacre and gc_massacre_no_constant all show work by you.
And a few sound like they've been acomplished already. Unless I'm confusing them with something else.
00:45 eternaleye joined
kthakore Tene: ^^ what number test segfaults? 00:46
Tene kthakore: Um... you tell me.
kthakore Tene: heeh
Tene: nopaste.snit.ch/22182
Tene kthakore: I have no idea whatsoever. Look at the output from the test before it segvs.
kthakore Tene: it Failed test 'puts' 00:47
Tene: but I want to run that in gdb
dalek rrot: r48142 | coke++ | trunk/DEPRECATED.pod:
Add library that slipped in in r47730
rrot: r48143 | coke++ | trunk/NEWS:
mention new experimental library
rrot: r48144 | darbelo++ | branches/strings:
Remove the 'strings' branch. It only contained pseudocode, the useful parts of which where lifted into the gsoc_nfg branch.
rrot: r48145 | coke++ | trunk/CREDITS:
update my CREDITS entry
Tene kthakore: grep puts t/pmc/stringhandle*.pir
Coke ack! 00:49
Tene :P
kthakore I did 00:50
Tene kthakore: by my reading of the .pl, there's only file that should have puts in it 00:51
Coke tene: muahahah.
kthakore Tene: yeah
Tene: I found the bug
-.-
Tene ah
kthakore I didn't account for PMC_IS_NULL
Coke (which test segfaults) btw, all the individual tests are saved as individual .pir files if the top level is a .t
er, is a perl file. 00:52
kthakore Tene: may I show you my work ?
Tene kthakore: Sure.
kthakore TT 1639
Tene kthakore: I have no idea what you're working on, but sure. :)
kthakore I updated the docs 00:53
err pathc
Tene: I am swaping STRING for StringBuilder in StringHandle.pmc
Coke looks like whoever released 2.3.0 did not follow the instructions for PBC_COMPAT. 00:55
Coke ponders how to retroactivate this. 00:57
cotto_work It doesn't really matter. Just call the new version 7.0 and nuke all the 6.x changes 00:58
Coke I'm generating a fake # for the 2.3.0 release. What could go wrong? 00:59
cotto_work The end of the world. 01:00
purl well, the end of the world is Tue Jan 19 03:14:07 2038 UTC or perl -lwe 'print scalar localtime 2**31 - 1' or maybe April 9 2030 or And there were voices, and thunders, and lightnings; and there was a great earthquake, such as was not since men were upon the earth, so mighty an earthquake, and so great. or DateTime::Calendar::Mayan->new( baktun => 13) or when we'll be burning one or when my master kills purl.
cotto_work It should be documented better. 01:01
nopaste "cotto" at 192.168.1.3 pasted "documentation fix" (13 lines) at nopaste.snit.ch/22184 01:02
cotto_work ex post facto?
Coke rejected.
after the fact.
cotto_work that "ex" seem superfluous, but I don't know Latin. 01:03
Coke basically, retroactively.
DE GUSTIBUS NON DISPUTANDUM EST!
also: 01:04
cotto_work If I'm nopasting things like that, it's a pretty sure sign that I need to get out of here.
cotto_work decommutes
dalek rrot: r48146 | coke++ | trunk (9 files):
prep for 2.6
rrot: r48147 | coke++ | trunk/PBC_COMPAT:
Clear out PBC_COMPAT for 2.6 release;

entries that slipped out in the 2.3 release that were dev-only.
Coke Hab SoSlI' Quch!
leaving? 'arlogh Qoylu'pu'?
Tene kthakore: you said you figured out where it was segving? 01:05
Coke zzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz building parrot on this laptop. 01:06
kthakore Tene: one of them 01:14
purl one of them is an assignment, the other behaves like the comma operator (is this actually documented?)
kthakore Tene: let me put up more details after dinner 01:17
:D
Tene kthakore: I was just going to say, let me look at it more after dinner.
cotto ~~ 01:20
dukeleto runs 'make smoke' on trunk
01:21 jsut_ joined
Coke thanks, leto 01:21
I am still trying to remake the native pbcs. insanely slow.
not a good sign that making the native pbcs throws warnings about the wrong bytecode version. 01:27
dukeleto Coke: hmmm. yeah, i am seeing all the packfile tests fail 01:30
Coke: smolder.plusthree.com/app/projects/...ails/34982 01:31
Coke: but you are probably seeing the same
Coke dukeleto: there is an update to the pbcs that you probably need.
01:31 plobsing joined
dalek rrot: r48148 | coke++ | trunk/t/native_pbc (4 files):
native_pbc platform updates
 01:36
dukeleto updates and runs make smoke again
mikehh Coke: nope still failing at r48148 01:39
darbelo Coke: You missed mk_packfile_pbc, didn't you? 01:40
mikehh Coke: both make corevm/make coretest, make world/make test - PackFile_unpack: This Parrot cannot read bytecode files with version 6.21 01:41
Coke darbelo: what is "mk_packfile_pbc" ? 01:42
do you mean "mk_native_pbc" ?
darbelo mk_native_pbc's little brother 01:43
Coke darbelo: ... which is listed nowhere in teh release guide.
darbelo It got split up in the last release cycle.
# Also, don't forget to run tools/dev/mk_packfile_pbc to rebuild PBCs 01:44
# used for testing Packfile* PMCs.
It's mentioned in the comments at the top of PBC_COMPAT
whiteknight darbelo: Sorry about the delay. I'll dig through those branches tomorrow. Some of them can be deleted, I think
Coke ECRAPPYDIRECTIONS.
darbelo Patches welcome ;)
whiteknight but now, bed 01:45
Coke darbelo: entirely unhilarious.
so does that mean that tools/dev/mk_native_pbc should NOT be run?
darbelo No. You need to run Both.
Coke in which order?
does that matter? 01:46
darbelo It doesn't matter AFAICT. They handle different files.
I only remember it because it bit me when I bumped PBC_COMPAT in the nfg branch.
Coke UGH; they're both convolunted shell scripts that look like they do teh same (#*&$ thing. 01:47
(perhaps, as you say, to slightly different files). What a cluster. 01:48
darbelo Then I forgot about it 'till today. I *think* they do the same (slightly wrong) thing to different parts of the tree, for similar purposes.
Coke well, it's going to take another half hour to run this script. back later. 01:49
01:57 snarkyboojum joined
Coke msg szabgab the screencast is nifty, but the colors in the editor make the code unreadable for me. need higher contrast. I look forward to more 'casts. =-) 02:03
purl Message for szabgab stored.
Coke AIGHHH 02:04
I ran mk_)native_pbc again.
mk_packfile_pbc is actually VERY quick. 02:06
dalek rrot: r48149 | coke++ | trunk/t/native_pbc (4 files):
run tools/dev/mk_packfile_pbc
 02:09
rrot: r48150 | coke++ | trunk/docs/project/release_manager_guide.pod:
Be sure to mention the OTHER, similarly named, similarly-functional script
rrot: r48151 | coke++ | trunk/docs/project/release_manager_guide.pod:
-except, use the actual name of teh other file.
Coke ... worst. release. ever. 02:11
bubaflub Coke++ 02:13
mikehh Coke: looks good 02:14
02:17 snarkyboojum joined
Util Coke++ 02:23
Coke re-re-re starting fulltest... 02:24
darbelo Coke++ 02:25
On that note. I'm going to catch some z's.
mikehh Coke: All tests PASS - (pre/post-config, make corevm/make coretest, make world/make test, fulltest) at r48151 - Ubuntu 9.10 amd64 (g++) 02:32
02:34 tcurtis joined 02:39 theory joined 02:42 LoganLK joined 02:47 bubaflub joined
Coke Gej? 02:48
purl Gej is logged in
Coke thanks purl!
cotto coke++ 02:53
02:54 janus joined 02:58 cognominal joined
cotto It there any reason to use two bytes for a Lorito opcode when we'll certainly have fewer than 255 ops? 03:05
Coke given a diffstat of 03:10
202 files changed, 32012 insertions(+), 34217 deletions(-)
how would you summarize "this many lines of code changed" 03:11
diff the insert/deletions?
ash_ cotto: even java doesn't have more than 255 op's...
Chandon Coke: Sum them?
Coke cotto: one byte WFM. 03:12
Chandon Hmm... that's wrong. Maybe take the bigger one.
Coke Chandon: now that I think about it, that contains changes to generatd files. I'llj ust leave that off. 03:13
Chandon lol. Parrot just has really productive programmers that happen to be perl scripts. 03:14
03:14 theory joined
cotto I smell a release. 03:17
It smells like rump.
;)
Coke, the entry in parrothist.pod should be starred. 03:18
kthakore Tene: ping 03:20
hi cotto 03:21
hi Coke
cotto hi kthakore
kthakore cotto: I am seriously stuck again :(
oh wait ...
purl i heard wait ... was that one of his other silly conventions? placing a 'G' in front of any account that he thinks is global?
kthakore cotto: do you know the pmc stuff?
cotto: TT 1639?
cotto:
cotto: trac.parrot.org/parrot/ticket/1639#comment:5 <-------- HALP 03:22
cotto hears a cry for halp
Coke 2.6.0 tarballs uploaded to the ftp server. please build and double check shas.
cotto: I am willing to let that be wrong in the release. =-) 03:23
cotto sure. not even close to a big deal
Coke, docs updated too? 03:24
Coke cotto: RSN. 03:25
cotto rsn wfm
kthakore oh well ...
Chandon Coke: sha sums look right
kthakore I spent many hours on this
bacek_at_work: also you lied when you said it was newbie ticket :D 03:26
bacek_at_work: way too much fun to be a newbie ticket
atrodo to me, newbie ticket != easy
kthakore atrodo: yes they LIE! 03:27
dalek tracwiki: v4 | cotto++ | LoritoOps
tracwiki: remove some unneeded ops, add say to the temporary i/o ops
tracwiki: trac.parrot.org/parrot/wiki/LoritoO...ction=diff
tracwiki: v5 | cotto++ | LoritoOps
tracwiki: separate coerce and set in an attempt to minimize magic
tracwiki: trac.parrot.org/parrot/wiki/LoritoO...ction=diff
03:27 nbrown joined
kthakore atrodo: now I am knee deep in seg faults and no mem 03:27
atrodo Yea, i saw your huge malloc attempt in the log 03:28
kthakore atrodo: where?
atrodo: what log?
purl o/` it rolls downstairs, alone or in pairs, rolls over your neighbour's dog, it fits on your back, it's good for a snack, it's log, log log! o/`
atrodo goodness, i love the random things purl knows
the irc log for this channel
kthakore atrodo: it is the mirror of this channel's ops
atrodo: they are all crazy
oops 03:29
cotto purl doesn't care about ops or oops
purl cotto: what?
kthakore se '@'atrodo
sees*
cotto purl listens to everyone
purl cotto: i'm not following you...
kthakore purl: go fuck yourself
purl kthakore: what?
kthakore purl: fucking bot!!! give me more memory!
purl kthakore: excuse me?
kthakore bah
cotto: HALP?
purl I CAN HAS SOMEBODY?
cotto Please avoid profanity when abusing the bot.
atrodo purl> botsmack
purl please! another hit! just one more.... *sigh*
kthakore cotto: oh ok .. 03:30
cotto thanks
kthakore cotto: but profanity helps me ...
atrodo cotto> looks like there was a good discussion about lorito today at ps
kthakore cotto: it cleanse my soul and pumps my evil black heart
cotto atrodo, one in a long series
kthakore When is the best time to come for pmc help? 03:31
oh well ... I am going to sleep
:\\
cotto kthakore, I'm taking a look at your patch, so now.
kthakore cotto: oh ok ...
cotto what timezone are you in? 03:32
atrodo Hope so. I was going to try and join in the discussion today, but forgot the company was going to a baseball game
kthakore cotto: Canada
cotto: hehe toronto
so 2331 right now
dalek rrot: r48152 | coke++ | tags/RELEASE_2_6_0:
tag the 2.6 release.
rrot: r48153 | coke++ | trunk/docs/parrothist.pod:
2.6 was a supported release
atrodo kthakore> Eastern then?
kthakore dalek: exterminate!
purl EXTERMINATE! EXTERMINATE! EXTERMINATE!
kthakore atrodo: sure 03:33
Chandon Coke: configure/make/make test works for me with both tarballs 03:34
Coke docs.parrot.org updated 03:37
atrodo cotto> Looking over the LoritoOps page again, so would it now be set_[insp]? 03:38
cotto atrodo, that'd be a better way to write that op 03:40
kthakore, what editor do you use? 03:41
In vim, you can pretty easily fix simple formatting issues by selecting the offending code and pressing = 03:43
Coke ... no 2.3.0 announcement on parrot.org ?
cotto It probably got unstuck after the 2.4.0 announcement 03:44
cotto wishes he were smart enough to remember to use corevm *before* running make
Coke cotto: it's nowhere to be seen. 03:45
cotto kthakore, one of the problems is that buffer_size checks buffer for nullness *before* setting it.
dalek rrot: r48154 | coke++ | trunk/docs/project/release_manager_guide.pod:
don't repeat yourself.
 03:49
cotto kthakore, I'm not sure about the exact fix for the other issue, but the problem is that you need to change new_string's encoding to be whatever the encoding ATTR is. 03:50
I'm out.
happy hacking
03:50 snarkyboojum joined
dalek rrot: r48155 | coke++ | trunk/tools/util/release.json:
fix JSON typo.
 04:06
website: coke++ | Parrot 2.6.0 "Red-rumped" supported release. 04:07
website: www.parrot.org/news/parrot-2.6.0-re...d-release.
Coke release is cut. 04:10
tcurtis Coke++ 04:13
bubaflub hey y'all, just sent out an email to the list about cross compilation support (TT #1195, TT #1516) 04:14
and feedback and input welcome
Coke Chandon, whiteknight - I will attempt to compose an email to list tomorrow to continue our conversation. 04:16
Coke wondes how one gets to the front page of blogs.perl.org 04:24
dukeleto Coke++ # yay, sha1 sums! 04:28
Coke dukeleto: someone added it to the makefile. easy peasy. 04:29
zzzzzzzzzzzzzzzz
dukeleto Coke: what kind of SHA sums are they? 04:31
Coke: they look SHA256ish
but i am just guessing
Coke yes. 04:36
moderator Parrot 2.6.0 | parrot.org Log: irclog.perlgeek.de/parrot/today | Nopaste: nopaste.snit.ch:8001 | 04:38
moderator Parrot 2.6.0 | parrot.org Log: irclog.perlgeek.de/parrot/today | Nopaste: nopaste.snit.ch:8001 | fix 'make html' (talk to Coke), update tutorial (talk to tcurtis)
dalek rrot: r48156 | Chandon++ | branches/gsoc_threads (10 files):
[gsoc_threads] More green threads API.
 04:39
04:54 Coke joined
dalek rrot: r48157 | Chandon++ | branches/gsoc_threads (47 files):
[gsoc_threads] Merge from trunk.
 04:56
05:36 Chandon joined 06:06 theory joined
dalek kudo: 3176182 | pmichaud++ | build/PARROT_REVISION:
Bump PARROT_REVISION to latest Parrot release. Frozen until #31 release.
 06:08
06:11 uniejo joined
PerlJam cd 06:15
dalek tracwiki: v1 | tcurtis++ | LoritoDesignQuestions 06:16
tracwiki: Create a list of questions to consider when thinking about Lorito.
tracwiki: trac.parrot.org/parrot/wiki/LoritoD...ction=diff
tracwiki: v6 | tcurtis++ | Lorito
tracwiki: Add link to LoritoDesignQuestions</a>
tracwiki: trac.parrot.org/parrot/wiki/Lorito?...ction=diff
06:16 baest joined 06:51 AndyA joined 06:57 fperrad joined 07:08 dalek joined 07:16 whiteknight joined
dalek TT #1710 created by pmichaud++: [BUG] exit of a Coroutine affects all of its sister clones 07:26
TT #1710: trac.parrot.org/parrot/ticket/1710
tracwiki: v2 | dukeleto++ | LoritoDesignQuestions 08:13
tracwiki: trac.parrot.org/parrot/wiki/LoritoD...ction=diff
rrot: r48158 | mikehh++ | branches/html_cleanup (27 files):
merge latest trunk into html_cleanup branch
 08:15
kudo: ac0987c | pmichaud++ | src/Perl6/Actions.pm:
Fix double-gather bug.
 08:28
08:40 AndyA joined 09:06 snarkyboojum joined 09:11 dalek joined
dalek parrot: f414dbf | dukeleto++ | plparrot.c:
Check that the trusted/untrusted interpreters were created correctly before
 09:16
parrot: ecebc43 | dukeleto++ | plparrot.c:
Better debugging
09:39 gaz joined
dalek kudo: c412170 | (Kodi Arfer)++ | (7 files):
Temporal goodness.
 09:48
kthakore cotto: hi 09:56
mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#34990), fulltest) at r48158 - Ubuntu 10.04 amd64 (g++) 09:57
10:09 gbacon joined
kthakore mikehh: hi 10:27
purl hey, kthakore.
mikehh kthakore: hi there 10:30
kthakore mikehh: no else seems up yet 10:31
:D
mikehh kthakore: it's all about timezones 10:32
clock
10:32 dalek joined
mikehh clock? 10:32
purl mikehh: LAX: Wed 3:32am PDT / CHI: Wed 5:32am CDT / NYC: Wed 6:32am EDT / LON: Wed 11:32am BST / BER: Wed 12:32pm CEST / IND: Wed 4:02pm IST / TOK: Wed 7:32pm JST / SYD: Wed 8:32pm EST /
mikehh kthakore: I'm on BST, how about you? 10:33
kthakore mikehh: EST 10:34
mikehh: oh ... man .... I am sooo stuck 10:35
2 tests to fix
:(
oh well
coffee :D
then fixage :D
mikehh kthakore: good luck
purl You'll need it.
mikehh needs to reeboot - brb 10:36
10:59 mikehh joined
dalek rrot: r48159 | gerd++ | trunk (6 files):
I take release 2.9; update Fedora spec file to current release
 11:07
11:33 kid51 joined
kid51 Although nopaste is on channel, it has not worked for me for last 10 hours. 11:34
Tried to paste this both last night and this morning: nopaste.snit.ch/22189 11:35
t/pmc/packfile*.t failures on Darwin/PPC at r48151 (trunk)
dalek rrot: r48160 | gerd++ | trunk/ports/suse (4 files):
update suse rpm build port
 11:39
11:49 Casan joined
Coke kid51: that was the release, no doubt. 11:57
I suspect it's the old bytecode file crap. (I did at least verify those passed on intel linux; I don't even have a ppc box anymore.) 11:58
I think the answer is: if it's failing on your platform, YOU have to update those files. I suspect those have been failing on ppc for a while. (I hope, anyway0 12:00
kid51 Coke: Darwin/PPC was passing make fulltest as recently as Sunday. 12:14
I have just now run 'tools/dev/mk_packfile_pbc' and gotten the t/pmc/packfile*.t tests to pass.
Shall I commit the modified t/native_pbc/*.pbc files?
kid51 has to leave for $job and Rakudo talk in 5 minutes. 12:15
12:15 dalek joined
kid51 Have to go. 12:17
dalek kudo: 10a2585 | moritz++ | README:
[README] shorten section on binary packages, and refer to the website instead
 12:21
12:31 bluescreen joined 12:33 mikegrb joined 12:36 whiteknight joined
Coke I presume you updated the _4 or _2 files; you're the only one left I know running PPC, so that'll be your job, I guess. =-) 12:38
until we sort out the bytecode stuff we kind of tabled at 1.0
free karma: remove branches/rm_doc_makefile (nothing's been done, it'll be easier to start over then merge if I do that project.) 12:52
12:56 dalek joined
kthakore cotto: hi 13:03
purl privet, kthakore.
kthakore oops
13:03 smash joined
kthakore Coke hi 13:03
smash hello everyone
kthakore smash: hi 13:04
13:11 lucian joined
Coke kthakore: hio 13:18
whoa, it's Sean Burke!
(not here)
kthakore Coke: hi
purl hello, kthakore.
kthakore Coke: I am seriously stuck with a problem
Coke: TT 1639 .. stil
Coke k. I'm at $DAYJOB, so I can not, currently, be your huckleberry.
kthakore Coke: ok ok
Coke: thanks
Coke but I appreciate your work on this. it'll pay off, trust me. =-)
kthakore Coke: really! there is money and or hookers in payement? 13:20
Coke ... in your increased understanding of the internals.
I've been working on parrot for 9 years and those sorts of things don't seem to happen.
kthakore Coke: ahhh ....
Coke: it will have happen
Coke: you have to *belive*
beleive
bah can't spell
Coke: don't worry if I was looking for $money I would be at $work 13:22
Coke: oh wait ... I am :D
13:34 NotFound joined
NotFound hi 13:34
purl salut, NotFound.
13:36 Casan joined 13:37 dalek joined
kthakore hi NotFound 13:38
13:42 Psyche^ joined
Coke hello, alphabetically first committer. 13:50
14:07 bubaflub joined 14:08 dalek joined 14:12 PerlJam joined, bubaflub left, tcurtis joined
tcurtis Hello, #parrot. 14:13
moritz hi
kthakore hi tcurtis 14:14
hi moritz
moritz: I learned pmc ... and learned that I can cause 'out of mem' 14:15
moritz kthakore: oh, that's easy :-)
kthakore moritz: hehe 14:16
moritz: now I have to find where .. I am gobling up memor
memory
atrodo tcurtis++, good lorito questions 14:17
dalek tracwiki: v3 | tcurtis++ | LoritoDesignQuestions 14:28
tracwiki: Add another question.
tracwiki: trac.parrot.org/parrot/wiki/LoritoD...ction=diff
14:41 mmcleric_ joined 14:42 mmcleric_ joined 14:51 dalek joined 14:55 mmcleric_ joined 15:02 Andy joined, davidfetter joined
kthakore hi 15:04
15:06 theory joined
kthakore how can I get the string in PMC's STRING ? 15:08
15:15 Casan joined, bubaflub joined
NotFound kthakore: VTABLE_get_string 15:16
kthakore NotFound: no that just returns a STRIN* 15:20
NotFound: I want to print out the value for the STRING*
NotFound: in gdb
like printf( "%s", cahr_from_string(STRING ) ? 15:21
NotFound: btw does get_string make a clone or pass?
moritz our strings are immutable 15:22
why should it clone?
kthakore moritz: that is my question
NotFound Maybe some PMC still does a clone, but after the switch to immutable strings it shouldn't. 15:23
kthakore get_string has a clone in it for StringBuilder
NotFound kthakore: StringBuilder isn't a String.
kthakore I know 15:24
NotFound ATTR STRING *buffer; /* Mutable string to gather results */
kthakore NotFound: I am working on TT# 1639
NotFound: that is going to PMC*
NotFound kthakore: the "Mutable" is the key here.
kthakore err
NotFound: I know
see the ticket please
NotFound It uses (or abuses) string internals, so it can't just return that string. 15:25
There is a comment about that in get_string, BTW. 15:26
kthakore NotFound: did you see my ticket?
NotFound: what is the comment? 15:27
purl somebody said the comment was pretty exhaustive, read it carefully
kthakore NotFound: I am trying to use set_native_string ....
and still running out of mem
NotFound -> /* We need to clone buffer because outside of StringBuilder strings */
-> /* are immutable */
kthakore NotFound: ok
NotFound: but ...
Fuck
NotFound: how do I get it not to run out of memeory? 15:28
NotFound kthakore: current StringBuilder abuses STRING internals, but it protects the outside world from its abuses.
kthakore right
that is why I am doing TT 1639
but see this 15:29
purl i think see this is a lesson you shouldn't get involved in things :)
kthakore in io/api.c
in readline
if we have a string handler
I do this
VTABLE_set_string_native(interp, stringhandle, result);
right
NotFound io/api.c also abuses string internals last time I looked at it.
15:29 bubaflub joined
kthakore but I am gettign it to use StringBuilder 15:29
stringhandle's attrbute stringhandle is now a stringbuilder 15:30
15:30 [1]Casan joined
cotto coke++ 15:30
kthakore NotFound: have you seen the ticket?
hi cotto
NotFound I's like to do some funny things to wich choose the name of that ATTR, with a chainsaw or something. 15:31
moritz I guess the IO stuff abuses the string API, because the string API itself doesn't expose the raw buffer for easy access, or so
kthakore it does now
jeez guys
have you seen the ticket?
moritz hasn't
kthakore moritz: TT 1639
moritz TT #1639 # now the IRC logs turn it into a link, and I can access it more conveniently :-) 15:32
kthakore moritz: oh ok 15:33
moritz: wait you do irclogs.perlgeek.de right?
moritz right
kthakore moritz: can I hack a feature in your code?
moritz: it is my biggest pet peeve
moritz kthakore: which one?
kthakore moritz: new stuff on top
moritz: so reverse the irc
moritz: so on iphone
NotFound kthakore: that code is abusing StringHandle internals. If you change StringHandle internals, it booms.
moritz kthakore: works for me as an option, but not as the default
kthakore moritz: I don't have to scroll 10 gazillion light years 15:34
moritz: yeah
moritz: as an option
NotFound: yup
NotFound: now I am fixing said boom
NotFound: first booms are gone
NotFound: but readline is killing me
moritz kthakore: bonus points for doing it in client-side js, so that the server doesn't have to double the cache size
kthakore moritz: ugh ... js
moritz: ... I left that scene 6 years ago
moritz: sure fine 15:35
moritz kthakore: I said "bonus points", ot "required"
s/ot/not/
kthakore moritz: whar be the code?
moritz: I will see
moritz: please let the code be on git hub
moritz kthakore: github.com/moritz/ilbot/
kthakore :D
WOOOT!
moritz++ for seeing the light
moritz: thanks man
moritz: I will see it soonish 15:36
NotFound: is there a better way to do readline with StringBuilder?
NotFound: there should be
I mean StringBuilder should handle all the stuff in io/api.c for us
no?
moritz kthakore: you have commit privs to ilbot now. Let me know when I should pull anew 15:37
kthakore moritz: oh ok
where is the js stuff this?
NotFound kthakore: I'll kill all that code in io/api.c and forward it to a method in StringHandle
moritz kthakore: no idea, I didn't do it... probably one of the templates 15:38
kthakore NotFound: yay!
NotFound Best way to handle abuse of internals is to stop doing it.
kthakore NotFound: can you put it in my patch
NotFound: yes
NotFound: then we can handle all this junk in StringHandle
internall?
NotFound: so we just do
NotFound Internals of StringHandle abused from io/
kthakore VTABLE_readline(interp, pmc)
right? 15:39
and it checks for PMC_IS_NULL and what not?
NotFound: I just dont know how to do that yet
NotFound: first non trivial ticket that lied and said it was newbie
:D 15:40
NotFound kthakore: when people mark tickets as appropiate for newbies usually mean "I don't want to do it"
kthakore well I want to 15:41
but I have no clue ...
I just do what I tried to clean from src + docs
and the docs ... don't help
15:41 [1]Casan joined
NotFound kthakore: The StringHandle readline method currently calls Parrot_io_readline. IMO the first step will be to sanitize that by doing it the other way: let Parrot_io_readline use the StringHandle method. 15:44
When the outside world is no longer abusing StringHanlde internals, the task will be easier. 15:45
kthakore yes
NotFound: can you put some hints? cookie crumbs? in the ticket
kthakore crosses fingers
15:47 macroron joined
NotFound kthakore: the hint is easy: delete lines 420-441 in io/api.c and put that functionality in StringHandle.readline 15:47
kthakore ok
...
but how do I make VTABLE thiny?
NotFound readline is a method, not a vtable function.
kthakore ah ok 15:48
so METHOD?
kthakore looks thorugh docs
NotFound The method already exists, line 288 of stringhandle.pmc 15:49
In trunk
kthakore ah ok 15:55
NotFound: ermm have you seen it?
NotFound: it just calls io_readline
hha
hahaha
that is hilarious 15:56
NotFound kthakore: that was the point. Instead of calling io_readline from the method, call the method from io_readline
kthakore yeah
I get it now 15:57
thank you
15:58 cotto_work joined
NotFound If you fix that, and any other abuse of StringHandle internals that may happen elsewhere, refactoring StringHandle will be far easier. 15:58
cotto_work ~~
mibbit--
kthakore NotFound: there some more places
NotFound: in the mean time ... can I please accept that ticket? 15:59
NotFound: or maybe just assign me to it?
NotFound kthakore: Don't know what permissions are needed to accept. If you can, do it. If not, tell me and I'll try to assign. 16:00
kthakore I can't do it
I can just comment
NotFound What's your trac account name? 16:01
16:01 lucian joined
NotFound NM, it's in the comments. 16:01
particle i have track_admin... i can give kthakore whatever perms necessary. NotFound, can you? 16:02
if not, i'll give you trac_admin
NotFound Assigned.
particle: I'm not sure if I want to learn more trac administration %-) 16:04
particle kthakore: i changed you from 'TICKET_APPEND' to 'developer' in trac
NotFound: what's your trac id? 16:05
NotFound particle: NotFound
purl somebody said NotFound was Julián Albo <mailto:julian.notfound@gmail.com>
particle ah, sorry missed it, it's not alphabetical for some reason, just *mostly* alphabetical. sigh. 16:06
NotFound thinks about creating a mostly_sort opcode
particle NotFound: i'm sorry, you're a trac admin now ;) 16:07
NotFound I'll have to live with that, I suppose ;)
kthakore particle: oh no! I am developer.... but I don't feel like it 16:09
atrodo gist.github.com/484399 # So far, lorito is turning out magic
kthakore particle: :p
particle: thank though
NotFound The trac admin page looks better now, sure.
kthakore atrodo: does lorito change string stuff?
atrodo: liek get_string ? 16:10
atrodo don't know yet
kthakore atrodo: I thought I saw a question about it in wiki
atrodo Yea, there's some questions about do we make strings into objects, and if should use strings or symbols 16:11
kthakore trac.parrot.org/parrot/wiki/LoritoRoadmap 16:12
NotFound I read that question as: "Do we make strings mutable again? And switch again to immutable later?" }:)
kthakore NotFound: yes ... 16:13
NotFound: I see more breakage
and I just started 2 days ago
..
16:13 pmichaud joined
kthakore ok lunch 16:13
NotFound: tahnk
moritz: thanks
NotFound kthakore: Don't worry, my few weeks in parrot were mostly locating and deleting shit.
kthakore NotFound: yeah ... 16:14
purl doo wop doo wop // shang a langa shanga langaa
kthakore that is sooo cool
I just broke more sutff
hehe
I will do it after lunch
particle kthakore++ you have the right attitude to work on parrot :)
NotFound kthakore: well, the beauty of deleting unused code is that you hardly break anything :D
atrodo NotFound> I hope that's not the case. 16:15
cotto_work bubaflub, fwiw I'd love to have a separate build dir.
atrodo But, remind me, the string primitive is immutable right now, right?
NotFound (unless you fail to detect it's really unused)
dalek tracwiki: v6 | cotto++ | LoritoOps
tracwiki: rename set, atrodo++ for noticing
tracwiki: trac.parrot.org/parrot/wiki/LoritoO...ction=diff
bubaflub cotto_work: sweet! do you have a preference between these two options
perl ../Configure.pl 16:16
perl Configure.pl --build-dir=build
NotFound atrodo: is immutable, except in a few places that don't let the outside world see what are doing.
atrodo NotFound> That's what i thought
particle atrodo: spare me a commit bit for lorito? 16:17
cotto_work bubaflub, the first is a little simpler, but if you can make either work it'll be great.
bubaflub cotto_work: yeah, i'm leaning towards the first as that is how i've seen it usually done
atrodo particle> done 16:18
cotto_work That's how firefox does it and iirc llvm
particle bubaflub: the first forces perl to be run from the root of the build dir
or, expects perl to be run from...
bubaflub particle: so is that a vote for the first one?
NotFound bubaflub: the fisrt way is more autotools-alike. 16:19
particle i'm not sure, actually. i hate autotools
ok, let's look at it like this:
NotFound particle: that is a point that autotools handle well... mostly.
particle the source directory is a resource, not a place to run any code from 16:20
that statement is somewhat true.
i'd like it to be fully true.
so, that means, we shouldn't be running Configure.pl from the source dir 16:21
which means, it should be run from the build dir, so perl ../Configure.pl makes sense
does anyone disagree?
NotFound Makes sense to me. 16:22
cotto_work +1
purl 1
Coke cotto_work: ?
16:22 eternaleye joined
moritz but please also make it work from within the root dir 16:22
bubaflub moritz: of course. "normal" building won't be affected.
cotto_work Coke: what's confusing? I was responding to particle 16:23
bubaflub if configure dir == build dir then everything will still work
Coke (newbies) - honestly, notfound, I like doing those myself. I don't like sharing them. =-)
cotto_work: no, you karma'd me like an hour ago. 16:24
cotto_work release
you have to get used to that the day after a release 16:25
NotFound BTW Coke++ 16:29
cotto_work see 16:30
Coke note to self: never use a netbook for a release.
cotto_work ouch 16:31
no wonder you were complaining about the speed
Coke it worked! 16:32
purl What do you mean it worked? Did it run to completion? Did it bomb out early? Did it finish the job early? Did it tell your girlfriend "let's just be friends"? Be specific!
cotto_work If it weren't a bad idea for other reasons, I'd probably try using a 48-core machine here at work to cut a release. I'd be done in 5 minutes.
NotFound Next time use tablet, fits better with current fashion trends. 16:33
cotto_work releasing from an iPad would be awesome 16:34
mikehh Coke: Hi - recovered from release No 1, ready for release No 2?
16:35 khairul joined
cotto_work clock? 16:35
purl cotto_work: LAX: Wed 9:35am PDT / CHI: Wed 11:35am CDT / NYC: Wed 12:35pm EDT / LON: Wed 5:35pm BST / BER: Wed 6:35pm CEST / IND: Wed 10:05pm IST / TOK: Thu 1:35am JST / SYD: Thu 2:35am EST /
cotto_work good evening, khairul 16:36
khairul good morning to you cotto
Coke mikehh: yah. I'll be doing that one from my slightly better powered imac. =-) 16:43
16:44 pmichaud joined
Coke mikehh++ #updating the branch. 17:02
17:03 dukeleto joined
bubaflub morning dukeleto 17:03
Coke creates a multiline, colored, svn and git aware, unicode using bash prompt.
dukeleto bubaflub: mornin! 17:04
Coke +1 to whoever posted their bash prompt in a npaste yestereday.
cotto_work Coke, want
dalek rrot: r48161 | coke++ | branches/rm_doc_makefile:
this branch has no commits - start with a fresh branch if we continue
 17:05
cotto_work btw, iwbn if the footer on docs.parrot.org wasn't unreadable
Coke cotto_work: you do any web work? should be a simple fix to the CSS in resources. 17:06
if not, open a ticket, assign it to me. 17:07
cotto_work When I do web work, it encourages real web developers to step in.
nopaste "coke" at 192.168.1.3 pasted "coke's bashrc" (38 lines) at nopaste.snit.ch/22196 17:08
17:08 darbelo joined
Coke hokay. ticket it is. 17:08
aw, it mangled my unicode!
cotto_work done 17:09
nopaste "coke" at 192.168.1.3 pasted "actual unicodey bits pasted in directly?" (1 line) at nopaste.snit.ch/22197 17:12
17:13 cotto_work2 joined
bubaflub who is the keeper of Configure.pl and related modules? 17:18
dalek TT #1711 created by cotto++: unreadable footer on docs.parrot.org 17:23
TT #1711: trac.parrot.org/parrot/ticket/1711
17:27 tcurtis joined
Coke bubaflub: kid51 does a lot of maintenance work on Configure.pl ; I'm not sure anyone owns it. 17:28
bubaflub Coke: thanks. 17:30
smash Coke: just commented with a quick possible solution to TT#1711, i don't think i have access to the resources files to fix it myself
Coke I can probably answer some questions if you need pointers
smash: lost your commit bit?
smash Coke: is the CSS file on the repository ? 17:31
bubaflub Coke: i'm working on getting out of directory building working, and the first step is getting Configure.pl to include the right files and what have you
cotto_work docs/resources/parrot.css 17:32
bubaflub I can use FindBin on Configure.pl itself ENV{PWD} for the build path, but was wondering how i should propagate that information to the other configure modules
Coke the build dir is already in the config. 17:33
smash cotto_work: cool, thks
Coke (after a build, do "./parrot_config build_dir" 17:34
dalek rtcl: 1ccfaa1 | Coke++ | config/PARROT_VERSION:
we work with parrot-2.6.0, bump to that version.
smash Coke: fixed the file css file, do i need to do anything else besides commiting it ? 17:35
bubaflub Coke: i mean being able to build something completely out of the source directory
say i've got my source in /parrot, i can make /parrot/build and then run `perl ../Configure.pl && make&
whoops, that last & should be a `
17:36 gbacon joined
Coke bubaflub: well, that would fail today, as make needs to be run int he same dir as the config.pl 17:37
so I presume you know that and want to deal with it. =-) 17:38
bubaflub Coke: exactly. it's for my GSoC with RTEMS
dalek rrot: r48162 | smash++ | trunk/docs/resources/parrot.css:
* simple fix for TT#1711
Coke well, step one is to find where build_dir is currently being set and override it.
Coke steps 2 through several thousand is making things respect the already_present build_dir setting in the Config. 17:39
bubaflub Coke: build_dir gets set in config/auto/defaults.pm
Coke see config/init/defaults for the setting part.
bubaflub whoops, init 17:40
but i think long before that i need to make some modifications
Coke ok. so you're already a step ahead of me. what are you stuck on, now?
smash bbl
cotto_work: just changed the CSS file, not sure if i need to do anything else for it to show correcly in the website
i'll check it later 17:41
bubaflub Coke: before it gets to running those steps it dies on trying to open the VERSION file
Coke: i can easily figure out what the source directory is (FindBin on Configure.pl) and can also easily figure out where I'm building from (some dancing with $0)
i'm just wondering what the best way is to get that information to all the Parrot::Configure modules
for example, some of them are doing 'use lib'
Coke bubaflub: ... by putting it the config object with build_dir. 17:42
and then fix EVERYTHING that assumes build_dir=src_dir. which is pretty much the whole build process.
bubaflub Coke: i think the script is dying before we reach even the config object
Coke for your VERSION thing - sounds like you want to delay the VERSION check until you've sorted out src_dir vs. build_dir.
bubaflub Coke: some of the Parrot:Configure modules are including other files that assume 'lib' is one folder away 17:43
Coke bubaflub: the VERSION check looks like it's also happening in init::defaults
(once you put thigns in the config object, you have that at every step of the configure. You probably need to rewrite those use lib's as hand-written updates to INC basec on $config 17:44
bubaflub okey dokey 17:45
Coke You can probably even split out defaults into defaults and ... prelude or something.
and move the build_dir stuff into that new step (and make it the first thing called)
if that's easier than juggling the defaults code.
bubaflub Coke: this is the error I'm getting (after fixing the include stuff in Configure.pl with FindBin) gist.github.com/484834 17:50
it looks like way before a step and inside an include
that something is getting eval'd
17:51 jsut joined
dalek TT #1712 created by doughera++: Packfile failures in 2.6.0 on SPARC 17:56
TT #1712: trac.parrot.org/parrot/ticket/1712
Coke bubaflub: I have no idea why "parrot_revision" is being called outside the step. move it in to just before it's used. 17:59
bubaflub Coke: okey dokey.
18:11 rurban joined
dalek tracwiki: v4 | cotto++ | LoritoDesignQuestions 18:14
tracwiki: make questions into headings so they'll be easier to distinguish from answers
tracwiki: trac.parrot.org/parrot/wiki/LoritoD...ction=diff
dukeleto bubaflub: defining build_dir basically needs to be the first step in the entire build process, and then everything else needs to use it instead of assuming $PWD 18:22
bubaflub: may the farce be with you
bubaflub dukeleto: yep. i think i can sneak it in very early in the build setup, throw it in the $conf->options hash
dukeleto: and also with you 18:23
atrodo Do I need a bit flipped in trac to make edits to the wiki? 18:24
dukeleto atrodo: i don't think so, have you tried? 18:25
atrodo I'm looking at the LoritoDesignQuesitons page, logged in, and I don't see an edit button 18:26
dukeleto bubaflub: it also sounds like you want to write a function that does something like this : sub build_dir { File::Spec->catfile($build_dir,@_) }, so you don't have to litter the code with File::Spec calls and save yourself some typing 18:28
atrodo: very bottom left? you may have to scroll 18:29
atrodo: i don't like it, there should be edit buttons on the top, but that is where Trac puts them
atrodo dukeleto> Nope, nothing down there
dukeleto Coke, cotto_work : do y'all need to flip a trac bit for people to edit the wiki? 18:31
atrodo: this may have been caused by getting lots of trac spam
Coke yes.
to bothm. moment.
atrodo dukeleto> that would make sense
dukeleto Coke++
cotto_work Ok. I though you just had to make sure the person responded to the welcome email.
p&w? 18:32
purl p&w is tinyurl.com/23dfwut
Coke The user atrodo has been granted the permission WIKI_MODIFY.
cotto_work: except that all our trac spammers have valid, working emails. :P
dukeleto ponders that we need a bot like hugme to do stuff like giving people wiki privs and commit bits
atrodo There it is. Coke++, dukeleto++ thanks
cotto_work dukeleto: sounds like a job for aloha
dukeleto aloha? 18:33
Coke dukeleto: That would be spiffy.
cotto_work bacek's bot
Coke ... or hugme.
but it happens so often, it's not worth my time. =-)
I'll happily create an admin account for this if someone gets it working securely, though. 18:34
dukeleto maybe someone can ask nicely about hugme hanging out in #parrot ?
dalek tracwiki: v5 | cotto++ | LoritoDesignQuestions 18:47
tracwiki: add some answers that I'm reasonably sure about
tracwiki: trac.parrot.org/parrot/wiki/LoritoD...ction=diff
18:48 theory joined 18:49 patspam joined
particle hugs dukeleto 18:51
...until hugme arrives.
kthakore how do I gett and atrribute of a pmc in another pmc 18:52
I am doing
GETATTR_StringBuilder_buffer
StringBuilder has pmclass StringBuilder provides string auto_attrs { ATTR STRING *buffer; 18:54
:tabp
cotto_work That should work. You just need to include the proper header
and make sure that the makefile dependencies are updated 18:55
kthakore ah ok ...
what header do I need?
cotto_work can't wait until we have Lorito and can do all this junk programatically instead of via stupid hacks. 18:56
kthakore cotto_work: src/pmc/stringbuilder.pl has pmclass StringBuilder provides string auto_attrs { ATTR STRING *buffer;
pmc
cotto_work: oooh magical mystical lorito
cotto_work: we don't have it now though
-.-
atrodo we should tho. Magic should just happen
kthakore where is the make file deps stuff for stringhandle.pmc add? 18:57
cotto_work pmc/pmc_x.h
kthakore atrodo: we shod
so ../src/pmc/pmc_stringbuilder.pmc" ?
cotto_work just pmc/pmc_stringbuilder.pmc 18:58
kthakore ok
where doese make deps goo?
cotto_work the generated headers are in include/pmc
kthakore ah ok
cotto_work config/gen/makefiles/root.in 18:59
purl well, config/gen/makefiles/root.in is what generates parrot/Makefile
cotto_work pats purl onna head
kthakore yay more working
NotFound kthakore: I'll try to avoid that. The buffer should be a StringBuilder internal detail that no one outside should care.
kthakore n\tbut it mutates
when I do get_string
it does clone on it 19:00
OMG IT WORK! 19:01
WORKED HUGHGHGHGHHG
kthakore hugs everyone! 19:02
wooooooth!!!
cotto_work the boy seems exicitable
kthakore cotto_work: WOOOOT!
\\m/
one last testcase to fix 19:03
... ugh
stringbuilder has fixed_8
not utf8
are they the same?
cotto_work no
kthakore crap
crappity crpa
crap
cotto_work in some cases you can pretend that they're the same but in general it's a Bad Idea 19:04
kthakore but stringbuilder uses fixed_8 19:05
sigh
cotto_work darbelo would be good to ask about that 19:06
kthakore darbelo: ping. i have free hug ready just to give away :D. 19:08
Coke backs away slowly.
atrodo huh, this room is now uncomfortable for me 19:09
kthakore huggles Coke and atrodo real quick. Ninja liek :D
atrodo drat, and me without my ninja repellent today 19:10
kthakore hugging Coke gives me a high!
cotto_work ok. That's weird.
kthakore cotto_work: well snorting Coke should too 19:11
cotto_work: any manner of contact with Coke can give you a high.
:d
:p
Coke O_O;
cotto_work I've met him. It didn't work.
no hugs though 19:12
Coke You don't want YAPC SARS, anyway. 19:14
kthakore hehe
cotto_work: well for some Coke brings you down 19:15
cotto_work: its sad really
cotto_work pmichaud: ping
kthakore wth ... 19:16
when I run it outside the harness I get utf8 19:17
-.-
19:20 AndyA joined 19:24 bluescreen joined 19:38 ash_ joined
kthakore um ... how is StringHandle open even ever called? 19:42
if I do
ifh = new ['StringHandle'] 19:43
ifh.open('temp_file')
it calles Perl_io_open
not the StringHandle.open methead
cotto_work That's odd. 19:45
seen allison or chromatic 19:46
seen allison 19:47
purl allison was last seen on #parrot 12 days, 22 hours, 47 minutes and 23 seconds ago, saying: NotFound: that removes the roadblock, so someone can work on it when inclined [Jul 8 20:59:36 2010]
cotto_work seen chromatic
purl chromatic was last seen on #parrot 21 hours, 37 minutes and 11 seconds ago, saying: TT #1639, dukeleto.
kthakore cotto_work: ok I will update the ticket more ... 19:48
cotto_work: hey that is the ticket I am working on 19:49
trac?
purl well, trac is a web-based software project management and bug/issue tracking system emphasizing ease of use and low ceremony. It provides an interface to the Subversion revision control systems, integrated Wiki and convenient report facilities. projects.edgewall.com/trac/ or Python, SQLite and ClearSilver or killing killtrac or a bug-tracking tool or at trac.parrot.org/parrot/ or slow or REALLY slow
cotto_work That may be the correct behaviour. I can't say for sure. It's just surprising to me.
kthakore cotto_work: me too 19:51
cotto_work kthakore: in Parrot_io_open, there's a special case for StringHandle 19:52
else if (new_filehandle->vtable->base_type == enum_class_StringHandle) {
kthakore I know
it just returns
nothing esle
19:53 ash__ joined
cotto_work that TODO comment isn't entirely clear 19:53
kthakore line 151 ?
there is not TODO there 19:54
cotto_work 136
looks like whiteknight's doing
kthakore but that makes sense ... 19:55
I am not sure why stringhandle even needs open 19:57
cotto_work me neither
That call to Parrot_io_open is a noop.
kthakore yeah 19:58
but why the hell was Stringhandle.open even made?
cotto_work so it has the same interface as FileHandle
If you use FileHandle, you can substitute a StringHandle for easy testing with no additional changes. 19:59
It's nice that way.
kthakore ok ...
that seems pointless to me ... but I don't know much about parrot 20:00
NotFound open sets the r/w mode
kthakore NotFound: but it is NEVER CALLED 20:01
see Perl_io_open
for stringhandle
NotFound kthakore: I doubt thatm
kthakore around line 151
NotFound: I walked trough the code line by line
....
brain hurts now
NotFound tapir2.ro.vutbr.cz/cover/cover-resu...e-pmc.html
kthakore ?? 20:02
cotto_work the Parrot_io_open might be there just for the flags.
NotFound 100% coverage
purl 100% coverage is impossible. 80% is trivial.
kthakore ok
cotto_work: so where is the encoding changing?
I don't see it
in readall?
cotto_work not sure 20:03
kthakore can open be called on null pmc?
what is the point of doing string_result = string_make(INTERP, "", 0, "unicode", 0); 20:04
if encoding_is_utf8 20:05
NotFound kthakore: because utf8 is the default encoding for unicode charset and people are lazy.
kthakore that shouldn't mess with encoding though 20:06
right
I mean we don't set the encoding to anything
NotFound And because string_make is obsolete and should be replaced.
kthakore *sigh*( 20:07
wth ...
ok
what should I use?
NotFound kthakore: unfortunately there is no good replacement. Just don't be too surprised by the ugliness of string creation. 20:09
kthakore ok .. 20:10
sigh
I am done
how do I leave messages? 20:12
maybe I shouldn't
oh well
cotto_work msg kthakore like this 20:13
purl Message for kthakore stored.
kthakore cotto_work will people hate me for leaving a silly message like 'HALP' at TT 1639 ? 20:14
cotto_work you should be as specific as possible 20:15
kthakore HALP PLEASE?
purl I CAN HAS SOMEBODY?
kthakore cotto_work: ok
wait who should I talk to about this stuff?
fixed_8 stuff
cotto_work svn blame will tell you who last modified any given chunk of code 20:16
if you can't find out any other way, that's a good start
kthakore cotto_work: but I don't even know where this is?
cotto_work ack for some relevant-looking code and ping whoever svn blame points at 20:17
kthakore where is it set?
cotto_work or grep
kthakore I did
ack -l fixed_8
only thing I see is string/api.c
but ... 20:18
cotto_work src/string/encoding/fixed_8.c
kthakore I know
but where is that called
I meah ...
I went true each line of code in gdb
don't see a change to encoding of the stringhandle
oh well 20:19
gtg home 20:20
will try tommorow
20:27 theory joined 20:35 tcurtis joined 20:47 eternaleye joined 20:50 jsut_ joined 21:00 [1]Casan joined
dalek nxed: r569 | NotFound++ | trunk/winxedst1.winxed:
make sure that float constants looks like that in generated pir
 21:06
21:19 snarkyboojum joined
dalek kudo: 1d4ba47 | (Martin Berends)++ | Test.pm:
[Test.pm] use pir::time__N for the fractional seconds removed from time
 21:28
21:40 theory joined
dalek TT #1713 created by coke++: helper script to find committers. 22:22
TT #1713: trac.parrot.org/parrot/ticket/1713
kthakore how can I get the value of a STRING* in gdb? 22:45
like
cotto_work p s->strstart
kthakore cotto_work: thanks man
cotto_work just don't do that in c code
kthakore cotto_work: what would I do with out you guys
cotto_work flail for a bit and figure it you a little less efficiently 22:46
;)
s/you/yourself/
kthakore haha 22:53
kthakore flails around
woopwoopwoopwooopwooopwopp
tcurtis cotto_work: concerning method dispatch in Lorito, you said that MMD would be implemented on top of Lorito. What about single dispatch, though? 22:54
cotto_work That might take an optimized path, but it'd go through the same code. 22:55
22:55 theory joined
kthakore cotto_work: um ... p s->strstart is a PASM 22:59
...
how do I get the "string"
tcurtis cotto_work: alright. 23:00
cotto_work kthakore: you're not making sense
kthakore ok 23:01
cotto_work: say I have a string
tcurtis kthakore: what does p s->strstart output?
kthakore like "foo"
tcurtis: ask cotto_work
then I set it to a STRING* 23:02
tcurtis kthakore: I mean, what is it outputting for you?
kthakore 0xf00000 (PASM)
dalek nxed: r570 | NotFound++ | trunk/winxedst1.winxed:
avoid generating and using unneeded temporary registers in some cases
kthakore ok screw this
I am not explaining
what I am doing
ok
tcurtis kthakore: I think that means that the contents of the string are "PASM". 23:03
kthakore tcurtis: cotto_work: I am working on (still) TT 1639
tcurtis: ah ok
I am trying to find when fixed_8 is set
as the encoding
23:04 theory joined
kthakore but it goes all the way into filehandle.pmc 23:04
cotto_work in that case, the contents of the string s are "PASM"
kthakore instead of stringhandle.pmc
cotto_work: ok
cotto_work That's an odd address though
kthakore cotto_work: trac.parrot.org/parrot/ticket/1639#comment:8 23:05
cotto_work: that was just random
cotto_work: I just made it 0xf00000
the test is t/stringhandle.t 23:06
23:06 [1]Casan joined
kthakore test is in regards to readall() - utf8 on opened stringhandle 23:07
I updated the ticket 23:10
cotto_work in the future, don't do that. A memory address doesn't reveal anything valuable about your computer 23:14
copy/paste is fine
kthakore, did you intend to change a test with that patch? 23:21
23:39 theory joined 23:40 theory_ joined 23:45 theory joined 23:57 Psyche^ joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!