Parrot 1.0 Released | parrot.org | 367 RTs left!
Set by moderator on 12 April 2009.
dalek rrot: r38139 | jkeenan++ | trunk/t/codingstd/svn_id.t:
Allow for non-testing of files which cannot contain an SVN Id tag. Cf.: trac.parrot.org/parrot/ticket/554.
 00:00
allison (I was developing the book outside the Parrot repository to avoid some of the integration issues, but it was getting difficult to keep up with changes on trunk, so just developing the book in trunk seems the less-bad approach)
dalek rrot: r38140 | whiteknight++ | trunk/src/hll.c:
fix up some of the documentation in src/hll.c. No substantive changes
 00:03
00:04 tetragon joined 00:06 jimk joined
allison back in a minute (rebooting) 00:06
00:07 jimk joined 00:08 jimk joined 00:09 AndyA joined
Whiteknight Tene_: If I can figure out where the PMCs need to be getting found for the HLL, I'll be able to figure out how to get them there 00:09
jimk OT: If I've had an interruption in my Net connection and my IRC client switches to an alternative nick because my primary nick is still "in use", how can I get back to my primary nick? 00:10
jimk => kid51
Whiteknight jimk: Some servers have an ability to ghost an unused nick, I dont think irc.parrot.org 00:11
jimk wants my identity back! Stolen by Verizon! 00:12
Whiteknight jimk: kid51 has already been kicked 00:13
jimk Whiteknight: thanks!
00:13 kid51 joined
kid51 breathes sigh of relief 00:14
00:16 bacek_ joined 00:20 allison joined
Coke allison: =for PIRTODO ? 01:24
=for INVALIDPIR ? 01:25
=for ENOTPIRYET ?
allison Coke: those work 01:26
though, were most of them out-of-date PIR, or PIR that's a fragment and can't run alone?
Coke: it may be PIRFRAGMENT for things that won't change in the text, but can't be tested directly 01:27
Coke: saving PIRTODO for code examples that need to be changed
Coke I think I had only marked things that were broken. 01:28
as opposd to things that were not yet testable.
allison Coke: okay, that makes sense
Coke as long as ones marked PIR are still marked PIR, we won't backslide.
I'm just trying to raise the bar a bit and insure we don't backslide.
allison Coke: yes, I didn't remove the PIR markers (I just added PIR as a valid 'target' so they would appear in the generated LaTeX) 01:29
Coke sehr gut
allison so, I'd add PIRTODO as another valid target
how about the extra text?
I think you included the ticket number for some examples on the =begin line? 01:30
Coke nice to have, but meh.
allison Coke: well, you could always put a =for TODO
Coke you could put it in as a ... right.
allison just before the =begin
that'll be completely skipped by the formatter, and won't appear in the text 01:31
Coke that's fine.
if the test was clever, we could key off that instead of needing a separate target for PIR good and PIR bad.
cleverer.
allison certainly possible, yes would likely need more work 01:32
01:34 mikehh joined
nopaste "kid51" at 71.247.45.119 pasted "Can these lines from the end of lib/Parrot/Op.pm be deleted?" (18 lines) at nopaste.snit.ch/16284 01:39
allison kid51: yes, we no longer use LICENSE and COPYRIGHT in the Pod sections 01:40
dalek rrot: r38141 | jkeenan++ | trunk/lib/Parrot/Op.pm:
LICENSE and COPYRIGHT no longer used in Pod sections.
 01:42
kid51 t/codingstd/copyright.t has a SEE ALSO docs/pdds/pdd07_codingstd.pod. But I can't find any guidelines about copyright notices in that file -- nor in any other docs/pdds files. 01:48
Is our policy documented somewhere?
allison kid51: codingstd.pod is the canonical documentation location 01:49
kid51: IIRC it used to have LICENSE and COPYRIGHT sections required in the Pod, and we removed them 01:50
kid51: but, we didn't add a notice to explicitly *not* include LICENSE, COPYRIGHT, and AUTHOR sections
kid51 But currently there is nothing whatsoever about copyright in pdd07_codingstd.pod. I simply want to remove or correct the SEE ALSO in copyright.t. 01:52
allison kid51: docs/pdds/pdd07_codingstd.pod doesn't say anything at all about copyright? 01:54
kid51: not even the standard policy of including the copyright statement as the first line of the file? 01:55
kid51 Not AFAICT. I searched for Copyright copyright and right -- and found only one instance of the latter, as a word by itself.
dalek rrot: r38142 | allison++ | trunk/docs/pdds/pdd07_codingstd.pod:
[pdd] Add a note on License and Copyright sections in Pod documentation.
 01:56
allison kid51: you're right, it doesn't. It should. I'll add that too. 01:57
(dalek is slow, I made that commit several minutes ago)
kid51: okay, I added a section on copyright notices to pdd07_codingstd.pod 02:05
dalek rrot: r38143 | allison++ | trunk/docs/pdds/pdd07_codingstd.pod:
[pdd] Add a description of our coding standards for copyright notices in
Coke allison: shebang trumps copyright. 02:09
02:09 wayland joined
dalek rrot: r38144 | jkeenan++ | trunk/t/codingstd/copyright.t:
Refine test for duplicate copyright statements to exclude files where the

failing files to the four files with Melvin Smith copyright notices.)		 02:28
moderator Parrot 1.0 Released | parrot.org | 365 RTs left! 02:29
kid51 Coke: Do you want we to fix that shebang thing? 02:30
02:35 janus joined
Coke a clarification couldn't hoit. 02:36
02:39 eternaleye joined
dalek rrot: r38145 | jkeenan++ | trunk/docs/pdds/pdd07_codingstd.pod:
Copyright statement goes on line 2 if line 1 is a shebang line.
 02:42
02:49 Tene joined 03:14 bacek joined
dalek rrot: r38146 | pmichaud++ | trunk/compilers/nqp (3 files):
[nqp]: Allow argumentless method calls w/o parens.

near future -- this just solves an immediate need.)		 03:38
03:47 MariachiElf joined, MariachiElf left 03:54 MariachiElf joined 04:09 MariachiElf left 04:11 tetragon joined, uniejo joined 04:39 msmatsko joined 04:57 TiMBuS joined 05:14 iblechbot joined
dalek rrot: r38147 | pmichaud++ | trunk (1 files):
Add tools/dev/create_language.pl, a substantially modified version of tools/dev/mk_language_shell.pl .

   * Languages now go to any path, not just languages/foo/
   * No longer relies on Parrot's configuration tools
   * Uses an independent t/harness (no longer requires Parrot::Test::Harness)
   * Supports the --gen-parrot and build/PARROT_REVISION options (from Rakudo)
   * Switches src/parser/* to be src/pct/*
   * Converts actions.pm to use $<match>.ast instead of deprecated $($<match>)
There are still updates needed to make this work well with installed parrot and provide a 'make install' target -- these will be added shortly.		 05:23
rrot: r38148 | pmichaud++ | trunk/tools/dev/create_language.pl:
Remove rakudo-specific item from create_language.pl .
 05:39
allison Coke: ah, good point on shebang. Thanks kid51 for the fix. 05:46
05:54 ruoso joined
Tene pmichaud++ 06:09
06:15 japhb joined 06:49 flh joined 06:56 iblechbot joined 07:34 iblechbot joined 07:52 Counterspell joined 08:32 Counterspell joined
eiro hello 08:32
08:34 Counterspell left
cotto Hi, eiro 08:37
runtime/parrot/dynext/digest_group.so is surprisingly musical when catted to /dev/audio 08:46
wayland It's very loud. I jumped out of my skin, and getting my skin back on is proving to be difficult 08:48
cotto Yeah. You'll want to turn your volume way down.
Sorry I didn't mention that.
At least you're awake now. 08:53
08:54 braceta joined
wayland But I'd like to be able to concentrate :) 08:57
cotto Are you at work (or otherwise working)? 09:03
wayland Sort of, yes :) 09:08
But I'm at home.
Does that answer the question? :) 09:09
dalek rrot: r38149 | fperrad++ | trunk/lib/Parrot/Docs/Section/Parrot.pm:
[docs] update book section
 09:10
09:34 Theory joined
dalek rrot: r38150 | fperrad++ | trunk/lib/Parrot/Docs/Section/Parrot.pm:
[docs] fix book reference
 09:49
rrot: r38151 | fperrad++ | trunk (4 files):
[install] linking with an installed Parrot
09:51 flh joined
bacek good evening. 09:55
Infinoid: packfile attributes sets to default values by PackFile_new. And because it's impossible to write PBC in format other than "native" almost all setters are useless... 09:59
Infinoid: except uuid_type. I just forgot to propagate it from PMC to PackFile... 10:00
Infinoid: (about "extend Hash") There is few TODO items to check actual types of argument on set_pmc_keyed_*. And for most pmcs only one set_SOMETHING_keyed_KEY makes sense. So it's easy (and better from point of view) implement is in terms of "has-a", not "is-a". 10:03
10:04 megazooch joined
dalek rrot: r38152 | bacek++ | branches/packfile_revamp/src/pmc (6 files):
Some minor documentation stuff. Infinoid++
 10:06
10:06 bsdz joined
dalek rrot: r38153 | fperrad++ | trunk/docs/compiler_faq.pod:
[book] fix POD format
 10:22
10:33 TiMBuS joined
bacek summon jonathan 10:41
wayland I think all the magic is in Oslo, and we'll have to do without :)
bacek wayland: yeah. But I really want his sign-off on pdd change before brake everything :) 10:44
11:12 iblechbot joined 11:52 NotFound joined
NotFound hi 11:52
Util Hi, NotFound 12:04
dalek rrot: r38154 | NotFound++ | trunk/src/string/encoding/ucs2.c:
some refactoring of encoding/ucs2 for builds without ICU, TT #557
 12:05
rrot: r38155 | fperrad++ | trunk/lib/Parrot/Op.pm:
[codingstd] add missing code coda
 12:15
rrot: r38156 | fperrad++ | trunk/tools/dev/create_language.pl:
[codingstd] remove cuddled else
rrot: r38157 | fperrad++ | trunk/compilers/nqp/src/Grammar/Actions.pir:
[codingstd] remove trailing space
12:33 uniejo joined 12:36 rg1 joined 12:43 ruoso joined 12:53 amoc joined
Infinoid bacek: PackFile_new sets the attrs in the actual pbc header correctly, but nothing sets the packfile pmc attrs in some cases. When loading from a file (or string) they are correct, they are 0 the rest of the time. 12:55
I think that won't matter in most uses. So it's probably ok if we just document it 12:56
it was just something I noticed that looked a little weird. otherwise, the packfile pmcs in your branch look good :) 12:57
13:12 acajou joined 13:16 gryphon joined 13:40 iblechbot joined 13:58 wayland76 joined 14:00 Andy joined
wayland76 allison: if you have a moment, I've updated trac.parrot.org/parrot/ticket/503 14:04
15:00 uniejo joined 15:10 LylePerl joined 15:23 jq joined 15:31 acajou left 15:36 slavorgn joined 15:50 davidfetter joined 15:52 claes joined 15:59 braceta left
particle- Coke: ping 16:19
LylePerl Hi (anyone) 16:32
I've been trying to learn a bit of parrot and documenting what I've covered into a guide
mikehh any progress on TT#551 - I am still failing t/pmc/nci.t at r38157 - Failed tests: 66, 69-70 - on Ubuntu Intrepid i386
LylePerl Looking at if statements in the examples 16:33
mikehh all 3 tests fail with a Segmantation Fault when I run them with ./parrot
LylePerl it seems to always be "if CONDITION goto LABEL" 16:34
So the else if just the next line...
Is that always the case? I couldn't find any docs on conditionals
mikehh LylePerl: thats PIR which is effectively syntactic sugar over PASM 16:36
LylePerl mikehh: Yes, sorry I meant I'm trying to learn PIR 16:37
PIR seems hard enough :s
mikehh It's essentially high level assembler 16:38
particle- lyleperl: there's also "unless condition goto label"
LylePerl particle: I saw that... 16:39
so it's just one or the other? Are there any docs on it?
particle- yes, of course!
have you tried make html?
LylePerl just running now. I'd only been looking at docs.parrot.org and google... 16:41
particle- see docs/book, docs/user/pir, examples/tutorial, etc. 16:42
LylePerl ahh yes. I'll read through these. Thanks :) 16:44
mikehh BTW a lot of the book stuff has been completely revised in the last 48 hours or so by allison - I have only just started going through it 16:49
LylePerl I'll make sure I grab the latest 17:04
Coke particle-: pong, catching up in review. 17:23
I am reminded we want to fix make html. :| 17:24
er, not fix, but run it daily at least on docs.parrot.org 17:25
particle- yeah, there you go. 17:26
also, i'm thinking about a drupal plugin to display current test stats for rakudo on rakudo.org 17:27
you have some experience with the google calendar plugin on parrot.org
17:27 wayland76 joined
Coke that's just an html snipet. 17:29
"snippet
particle- ah
i need something that will, say, read the last line of a csv 17:30
then parse and emit html
PerlJam if you're an admin-type on drupal you can write arbitrary PHP in various places. reading CSV and emitting HTML isn't that big of a deal in PHP :) 17:36
Coke Sounds like PJ needs admin rights. 17:42
particle- is there a unix command to read a file into the current copy buffer? 17:49
wayland76 You mean xwidows copy buffer? 17:50
particle- well, hrmm. i'm using putty
guess it won't work anyway.
wayland76 not with putty 17:51
particle- lemme see if i can convince vim to insert the contents of a file into my current location
wayland76 have you considered psftp/pscp?
I'm pretty sure you can do that
isn't it :r <filename> or something?
I haven't used vim in years :)
Infinoid if you're sshed into a unix box, screen has builtin cutpaste too.
wayland76 Oh, yeah, that would be useful. I don't use screen, but I know why people like it :) 17:53
Coke wayland76: yes, :r
wayland76 I think I was 18 when I last used that :)
Coke what, vi?
I'm old school and use it for most of my editing. 17:54
wayland76 Nono, the :r command
I've used vi on and off
I use nano for editing, and pine for e-mail
Coke we were forced to learn it to work on the mainframe. =-)
wayland76 I can use it if I need to, I just don't make a regular habit of it
I once worked on a machine without even vi 17:55
using ed for editing was no fun :)
Oh, no, that's right, it had vi, but something about the terminal was broken 17:56
so vi didn't work
Infinoid such fun. 17:57
Coke looking for perl work? jobs.perl.org/job/10462 18:09
allison particle: hmmmm... I always use :split and then Y in one file and, P in the other file (Ctrl-w-w to switch windows)
Coke (note the email contact)
PerlJam go catalyst-porn 18:10
wayland76 allison: if you have a moment, I've updated trac.parrot.org/parrot/ticket/503 18:12
allison wayland76: okay, thanks, taking a look 18:14
wayland76 I may have confused my svn, though, in which case the patch could be incomplete 18:15
dalek rrot: r38158 | allison++ | trunk/docs/book/ch02_getting_started.pod:
[book] Update names of chapters in listing.
allison wayland76: it looks like this patch thinks ports/suse/parrot.spec and ports/fedora/parrot.spec already exist 18:24
wayland76 Yeah, that's what I thought. My svn checkout must be confused. I'll fix that and do it again. 18:25
18:28 bsdz_ joined
allison wayland: you can just attach the three files if that's easier 18:43
18:46 bsdz_ joined
wayland76 Well, I'm almost done now :). 18:46
allison wayland76: it's good svn experience :) 18:49
wayland76 Exactly. But I just moved the old one aside, and did a new checkout :) 18:50
bsdz_ hi, any one know if i can add a vtable to a builtin pmc. nopaste to follow? 18:51
nopaste "bsdz" at 78.105.78.157 pasted "add vtable function to a builtin integer PMC" (23 lines) at nopaste.snit.ch/16291
GeJ Good morning everyone 18:53
bsdz_ good evening GeJ 18:54
allison bsdz_: you'll need to subclass Integer 18:58
bsdz_ okay thanks. 18:59
nopaste "NotFound" at 213.97.96.43 pasted "add vtable function to a builtin integer PMC" (24 lines) at nopaste.snit.ch/16292 19:02
NotFound bsdz_: something like that? 19:03
19:03 AndyA joined
bsdz_ NotFound: looks good. i'll give it a try 19:03
NotFound bsdz_: note the 'self' parameter in the function 19:05
wayland76 allison: Try it now :)
bsdz_ NotFound: thanks. is there any way i can code to call my new class in place on Integer without having to replace Integer for 'NewClass' name? 19:07
NotFound bsdz_: if you are writing a language you can set an hll mapping 19:09
bsdz_ oh right. is there any pointers anywhere? 19:10
er, i think i need to brush up my english! i mean where can i find more info on hll mappings? 19:11
allison waylan76: thanks, looks good, applying 19:12
bsdz_ ah found it in HLLs and Interoperation
NotFound bsdz_: you can take a look at t/op/box.t for example
bsdz_ NotFound: thanks again
19:14 claes joined
dalek rrot: r38159 | allison++ | trunk (7 files):
[cage] Split out separate spec files for different distributions.
 19:16
19:40 particle1 joined
wayland76 allison: I've just discovered another reason why doing what we just did with the spec files is a bad idea 19:49
allison wayland76: what's that?
purl That's the money you could be saving with Geico.
wayland76 When you run rpmbuild -ta parrot-$version.tar.gz, it tries to use all the spec files at once 19:50
allison wayland76: but none of the spec files are in the tarball
oh, the rpmbuild target is broken
should be deleted
wayland76 Well, I just build from SVN anyway :)
I guess I'm evil :) 19:51
allison wayland76: ok, you're not using 'make rpms'? 19:52
wayland76 no, not at all
My theory is that I should be able to take a distribution tarball, and run "rpmbuild -ta parrot-$version.tar.gz" on it, and have it work 19:53
That's what RPM-using people expect to be able to do
allison wayland76: rpmbuilders shouldn't be working from trunk, they should be working from a release tarball
wayland76 Exactly
allison that is, a release tarball that they downloaded from parrot.org 19:54
wayland76 That's what I meant by a "distribution tarball"
exactly :)
allison then, I don't see the problem? you grab the tarball, grab the spec file for your distribution, and run rpmbuild
wayland76 Ah, but the spec should be in the tarball, and we should be able to rpmbuild -ta $tarball :) 19:55
(at least from an RPM user's point of view :) )
allison wayland76: debian has a similar vanity, but it doesn't work because the package specifications are different for debian and ubuntu, but both expect a "debian/" directory at the top level 19:56
so, we're distribution agnostic. we don't include any packaging files in the releases 19:57
wayland76 Hmm. Well, I'm not so keen on that, but I can live with it for now, I guess :) 19:58
19:58 Khisanth joined
PerlJam wayland76: If your expectations are common, then that sounds like FAQ fodder. (or at least a mention in the README) 19:59
wayland76 Well, a common specfile that works across distributions (works, but may not be "ideologically" sound) is what others seem to be doing 20:00
And if people don't like it, they can package it into a distro and write a separate spec file
PerlJam wayland76: Like in a contrib dir or something? 20:01
wayland76 My understanding of the distros is that they have their own packaging files that don't necessarily get integrated upstream
Infinoid So where would such a common specfile go? 20:02
wayland76 ie. Parrot provides a spec file, and Fedora ignores us, and makes their own different one
Infinoid: Anywhere you like. I had 3 in 3 different directories (supposedly for different distros), and it tried to use all of them, which didn't work 20:03
allison wayland76: you must have been trying to build from parrot trunk 20:10
wayland76: we can rename the files to 'fedora.spec', 'suse.spec', and 'mandrivia.spec', if that at least means that rpmbuild doesn't try to do the wrong thing 20:11
wayland76 Yes, and I know I'm not supposed to do that. I'm just making the point that RPM is quite capable of finding the specfile anywhere in the tree
I'd recommend I investigate what RPM does under those conditions before we try that 20:12
allison wayland76: yes, some general investigation is in order 20:13
wayland76: especially, talking to the distributions to see what they want (if they're willing to contribute back their spec files, etc) 20:14
wayland76 I think they'd be happy to if we wanted them, as long as they didn't have to allow for other distributions 20:16
mikehh doesn't Opensuze provide a build service where you can build for virtually all platforms - certainly Linux?
wayland76 Someone on #gtk-perl was enthusing about that 3 minutes ago
allison wayland76: yes, that's what I expect, and having a copy of the actual spec files seems quite useful 20:17
mikehh: someone offered that for Parrot...
mikehh: have to dig back through messages
yes, see the message "Parrot RPMs for SUSE" on march 19th 20:18
wayland76 The guy on gtk-perl said he was using the SuSE build system on Ubuntu :) 20:19
20:19 msmatsko joined
allison wayland76: would you be interested in working with Michal setting up the build service? 20:19
mikehh AFAIK they offer it for open source projects, maybe others, but you can also get Debian, Ubuntu, Red Hat etc. 20:20
allison wayland76: it would certainly supply us with more reliable SuSE packages
wayland76 allison: Don't ask me now -- I've been up all night, and don't feel like doing anything at the moment :)
allison Ubuntu already has a package testing service in their PPMs
wayland76: well go get some sleep! :) 20:21
wayland76: I'll just let him know we're interested and will look for someone to volunteer
wayland76 Basically, once I get a Parrot RPM that I can build a Rakudo RPM on, and have contributed that back to Parrot (and cleaned up any messes I leave), I expect to disappear from the Parrot community 20:23
I only have a limited amout of time, unfortunately
But I might be able to help in bits and pieces 20:24
claes in PGE, is there a predefined rule or other way to match that a character belongs to a unicode category? 20:26
allison wayland76: makes sense
wayland76 If I'm needed after I disappear, look for me on #perl6 20:27
Coke claes: I am not sure if PGE has that yet, but the syntax will no doubt be the same as svn.pugscode.org/pugs/docs/Perl6/Sp...-regex.pod (trying to find it in there now) 20:28
claes Coke: I read that but couldn't find it
Coke looks like <+isLu> == "is an uppercase letter".
claes lovely, I'll try that.. thanks =) 20:29
davidfetter HAI
Coke claes: I know pmichaud has only recently beefed up unicode handling in PGE; pretty sure there's more left. 20:30
(to do)
claes Coke: ok, I can ask him tomorrow if it doesn't work. He's here at NPW 20:31
Coke tell him I said hi. =-) 20:33
PerlJam claes: ask him when the first rakudo 1.0 release is going to be ;> 20:34
claes I went to his PCT talk and felt a strange urge to play with it afterwards
wayland76 allison: Renaming the files doesn't help. It found "whats_this_oddity.spec" just fine 20:35
claes: Are you calling us strange? :o) 20:36
claes wayland76: strange in the the positive way
wayland76 Because we all have that urge here 20:37
Well, maybe I exaggerate
claes so as an excersice I'm writing a JS compiler for it 20:39
mikehh en.opensuse.org/Build_Service 20:42
Coke wayland76: perhaps rename them to spec.whats_this 20:43
or whats_this_spec
wayland76 Yeah, I su"spec" that would work :) 20:44
Yup, that works
Coke I can't believe you'd use that pun on this unsuspecting crowd. 20:45
wayland76 It presumably finds *.spec
allison so ports/suse/parrot.spec.suse?
wayland76 Well, I believe in pun-ishment :)
Coke next time be more circumspec.
wayland76 allison: I'll try it now
20:45 uniejo joined
wayland76 I exspec I might be able to manage that 20:45
allison: That works. I'll presume you don't need me to send a patch to move those files? 20:48
allison wayland76: yeah, I'll just move them directly in svn 20:51
wayland76 :)
20:56 slavorgn joined
dalek pp: 76648c9 | (Bernhard Schmalhofer)++ | t/embed/eval.t:
Simplify path to included lib.
 21:01
pp: 6c7b42d | (Bernhard Schmalhofer)++ | docs/internals.pod:
Fix typo. s/und/and/.
shorten dalek's url is at xrl.us/bepagy
shorten dalek's url is at xrl.us/bepag2
dalek rrot: r38160 | allison++ | trunk/ports (0 files):
[cage] Renaming the various RPM .spec files so running 'rpmbuild' from

the fix.		 21:10
purl well, the fix is to just override new
rrot: r38161 | allison++ | trunk/config/gen/makefiles/root.in:
[cage] Remove the 'make rpms' target.
Infinoid poor bot. 21:11
21:18 slavorgn joined
bacek good morning 21:43
Infinoid hi bacek
bacek hi Infinoid
Next commit special for you :)
Infinoid ooh 21:44
bacek r38162
dalek rrot: r38162 | bacek++ | branches/packfile_revamp (2 files):
Initialize Packfile attributes to system default as suggested by
 21:46
purl hmmm... infinoid is Mark Glines <mailto:mark@glines.org>
bacek poor bot... 21:47
Infinoid purl: ignore dalek
purl Infinoid: huh?
Infinoid purl: please ignore dalek? 21:48
purl infinoid: i haven't a clue
Infinoid purl: sudo ignore dalek!
purl Infinoid: excuse me?
Infinoid oh well.
wayland76 purl: get a clue
purl wayland76: what?
Infinoid purl: clue?
purl i guess clue is "one misfit", what're the available letters or amagosa's enemy or www.bbc.co.uk/radio/aod/radio4_aod....adio4/clue or www.bbc.co.uk/radio/aod/genres/come...?bbc7/clue or owner of the rudest jokes on day time radio or www.ivorysky.com/isihac/samantha-gr...-36-39.php
wayland76 purl: help! 21:49
purl help! is, like, WATCH OUT THEY'RE COMING FOR YOU
Infinoid purl: Here's your sign. </JeffFoxworthy>
purl Infinoid: sorry...
wayland76 purl: what do you do?
purl wish i knew, wayland76
wayland76 purl: list
purl hmmm... list is at news.gmane.org/gmane.comp.lang.perl.moose
wayland76 purl: argh!
purl argh! is the wiki still down?
Infinoid hah 21:50
forget argh!
purl Infinoid: I forgot argh
21:50 bobke joined 21:52 amoc joined
rg who's running purl? the doc says ignore should have worked 21:52
Infinoid purl, owner
purl owner is Masque
bsdz_ where are purl's docs?
purl purl's docs are so opaque I couldn't tell
Infinoid that was actually enlightening. 21:53
rg masque?
purl it has been said that masque is figuratively speechless. or COOL WITH BURRITOS EVEN IF MINE DON'T SAY TOMMY HILBURRITO or Masquenfusion on AIM - USE THIS TO FIND HIM, IRC SUCKS or totally in love with warningsToBrowser, but forgetting to turn that off is something I fear. or awake. or Masquenfusion and Euqsam. or DJ Fresh Catnip or pleased with Masque's copy. or (see masque 2) or fond of used things that work. or a herring
dalek rrot: r38163 | bacek++ | branches/packfile_revamp (2 files):
Copy attributes from Packfile PMC to PackFile before packing.
Infinoid masque 2? 21:54
purl masque 2 is You sass that hoopy Masque? There's a frood who really knows where his towel is. or hellyeah sysop IIRC (or knows who is) or avocado-powered, baby
bacek That's it. Now packfile PMCs almost in line with PDD13.
rg purl's docs are also at www.infobot.org
purl okay, rg.
Infinoid mmm, avocados
bacek Few todo items left in sources.
bsdz_ thanks rg. i've always been curious about p*rl
wayland76 There are some in the specs :)
bacek wayland76: :) And I'm going to put more! 21:55
wayland76 For example, most of S17 is to do :)
bacek Infinoid: can I merge it to trunk? Or better wait till post release?
Infinoid Go ahead, we aren't deprecating anything here and your versions are significantly better than what's already in trunk 21:56
If you can, please consolidate the merge into one commit though, to reduce irc spam
Just prepare to be flogged if you broke anything :) 21:57
21:57 Whiteknight joined
bacek Infinoid: I can use "git merge --squash". Will it well behave with svn? 21:57
wayland76 bacek: If you break anything, just imitate a dead horse -- everyone will say "no point flogging a dead horse" :) 21:58
Infinoid wayland76: No, they'll just flog me instead as his mentor.
bacek We can imitate 2 dead horses!
21:58 contingencyplan joined
wayland76 And if we had some dogs eye, we could eat you :) 21:59
(that last line only works if you know rhyming slang)
rg it has been asked before (without response iirc), but is anyone going to do something about the nci failure on jit-capable platforms? 22:00
Infinoid which ticket, rg? cotto++ already fixed some of those, but I seem to recall there being more.
Whiteknight rg: I suggested a fix in the ticket, but no feedback 22:01
I'll be happy to fix it if a reasonable fix is found
Infinoid bacek: and about --squash, I have no idea, but I suppose so 22:02
git svn just commits anything that doesn't already have a git-svn-id, as long as it's at the top of the log, I think 22:03
bacek Infinoid: ah. Thanks. I have to fix couple of missed unit-tests before. 22:05
rg whiteknight: I have no idea about the details. the ticket reads like you know how to fix it, but not if there are downsides to your suggested fix. what's unreasonable about your suggested fix? 22:07
Infinoid I'm used to the stgit workflow, I rarely work with bare git. Is there a way to edit commit logs with git?
does it change the commit hash when you do?
rg also, there are now 3 failures reported in smolder, so either things got worse or the ticket doesn't cover all failures anymore? ;(
Whiteknight I don't think anything about it is particularly unreasonable, but it changes the semantics of the Pointer PMC
oh, I didnt realize there were more failures 22:08
JIT is such a damn mess
rg tell me about it ;)
Infinoid but at least it's an mmapped mess, now :)
rg according to TT#187 (deprecation notice), Pointer is unused. maybe remove it altogether? 22:09
bacek Infinoid: git commit --amend
rg ah no, i should read the whole ticket
wayland76 allison: it seems we didn't update the MANIFEST 22:12
allison wayland76: the ports directory isn't included in the MANIFEST 22:13
wayland76 No, but parrot.spec needs to be removed from it
allison wayland76: (though, I forgot that too, and ran the MANIFEST update script)
rg still, i think you can change the semantics, especially if it makes Pointer more useful. 22:14
allison wayland76: that's odd, because running the MANIFEST update didn't change it
wayland76 When I did a SVN up, I think it gave me an empty parrot.spec instead of none at all. Has it done that to you too? 22:15
allison wayland76: oh, I see, I just applied your patch, didn't explicitly svn delete parrot.spec
wayland76 (maybe I'm wrong about svn)
ah, of course :) 22:16
allison wayland76: deleted now
wayland76: no, you're right with svn, it's just that a diff file has no way to represent "svn delete" so it records it as deleting every line of the file 22:17
dalek rrot: r38164 | allison++ | trunk (2 files):
[cage] Actually delete parrot.spec.
allison wayland76: so the patch application faithfully deleted every line of the file, leaving an empty file checked into the repository
wayland76 I could see all that as soon as you said you didn't explicitly delete. 22:18
versioning systems aren't too bright. :)
I'll be sending at least one more spec patch out, to make it so that Rakudo can build on top of the Parrot RPM 22:19
I'll create a new issue, and we can argue about it :) 22:20
rg ah now i see. the thing cotto changed was the third failing test, not an attempt at a fix. 22:23
bsdz_ anyone know anyway i can debug getinterp::hll_map? seems to only work on classes local to the pir file its loaded in 22:25
rg whiteknight: i don't quite get how the pointer pmc figures into all this
is that what's used for the V arguments? 22:26
cotto rg, V arguments are void** 22:32
and Pointers are used with them. 22:33
mikehh when I run the .pir files on that test *the failing ones 66m 69 and 70) I get a Segmentation Fault
eg in ./parrot -t t/pmc/nci_66.pir - 28 invokecc P3 - Segmentation fault 22:36
in each case it calls invokecc with a P reg and then fails with a Segmentation fault 22:37
dalek rrot: r38165 | bacek++ | branches/packfile_revamp/t/pmc (8 files):
Move t/pmc/packfile_common.pir to t/pmc/testlib.
rrot: r38166 | bacek++ | branches/packfile_revamp/t/pmc/packfilefixuptable.t:
Reimplement t/pmc/packfilefixuptable.t in PIR
rrot: r38167 | bacek++ | branches/packfile_revamp/t/pmc/packfilesegment.t:
Reimplement t/pmc/packfilesegment.t in PIR
mikehh this is on Ubuntu Intrepid i386 at r38157 22:38
cotto mikehh, that's a side-effect of the PMC UnionVal removal. Nobody's managed to submit a patch so far, although it's a known problem.
I tried implementing Whiteknight's suggestion, but my patch only fixes one of the failing tests. 22:41
mikehh according to the ticket TT#551 its onlt failing on jit capable platforms (it passes on my Amd64 builds) 22:45
rg mikehh: it would probably pass if you compile with --jitcapable=0 22:46
cotto Bah. I'll just TODO the tests until they're all fixed. 22:47
mikehh maybe - does that mean in all our oyjer 10000+ tests this is the only ones that use this feature 22:48
s/oyjer/other/
cotto Right, but NCI is important for interfacing with external libraries too, e.g. mod_parrot. 22:49
mikehh I think we need to look at our test coverage a bot more carefully 22:51
cotto wonders how to test jit capability from Perl
mikehh * I STILL HATE THIS KEYBOARD *
cotto nm. found it 22:52
mikehh I was going to get another one yesterday, but somehow didn't get around to it
rg mikehh: i believe there is even a way to get coverage details. check for make cover. 22:53
mikehh I tried running some of the devel::cover tests - there is a make target but it failed spectacularly 22:54
cotto make cover works quite nicely, when it works.
mikehh I don't think anyone has run it for a while 22:55
Util Scanning src/ops/ops.num, I see 30 ops that take '_pc' params. I take that to mean "PMC Constant". 22:56
cotto I try to run it occasionally. it worked for me on r37626.
Util I also see that we support PMC constants in the bytecode. (packfile.h: PackFile_Constant and PFC_PMC).
Q: Do we have any PASM or PIR code that *produces* PMC constants?
cotto is awed by Util's typing speed. 22:57
Util is a terrible typist, and so pre-types (nearly) everything in an editor. 22:58
Infinoid there's a bunch of those generated in runtime/ from the contents of include/parrot/ 22:59
mikehh rg: BTW I noticed your comments/patch for TT#530 - makes sense - haven't worked with the 80x87 stack for a long while
Infinoid Util: grep for IPPROTO_TCP, for example 23:00
dalek rrot: r38168 | cotto++ | trunk/t/pmc/nci.t:
[PMC] TODO some failing nci tests that fail on jitcapable platforms
 23:01
mikehh Util: I think I need to try that - at least until I get a properly sized keyboard
Util Thanks, Infinoid! 23:02
mikehh: by pre-typing, I have saved much egg from my face. It does "up" the latency in this medium, though. 23:04
Infinoid and there's also some static ones we should get rid of (runtime/parrot/include/sockets.pasm)
23:10 Limbic_Region joined 23:17 tetragon joined
rg cotto: i think you're missing one todo 23:19
or is one block spanning two tests? 23:20
it is, never mind. 23:21
mikehh ok = I am giving make cover a try - let's see what happens - although I really should be able to pass make fulltest first 23:22
rg mikehh: update and apply my patch, then try fulltest ;)
mikehh I wll :-} 23:23
will
meed to get codetest running as well 23:24
meed
need - I did it again 23:25
I also think I need some sleep first
It's well after midnight for me 23:26
rg well my clock says the same, but i'm kinda living on us time. plus i don't want to miss asot 400 :) 23:27
mikehh well I am going to call it a night for now - try in the morning - cul8r and all 23:31
23:32 kid51 joined
rg kid51: would you help me close 4 rt tickets? 23:37
56012, 56024 and 56006 should be obsolete 23:38
43150 needs removal of one comment line from a file and then a close. 23:40
kid51 Do you have an RT account? If so, can you post to each of them why you are recommending they be closed? 23:41
rg i've already posted to 43150. i will also add a comment to the other 3. 23:42
kid51 Granted, I've looked at them in the past, but since I don't have access to Solaris or know much about it, we have to present a rationale for closing.
Thx
I'll probably allow for a few days for final comments, particularly from Andy Dougherty, who is knowledgeable about Solaris. 23:43
23:44 Gerd joined
kid51 I'll re-read each ticket after dinner. 23:45
Whiteknight cotto: At least the suggestion fixed one test! I'm not a complete bozo! 23:54
23:57 Theory joined
cotto We prefer incomplete bozos around here. 23:58

Powered by Cro and the Raku® Programming Language.

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

Thank you!