Parrot 0.9.0 | parrot.org/ | 566 RTs left.
Set by moderator on 31 January 2009.
00:09 AndyA joined 00:10 gerd joined 00:22 Theory_ joined 01:01 gravity joined
dalek tracwiki: v12 | allison++ | ChrootSetup 01:02
tracwiki: trac.parrot.org/parrot/wiki/Chroot...version=12
tracwiki: v13 | allison++ | ChrootSetup 01:03
tracwiki: trac.parrot.org/parrot/wiki/Chroot...version=13
01:13 gerd left 01:28 gravity joined 02:05 MariachiElf joined
GeJ Good morning everyone 02:07
Coke evening. 02:08
purl evening is probably when IRC is dead, TV is laden down with ads, and you're having my own dinner.
dalek rrot: r36345 | allison++ | trunk/docs/project/ubuntu_packaging_guide.pod:
[install] Adding an Ubuntu packaging guide for those steps beyond Debian packaging.
 02:16
02:22 particle joined 02:47 Fayland joined
s1n Coke: i see my name in the query builder :) 02:52
Coke ? 02:57
that's good, I suppose. 02:58
dalek rrot: r36346 | allison++ | trunk/docs/pdds/draft/pdd30_install.pod:
[pdd] Some cleanups/clarifications/decisions in the install PDD.
 03:02
03:20 jimmy joined
jimmy where's the query builder? 03:20
s1n jimmy: well, i was referring to Coke adding me to the rt parrotbugadmin group. i can now see my username listed on the search page 03:21
jimmy Will rt be never used? 03:23
03:25 Tene_ joined
s1n jimmy: i'm not sure i understand the question 03:26
jimmy parrot uses TT instead of RT. 03:27
s1n TT? trac? 03:29
jimmy yes.
Trac Ticket 03:30
s1n there are 456 open tickets on RT for parrot right now
Coke_afk: it worked, i must be in the right group, thanks for the help!
jimmy trac.parrot.org/parrot/wiki/TicketSprint
s1n jimmy: that also lists RT, and there are only 153 TT versus 526 RT 03:32
imho parrot should pick one, transfer the ones from the other system, and stick with it
jimmy yes, RT will be closed or moved to TT. 03:33
Am I right?
03:35 janus joined 03:37 kid51 joined
jimmy having lunch 03:38
kid51 In addition to the 'open' tickets in RT, there are approximately 70 others which are 'new' or 'stalled'. So the total which really need to be dealt with is, by my count, 524. 03:39
s1n kid51: is parrot really moving to TT? 03:40
kid51 We've asked committers to cease opening tickets in RT, but since parrotbug continued to point to RT, non-committers have continued to open tickets there.
03:41 TiMBuS joined
s1n why not just repoint parrotbug and export/import the tickets? 03:41
does RT even have the ability to export tickets for that matter?
kid51 I know a couple of folks who are following Parrot but are not committers (yet). I've emailed them directly advising them to open Trac accounts and file their next tickets there.
s1n is there an effort underway to transition all of the tickets over to trac? 03:42
kid51 Eventually parrotbug will probably be repointed or discontinued. Right now we're working on getting its substitute, email2trac, working. There's a TT about it; it kinda works.
03:42 ask_ joined
s1n but what about all of the existing tickets? 03:43
kid51 I don't think there's an organized effort yet to move tickets over. That has happened in individual cases. See my post in #parrotsketch today: The RT tickets really have to be evaluated by knowledgeable people. That's more important than worrying about which system they're in. But those "knowledgeable people" are, of course, are core developers, who have only so many tuits available. 03:44
Feel free to go through the new/open/stalled RT tickets and comment on any where you feel competent to do so. 03:45
s1n i'm not competent for any of them, i can just indicate whether they are still reproducable
kid51 There are *hundreds* of RTs where I don't know enough to say anything. 03:46
s1n if the parrot devs want the rt tickets that can be reproduced to be transferred, i can help with that
kid51 Well, that's one way you can help out. For example, over this past weekend I went through the list and found a few where the RT was old and the bug reported was no longer reproducible. I posted my passing result and asked for confirmation. Hearing nothing either way after 3 days, I was able to close a ticket. 03:47
s1n i'm currently doing that for rakudo, i could do that for parrot if so desired 03:48
kid51 I don't think they've asked for that. In any event, as mentioned above, we have bugs to work out in Trac. And we already have a growing backlog of open tickets in Trac.
s1n i just have basically no knowledge of parrot
well, if the parrot devs haven't asked for that and don't want it due to the backlog in trac, that's fine too 03:49
kid51 Well, when it comes to what is really the core of Parrot -- compilers, C, VMs -- I'm not far ahead of you.
s1n a large backlog means there may be too many chiefs and no indians
kid51 I think it will eventually happen, but, in addition to shooting for 1.0 on March 17, we have a lot of other infrastructural work going on. right now.
s1n it might benefit core devs to start spreading the knowledge so they can get more help 03:50
kid51 I hear what you're saying, but that's not the real problem.
s1n but i don't, i don't do any parrot development, so it's not like i'd be focusing on the wrong thing
what does a growing backlog imply to you? 03:51
kid51 The problem is that Parrot development has moved so rapidly in the last 2 years that it's difficult for people who don't have those core skills to keep up. I certainly can't, even though I've been a committer for 2+ years.
s1n well, i'm offering assistance with the tickets, take it or leave it 03:52
kid51 Most of what I read in this channel I don't understand. But I was originally recruited to write tests for Perl 5 code within the distro. That's mostly done now, and we really don't want to be adding more Perl 5 code to the distro if we can help it. 03:53
Because it all has to be torn out eventually.
At this point, it's more important to increase the number of people who are *reading* the tickets (regardless of which system they are in) and *thinking* about them. You can probably help there. 03:55
Do you know how to use gdb? 03:56
kid51 knows only the bare essentials. 03:57
Do you know how to do Smolder tests? What OSes/platforms are at your disposal? 03:58
We need at least daily Smolder reports on Darwin and Win32, but we're not getting them. That's an area where someone could help. 04:00
We need someone knowledgeable in the differences between OpenBSD and Free/NetBSD, so that we can resolve the tests which fail only on Open and not on the others.
We need someone intimate with Darwin to help resolve OS X-specific RTs and TTs. 04:01
04:07 Fayland_ joined
s1n kid51: i know enough gdb to be dangerous 04:17
i'm using gentoo amd64, i have arch and slackware VMs immediately available
i don't know anything about osx or windows and not much about the bsds 04:18
dalek rrot: r36347 | jkeenan++ | branches/closure_for_closure:
Very rough first draft of work to eliminate deprecated closure PMC. 'make' not yet completing.
 04:21
kid51 okay. If you follow this channel, you will see that from time to time, when tests fail, developers ask reporters for backtraces. That's the upper limit of my knowledge of gdb -- learnt only 3 weeks ago. So that's some place you can help.
kid51 must sleep
purl $kid51->sleep(8 * 3600);
04:39 szabgab__ joined, szabgab joined
Util must sleep 04:53
purl $Util->sleep(8 * 3600);
cotto Util++ 05:43
06:39 alinbsp joined 06:40 rurban_ joined 06:53 uniejo_ joined 07:13 uniejo joined 07:18 jrockway joined 08:01 HG` joined
lu_zero must nap 08:05
purl should consider nap, not just sleep =P
purl lu_zero: sorry...
08:09 iblechbot joined 08:43 alvar joined 09:19 ask_ joined 10:04 kj joined 10:21 braceta joined 10:24 Zaba joined 10:57 darren34 joined
darren34 hello everyone 10:57
im trying to display the contents of a text document, using perl. but it is not working :( 10:58
the output used to be blank but now im getting: "The specified CGI application misbehaved by not returning a complete set of HTTP headers."
i have checked my format against my other perl files that run correctly 10:59
and all looks ok
anyone got any suggestions please?
bacek darren34: one suggestion: join #perl at FreeNode. 11:00
darren34 how do i get there?
szbalint O_o 11:10
how did he get HERE?
:)
szabgab is map already implemented in Rakudo ? I can't seem to get it working. 11:24
hmm, now it worked 11:35
bacek szabgab: it is implemented 11:36
rakudo: map { say $^a }, 42..43
polyglotbot OUTPUT[42ā¤43ā¤]
11:53 masak joined
janus Infinoid: no luck with your patch and a bit hard to test as i have to bounce diffs to someone else with enough ram to compile, the 256mb in my u10 are unfortunately not enough 12:27
12:31 rg1 joined 12:32 kj joined 12:34 iblechbot joined
dalek rrot: r36348 | jkeenan++ | branches/closure_for_closure:
Branch was unsuccessful; removing.
 12:44
rrot: r36349 | jkeenan++ | tags/closure_for_closure-36342:
Remove tag corresponding to deleted branch.
 12:45
sjn wonders if there's a way to access the parrot SVN repository without https 12:52
13:08 bacek joined 13:13 Whiteknight joined
Util_away sjn: plain http worked before the move to svn.parrot.org, but does not post-move. 13:14
Whiteknight yeah, it's https only now 13:21
sjn yay 13:24
sjn upgrades his svn install 13:25
Whiteknight if you really want http: instead of https: raise the issue on the list 13:28
they didn't add it because they didn't think enough people were using it
Coke_afk (move tickets) that's a waste of energy. Let's just /resolve/ the old tickets. 13:59
msg kid51 - there are daily smolder reports on darwin/x86, because I'm submitting them. 14:01
purl Message for kid51 stored.
Coke s1n: I'll have a clarification on the dual-ticketing system in a writeup due this weekend. 14:08
but, basically, we're in the /middle/ of migrating. Try to use tt for new stuff. Try to close out the old stuff. 14:09
dalek rrot: r36351 | fperrad++ | trunk/languages/lua/t:
[Lua] fix tests (portable Inf & NaN)
 14:11
rrot: r36352 | whiteknight++ | trunk/docs/book/ch12_opcodes.pod:
[Book] Add more information about opcodes, dynops, and runcores.
 14:15
14:27 gryphon joined
dalek rrot: r36353 | whiteknight++ | trunk/docs/book/ch12_opcodes.pod:
[Book] a little bit more about opcode parameters
 14:29
14:42 rurban_ joined
Infinoid janus: Ok, thanks for trying. 14:50
14:50 bkuhn joined 15:52 Tene joined
particle so, in my .git/config, i have url = git://github.com/rakudo/rakudo.git 16:00
this allows me to git pull
but i can't git push, it says it wants git@...
16:05 NotFound joined
NotFound hi 16:05
purl hey, NotFound.
Infinoid good morning 16:12
dalek rrot: r36354 | NotFound++ | trunk:
update headerizing
 16:13
pmichaud particle: github.com/rakudo/rakudo/tarball/master 16:19
oops, wrong
particle:
jonathan morning pmichaud
pmichaud git@github.com:rakudo/rakudo.git
hello, jonathan.
jonathan How's things? 16:20
16:20 alinbsp joined
jonathan Snowed under? 16:20
pmichaud figuratively, yes. :-)
I'm working on an updated Configure.pl now. About to push something to github.
jonathan Great.
Haven't tried my first github push yet!
pmichaud I suspect it won't work under windows, though -- don't have an easy way to test.
well.... it'll work the same way it always has
jonathan Just push it, I'll fiddle if needed.
pmichaud yeah, there's quite a bit of stuff we have to resolve in general. 16:21
should go quickly once the basic pieces are in place, though.
jonathan fwiw, the Configure.pl that is there now, if I check out to languages/rakudo/, works on Win32.
pmichaud right. I'm keeping that portion. 16:22
jonathan OK
pmichaud essentially, the updated Configure.pl asks "can I reach ../../tools/dev/reconfigure.pl ? If so, use that."
jonathan Simple. :-)
pmichaud if not, it starts looking for other parrot config information (which is the part that probably won't work under Win)
jonathan OK.
What does it fall back to trying? 16:23
Looking for an installed one?
pmichaud it starts looking for the "parrot_config" executable.
and then uses that to obtain the rest of the configuration information.
jonathan Is that something expected to be in the path? 16:24
pmichaud it first looks in parrot/parrot_config, then it does a generic "parrot_config", and after that it carps.
one can specify --parrot-config=path/to/parrot_config to request a specific one.
jonathan One bit of feedback - it would be nice if Configure.pl spat out a message telling you it was successful, when it was.
pmichaud it does that now.
(in the version I'm about to check in)
jonathan OK, great. 16:25
pmichaud it gives a lot of information about what it's doing and what the results are.
jonathan I was showing someone Rakudo at the weekend and they can Configure.pl and was like..."so did it do anything?"
Great! :-)
pmichaud I'm still trying to figure out: given that someone gets a copy of Rakudo, how do we then get them the appropriate copy of Parrot?
with this Configure.pl they can checkout and build to a parrot/ subdir, but I'm wondering if we want to be even a bit more automated than that. 16:26
jonathan Are we going to track a particular copy of Parrot that we work against?
pmichaud I think we'll have to track Parrot, yes.
jonathan like a PARROT_VERSION file?
pmichaud that, or a line in Configure.pl
jonathan Or maybe PARROT_REVISION
pmichaud yeah, PARROT_REVISION makes sense. 16:27
jonathan I guess separate file is nice in that you can cat it without having to go dig through Configure.pl
To quickly know which Parrot revision we're working against.
pmichaud I'm wondering if Configure.pl should try to automatically grab/build the correct version of Parrot, or if we should require folks to do that explicitly.
16:28 particle joined
jonathan It should maybe give the option of doing that. 16:29
pmichaud particle: your thoughts on irclog.perlgeek.de/parrot/2009-02-04#i_885604 would be helpful.
yeah, I thought about making it an option also.
rurban pmichaud: RT #40806 HLL setline: I believe parrot is correct. If you really want to display the right lineno for $a = 9876;\\nnsf($a + 1); then your compiler should insert all the setline ops. But I cannot blame parrot for correctly calculating the lines
dalek rrot: r36355 | fperrad++ | trunk/languages/lua/src/pmc:
[Lua] some consting
 16:30
jonathan rurban: setline is not for HLL debug info
NotFound rurban: this is not obsoleted by code annotations?
pmichaud #40806 was resolved months ago, I thought.
rurban true, so rt.perl.org/rt3/Ticket/Display.html?id=43269 shoudl be rejected.
jonathan I thought I closed the related RT tickets. 16:31
But may have missed some too.
rurban .annotate is already there
jonathan yes, I did resolve it.
pmichaud 43269 was resolved in early Jan.
jonathan That ticket is resolved. No action needed.
rurban Ok, so I'll rewrite the TODO t/op/debuginfo.t tests
jonathan Aha, OK. 16:32
rurban: But note that there are already t/op/annotate.t
rurban t/op/debuginfo.t 7 and 8 still fail, but should pass
pmichaud since setfile/setline are now obsolete for hll stuff, some of the tests in debuginfo.t can probably be removed altogether.
Infinoid rurban: in the context of register window flushing for GC (TT #271), should sparc v10 be handled like v9, or like v8? 16:33
rurban Infinoid: like v10 I suppose, because v8 worked for us, and v10 came up with the errors.
sorry, Imixed it up.
I have no sparc to test but someone posted the manual yestreday 16:34
NotFound PacoLinux may have some sparc available
particle pmichaud: my first priority is getting a commit via git working 16:36
Whiteknight Infinoid: I've been looking for details all day. I can't find a v9->v10 changelog of any type
rurban me neither.
dalek rrot: r36356 | rurban++ | trunk/t/op/debuginfo.t:
- remove TODO's and fix tests.
Infinoid Ok. I've got a patch that simplifies things greatly according to Andy D's suggestions. But I wasn't able to get a confirmation on whether it works, janus said no luck
I don't have anything to test with either. 16:37
(my failed patch: nopaste.snit.ch/15493)
rurban I really should cleanup my harddisc to add a big-endian 64-bit machine, like a sparcv10 16:38
Whiteknight Infinoid: I've been looking at the Bohm-Demers-Weiser collector source code too. They have these routines for most of our supported platforms I think
pmichaud particle: you need to use git@github.com:rakudo/rakudo.gif
particle: you need to use git@github.com:rakudo/rakudo.git
rurban libffi should have it also
pmichaud (for git push)
rurban nope, libffi only has v8 and v9 16:39
Whiteknight See this: bdwgc.cvs.sourceforge.net/viewvc/bd...iew=markup
shorten Whiteknight's url is at xrl.us/beewkx
Infinoid Actually... come to think of it, we have some random sparcstations laying around here at work. Wonder if there's enough hardware to build a complete system.
Whiteknight And this: bdwgc.cvs.sourceforge.net/viewvc/bd...iew=markup
shorten Whiteknight's url is at xrl.us/beewkz
Infinoid (won't be sparc64, but it's better than nothing)
rurban beware we have two issues 32-bit vs 64-bit and v9 vs v10 16:40
particle Enter passphrase for key '/c/Users/particle/.ssh/id_rsa':
que es eso?
pmichaud did you set up a ssh key on github? 16:41
it's asking for the passphrase associated with that key.
particle ok, i must have, no idea. hrmm
pmichaud you could set up a new key 16:42
dalek rrot: r36357 | NotFound++ | trunk:
reimplement stdhandle method, TT #264
particle i might have to. 16:43
jonathan pmichaud: I didn't add a key at github. Is it a requirement though?
rurban sources.redhat.com/cgi-bin/cvsweb.c...oot=libffi has the v8 vs v9 and 32 vs 64 bit issues seperated
shorten rurban's url is at xrl.us/beewmw
pmichaud jonathan: for doing a push, it may be.
jonathan Or can I just authenticate with username/password like I did with svn?
pmichaud jonathan: I don't know if username/password works by itself. I set up an svn key for me.
moritz s/svn/ssh/ ? 16:44
pmichaud yes, ssh, thanks.
github.com/guides/how-to-not-have-t...every-push
shorten pmichaud's url is at xrl.us/beewmy
rurban I found no jit support for sparcv10 on the net at all 16:46
Infinoid I love how wikipedia describes the changes in v9, but doesn't mention v10 at all. 16:47
dalek kudo: db33076 | particle++ | .gitignore:
add win32 extensions for generated files to .gitignore
 16:48
NotFound Sparc Me? X-)
particle ok, yes, key is required at github, i generated a new one, et viola
shorten dalek's url is at xrl.us/beewnk
dalek kudo: c3705e4 | particle++ | :
Merge branch 'master' of git.rakudoperl.org/rakudo/rakudo

Conflicts:
  \t.gitignore
shorten dalek's url is at xrl.us/beewnn
particle jonathan: github.com/guides/providing-your-ssh-key#msysGit
conflicts?
purl conflicts are conflicts
rurban I've only found that v10 removed some tagged instructions from v9: www.realworldtech.com/forums/index....p;roomid=2
shorten rurban's url is at xrl.us/beewn7
Infinoid purl: wow, enlightening.
purl Infinoid: huh?
Whiteknight Infinoid: I think we can definitely adapt what the Boehm collector is doing here. They are using the same code for v9 as they are using for v10. They're also using the same ASM for 32- and 64-bit variants 16:49
16:49 Theory joined
Whiteknight If we can tweek the configuration step to add a new .s file to the build for sparc systems, I think I can fix the problem (with testing, of course) 16:50
Infinoid particle: If your commit is ok, it might indicate something I need to fix in dalek's parser.
pmichaud github says 'conflict' in its message also.
particle right.
pmichaud anyway, the file looks okay. 16:51
particle runs test_summary.pl with optimized parrot
dalek kudo: 7d7f1af | (Patrick R. Michaud)++ | Configure.pl:
Update Configure.pl with some help messages and options.
 16:55
shorten dalek's url is at xrl.us/beewo8
dalek kudo: 8449537 | (Patrick R. Michaud)++ | (2 files):
More Configure.pl updates.
shorten dalek's url is at xrl.us/beewpe
dalek kudo: 183814a | (Patrick R. Michaud)++ | .gitignore:
Merge branch 'master' of git@github.com:rakudo/rakudo
shorten dalek's url is at xrl.us/beewpg
PerlJam good morning #parrot 16:57
Infinoid By the way, those "merge" commits are caused by pull, and I think they can usually be eliminated by doing a rebase before the push
szbalint good evening
Infinoid hai PerlJam
pmichaud okay, if folks want to test Configure.pl a bit, that would be helpful. 16:58
particle i need a manual.
pmichaud right now there are two forms it (should) understand.
1. Check out rakudo into languages/rakudo of a parrot build tree 16:59
2. build Parrot in the parrot/ subdir of the Rakudo tree
I haven't tested the second form fully yet.
PerlJam pm: automagically or based on options.
?
pmichaud pj: 1 is automagic 17:00
pj: 2 automagically detects if there's a built copy of parrot in the parrot/ subdir, but it doesn't automagically build on for you.
particle parrot\\rakudo\\languages>git clone git@github.com:rakudo/rakudo.git rakudo-configure
hrmm i'll nopaste
pmichaud *one 17:01
nopaste "particle" at 76.121.106.245 pasted "doesn't work from languages/rakudo-configure" (43 lines) at nopaste.snit.ch/15495
pmichaud particle: it has to be languages/rakudo
languages/rakudo-configure won't work, because parrot's reconfigure isn't that smart. 17:02
particle ok, that's a problem for testing, then.
ok, i see.
i'll try the other method now
pmichaud maybe we can make the Configure.pl script smart enough to use the basename of the current directory.
and pass that to parrot's reconfigure.pl step.
particle 'parrot' is not recognized as an internal or external command, 17:03
need to look for parrot.exe, checking 17:04
pmichaud parrot.exe?
should be parrot_config, I think.
Infinoid Ok. I built parrot in parrot/, and then Configured rakudo successfully.
But when I try to build:
make: *** No rule to make target `../../parrot', needed by `perl6.pbc'. Stop.
particle parrot_config.exe, then
pmichaud sounds like a root.in error 17:05
I have to run for a while -- feel free to adjust the makefiles and Configure.pl for a while
Infinoid That's on linux/x86-64
Ok, I'll see if I can tweak it to work
particle my @parrot_config_exe = map {($_, $_ . '.exe') }
qw(parrot/parrot_config parrot_config);
PerlJam When I run Configure.pl in a fresh parrot checkout (with rakudo in languages/rakudo), the message it gives is a tad conflicted: "You cannot use --step until you have completed the full configure process" Okay, that looks like an error, but then: You can now use 'make' to build Rakudo Perl. I can? Somehow I doubt it :) 17:06
pmichaud PerlJam: okay, we need to check the status or make sure a Makefile got created.
particle patches welcome!
purl patches welcome is ponies welcome or Set Objectives, Achieve Results! or swahili for "Put up or shut up."
pmichaud Rakudo's Configure.pl depends on having a completely built parrot.
particle most of my tuits are job-interview-shaped today
PerlJam particle: good luck! 17:07
purl You'll need it.
particle thanks, purl.
purl no worries particle
particle thanks, pj!
jonathan Anyone else getting Rakudo failures in some of the meta-op tests? 17:09
moritz it doesn't even build for me
jonathan e.g. t\\spec\\S03-operators\\reduce-metaop
moritz ../../parrot: symbol lookup error: /home/moritz/src/parrot/runtime/parrot/dynext/perl6_group.so: undefined symbol: string_append
jonathan Oh, that'll be a result of allison's fiddling. :-)
NotFound moritz: Wrong parrot rev?
jonathan moritz: Try building against an older Parrot, not head... 17:10
moritz NotFound: current parrot rev, don't know if that's wrong
jonathan moritz: I'll get latest Parrot head and try and fix it though.
NotFound moritz: rakudo is wrong, not updated to string functions recent rename
jonathan (not tracking Parrot head)++ ;-) 17:11
moritz don't we usually have a deprecation cycle when renaming API functions?
or is that not an API change?
PerlJam My freshly checked out parrot and freshly cloned rakudo both built fine. 17:12
particle jonathan: i'm looking at them now
jonathan: i suspect the tests were changed to match the new spec
jonathan "near "X, 1,2 X, "
That's not the syntax Rakudo expects...
OK.
So X,X is now just X,
?
moritz yes
jonathan hmm
And reduce?
purl reduce is reduce or not oxidize or donate electrons to
NotFound moritz: I think allison decided that the string nightmare is an exception. 17:13
particle yes, it's X, or X[,] (for disambiguation)
jonathan OK, that explains the cross meta-op failures. Still curious why we fail the reduce meta-op tests. 17:14
But guess that means my currently local diffs didn't cause this breakage.
dalek kudo: d38a4ac | (Mark Glines)++ | config/makefiles/root.in:
[root.in] Fix the path to parrot(.exe), for the "parrot is in parrot/" case.
 17:15
particle t\\spec>svn diff -c 25140
moritz it changed on Feb 01
shorten dalek's url is at xrl.us/beewqy
NotFound Someone has built r36357 ?
moritz it seems that a few tests were deleted without adjusting plan
jonathan ah, OK
jonathan doesn't worry about it much for now 17:16
I'll patch the meta-ops for the new syntax though.
erm
the cross-ops I mean
PerlJam NotFound: I just did.
NotFound Good, looks like this time I don't broke anything :) 17:17
jonathan I want to check in just the changes in the src directory 17:18
git push src
Doesn't seem to do it.
fatal: 'src': unable to chdir or not a git archive 17:19
fatal: The remote end hung up unexpectedly
Infinoid jonathan: "git add" the files you want to include in your commit. (This adds them to the index.) 17:20
jonathan Infinoid: They're not new files, I just changed them.
Infinoid jonathan: "git add" != "svn add"
It adds them to the index, not to the repository
jonathan OK, but surely there's just a command to commit the changes? 17:21
Infinoid The index is what determines which files get committed
jonathan Like I'd do a svn ci -m "..." src
?
particle git checkout != svn checkout # also confusing
git commit -m "..." ; git push origin master
jonathan particle: and can I commit with just a given directory? 17:22
particle sure, with commit dir/ -m ...
Infinoid I'm probably not much help here, actually. I use a wrapper that hides all that index stuff so I never had to learn it.
jonathan git. SO GOOD you have to write your own wrappers to not have to type a load more than under svn... 17:23
Infinoid I didn't write my own, I use stgit :)
It's a bit more svnish
PerlJam git tracks content changes rather than file changes 17:24
particle oh crap, i missed the oscon cfp 17:25
pmichaud oh darn, so did I
I can't believe it's feb already 17:26
dalek kudo: a5e405e | (Mark Glines)++ | (2 files):
Fix a couple of other things related to building with a parrot in parrot/.
shorten dalek's url is at xrl.us/beewsh
Infinoid pmichaud: Now I'm getting some build failures due to hardcoded "../.." include paths in C sources.
pmichaud Infinoid: which C sources?
Infinoid perl6_ops_switch.c:30:44: error: ../../../../src/pmc/pmc_object.h: No such file or directory
jonathan Hmmm...anyone know what this means:
To git@github.com:rakudo/rakudo.git ! [rejected] master -> master (non-fast forward)
Infinoid perl6_ops.c:29:44: error: ../../../../src/pmc/pmc_object.h: No such file or directory
jonathan error: failed to push some refs to 'git@github.com:rakudo/rakudo.git'
pmichaud jonathan: you might need a 'git pull' first 17:28
Infinoid If your tree has local changes, I'd suggest "git fetch; git rebase origin" instead
pmichaud I think it means your local repo is out of date wrt the master
or what infinoid said. :-)
jonathan pmichaud: git pull did it.
You know, git may be more sophisticated than SVN. But at least for simple operations, I knew what it was going to do! 17:29
;-)
dalek kudo: 83a3599 | unknown++ | (4 files):
Merge branch 'master' of git@github.com:rakudo/rakudo
pmichaud yeah, it does involve re-learning a few of the basics.
shorten dalek's url is at xrl.us/beewsq
Infinoid rebase prevents useless "Merge branch 'master'" pseudocommits
purl, karma unknown?
purl i don't know, infinoid
Infinoid purl, karma unknown
purl unknown has karma of -23
jonathan Hmm. Was that my commit?
Infinoid It looks empty 17:30
pmichaud it looks empty to me also.
jonathan Why would an unknown user be able to commit? :-S
erm, push 17:31
PerlJam jonathan: you need to setup your local git config to identify yourself to git.
nopaste "jonathan" at 85.216.157.73 pasted "here's what happened when I pushed" (7 lines) at nopaste.snit.ch/15496
PerlJam jonathan: github.com/guides/tell-git-your-use...il-address
shorten PerlJam's url is at xrl.us/beewsw
pmichaud afk, lunch 17:32
jonathan PerlJam: Thanks, done that. 17:33
Now what I'm curious about is - what happened to the stuff I thought I'd pushed?
Ah 17:34
github.com/rakudo/rakudo/commit/c8d...1a0b0c7f3c
shorten jonathan's url is at xrl.us/beews6
jonathan There it is.
So why didn't the other link point to that?
Infinoid Your other link was because of "git pull", and didn't have any content
jonathan :-|
Infinoid dalek dropped the real commit because the commit date was less than (numerically) my commit
That's something I'm gonna need to fix 17:35
dalek kudo: a888fb3 | jnthn++ | t/spectest.data:
Add S12-role/crony.t to spectest.data.
 17:36
shorten dalek's url is at xrl.us/beewta
jonathan Well, that one worked out a bit better... 17:37
Though I can't believe how much more I have to type. :-(
git-- # not making the easy things easy
rurban solaris currently cannot link shared. Problem is cc => CC, with cc -shared it works fine. => new ticket 17:38
dalek kudo: 78650e7 | (Mark Glines)++ | src/ops/perl6.ops:
Since -I<parrot>/src/pmc is in the compiler command line, we don't need all that

building successfully when parrot is in a subdir.		 17:39
shorten dalek's url is at xrl.us/beewte
PerlJam Infinoid: dalek outputs the full commit message? 17:40
Infinoid PerlJam: Yeah. Shouldn't it?
It's consistent with its other log output 17:41
PerlJam There's a convention in git-space of first line summary, blank line, full message. I wonder if we should adopt that convention and just have dalek output the first line.
Infinoid Yeah, I've seen that too. If rakudo adopts that convention, I'll adapt dalek too. Deal? 17:42
I tried "make perl6" to make a fakecutable, but pbc_to_exe seems to have hung
PerlJam sure :) 17:43
17:44 barney joined
Infinoid Oh, pbc_to_exe didn't hang, it's just *very* slow. 17:44
I also get a "double free or corruption" bactrace from glibc every time parrot hits a parse error 17:47
17:47 hercynium joined
particle error: Unable to find ab74be6e7b81783ff462a7ed994aa82b7e824340 under gith 17:48
ub.com/rakudo/rakudo.git
Cannot obtain needed tree ab74be6e7b81783ff462a7ed994aa82b7e824340
Infinoid I don't have that commit, either.
Where did that error come from? 17:49
particle i tried both git pull and get fetch 17:50
after yours and jonathan's commits
17:50 rhr joined
Infinoid Does "git show ab74be6" turn up anything? 17:50
particle yes, it screws up my terminal 17:51
Infinoid urk
that wasn't quite what I was hoping for, sorry about that. 17:52
particle WARNING: terminal is not fully functional
fatal: ambiguous argument 'ab74be6': unknown revision or path not in the working tree.
Infinoid Ok. So I'm still not sure why it's looking for that revision
particle how do i get out of this thing?
i keep getting (END)
Infinoid Do you have any local changes?
particle and beeps when i press keys
Infinoid oh, try q maybe 17:53
on linux, it runs "less" automatically, so you're probably in some kind of pager
particle ah, okay
i wonder if there's a way to configure that
i have less installed via msysgit 17:54
q<Enter> several times seems to have worked, thanks.
Infinoid [core]'s "pager" attribute in .git/config
It'll use that whenever you do a "diff", "show" or "log" 17:55
or "annotate", or any number of other things
particle heck, i could use vim then 17:57
rurban hell, linkflags is also not honored with libparrot.so 17:59
particle Infinoid++ # my own less.exe works much better, thanks for the tip! 18:03
Infinoid Oh, great. I hope you're keeping notes on all of this :)
So your current tree doesn't contain ref ab74be6, and neither does github. The question is why it wants it 18:04
Are there any merge commits in "git log" that mention it?
Are there any outstanding commits in your tree?
particle not in either of my trees 18:08
Infinoid I don't suppose you can tell me what you *did* do to these trees, so I can try to reproduce it, can you? :)
what's the global gitconfig called on win32? "~/.gitconfig" looks a bit unixish 18:10
particle ah, ok, it's just one tree that's wedged
global? iunno
i have .git/config files in my trees 18:11
ah, i do have /c/users/particle/.gitconfig
fab, better place for my info :)
PerlJam you don't need to know what it's called, just run "git config"
Infinoid particle: Great, I was just going to suggest that. :)
particle git config # usage page 18:14
PerlJam git help config # too 18:15
particle git isn't recognizing my global config :( 18:20
pmichaud (commit convention) I'm fine with adopting the convention. 18:26
Until now we've been using parrot's conventions, but we can switch.
jonathan pmichaud: What do you mean by commit convention? 18:29
I didn't prefix my last couple with [rakudo] because, well, it's a tad obvious now. ;-)
pmichaud 17:41 <PerlJam> There's a convention in git-space of first line summary, blank line, full message. I wonder if we should adopt that convention and just have dalek output the first line. 18:30
17:42 <Infinoid> Yeah, I've seen that too. If rakudo adopts that convention, I'll adapt dalek too. Deal?
Infinoid The first line is sort of a subject, or title 18:31
18:31 rurban joined
Infinoid I don't feel very strongly about it. But it does make log summaries look nice 18:31
particle it makes commit -m harder 18:33
jonathan Indeed. 18:34
particle however, if you can set editor = gvim and that works with git commit, i'm all good
jonathan fwiw, I've never written a commit message with a newline in.
And I don't especially feel inclined to start doing so now, when nobody has ever complained about my commit messages in the past.
NotFound Did we have conventions for commit messages? 18:35
jonathan :-D
Infinoid particle: I don't think it makes commit -m any harder; it just means you *only* have a title line 18:36
It's not a very rigidly enforced standard, in any case.
I think it isn't worth worrying about, at the moment. 18:37
pmichaud for Parrot, I've followed the convention: "[subsystem]: summary\\nlonger description"
NotFound I follow that, except that sometimes elid the subsystem part and never put the longer description ;) 18:38
pmichaud yes, I sometimes omit the longer description if there's nothing longer to describe.
Don't want to fall into the "i++; /* Add 1 to i */ " trap. 18:39
NotFound "i++; /* post-increment i */ " ;)
Infinoid I once saw a /* bigger is always better! */ 18:40
True story.
NotFound "i+=1; /* Add 1 to i */ "
Correcteness matter ;)
Whiteknight I read a hilarious story about when Netscape went open source and all the foul comments that it's engineers had in the code for it 18:41
like /*
like /* Microsoft sucks ass! This code is complete crap!!! */
confound url!
Infinoid Yeah, coding standards often don't exist for proprietary code. One of the reasons why I'm sick of the .com world 18:42
GeJ Good morning everyone 18:44
Infinoid morning GeJ
Whiteknight good morning GeJ
GeJ 5:45 AM. Waaaay too much morning-y is you ask me. 18:46
Infinoid Almost lunch time here.
rurban 8pm here 18:53
I've got shared libs working now on solaris, rpath was not set. 18:54
jonathan afk for a bit 18:55
Infinoid rurban++ 18:59
rurban I only got static libs. it worked fine, but then my harddisc space filled up. So I NEED a libparrot.so :) 19:00
Don't we have some common cc and ld guidelines somewhere for the most common compilers? 19:01
gcc is obvious, but llvm-gcc, icc, borland, msvc cl, sunpro CC and so on... 19:02
dalek kudo: 84b76c7 | pmichaud++ | PARROT_REVISION:
Add PARROT_REVISION to track wanted revision of Parrot for building Rakudo.
 19:03
shorten dalek's url is at xrl.us/beew6m
rurban where what a user search for it? perl Configure.pl --help, docs/compiler_faq.pod, docs/configuration.pod 19:04
PLATFORMS also maybe 19:05
NotFound wiki? 19:08
purl wiki is dev.catalyst.perl.org/wiki/
barney also docs/porting_intro.pod 19:16
rurban I thought adding the tested compiler flags to PLATFORMS might help 19:18
icc and llvm builds are also not mentioned in PLATFORMS 19:19
NotFound rurban: put a reference to the relevant wiki page
rurban borland also not
NotFound I think that the last time someone asked the answer was that no one builds with Borland 19:21
19:21 chromatic joined
Whiteknight I tried lazily to build with Borland and it borked 19:22
dalek kudo: b0713e5 | pmichaud++ | MAINTAINER:
Remove unneeded MAINTAINER file.
 19:24
shorten dalek's url is at xrl.us/beew9r
dalek tracwiki: v5 | rurban++ | ListOfPlatforms 19:28
tracwiki: trac.parrot.org/parrot/wiki/ListOf...?version=5 19:29
shorten dalek's url is at xrl.us/beew99
19:31 geof joined
Infinoid emails ask_ about shorten and dalek's URLs 19:33
cotto Infinoid++ 19:37
TimToady phone 20:00
20:02 register joined 20:19 mberends joined
dalek tracwiki: v6 | fperrad++ | ListOfPlatforms 20:29
tracwiki: trac.parrot.org/parrot/wiki/ListOf...?version=6 20:30
shorten dalek's url is at xrl.us/beexgb
dalek rrot: r36358 | rurban++ | trunk/config/inter/libparrot.pm:
[configure] Try some more libparrot.so candidates.
 20:41
rrot: r36359 | whiteknight++ | trunk/docs/book/ch04_pir_subroutines.pod:
[Book] Add info about Sub PMC, more stuff about LexInfo PMC, and info about Namespace PMC. Few small cleanups.
 20:44
rrot: r36360 | rurban++ | trunk/config/inter/progs.pm:
[configure] Improve wording for 2 Configure.pl --ask questions:

  - "And your cc for building shared libraries?"
   => "And your cc flags for building shared libraries?"		 20:46
Coke_afk it's not necessary to duplicate your change in the commit message. 20:47
Whiteknight but it's not forbidden! 20:50
rurban I thought it might be heated issue. 20:51
a heated issue
closing tickets: I think I can explain #37993: [BUG] optimized parrot behaves differently than non-optimized (win32-msvc) 20:52
Whiteknight really? I'd be interested to hear the explanation
rurban it's just the msvcrt runtime dll having these issues, mingw ditto
They don't support -0 as on posix, neither Inf/Nan 20:53
I tend to ignore that issue and document it as platform quirks 20:54
but I'll check first how that behaves on perl5 20:55
20:55 szabgab joined
Coke heated, no, just redundant. 20:57
Coke curses whiteknight for commit an it's/its bug so he felt compelled to troll through the book.
dalek rrot: r36361 | coke++ | trunk/docs/book:
fix the state of "it's" vs. "its", and one spelling mistake.
Coke *committing
Whiteknight Coke: Sorry! 20:58
my stupid fingers don't do what they should every time
chromatic Mr. Bender, your son has a case of what we professionals call "stupid fingers". 20:59
Whiteknight "We're sorry, the fingers you have used to dial are too fat. To order a special dialing wand, please mash the keypad with your palm now" 21:03
...and on that note, I'm heading home.
rurban How do I test signals as mentioned in PLATFORMS? perl t/pmc/signal.t => Skip no events yet 21:11
Sorry, I see it 21:13
dalek kudo: 1bbe7f4 | pmichaud++ | (2 files):
Move PARROT_REVISION to build/ (keep root dir uncluttered)
 21:14
kudo: cce3f9e | pmichaud++ | LICENSE:
Add LICENSE file.
shorten dalek's url is at xrl.us/beexog
shorten dalek's url is at xrl.us/beexoi
rurban pmichaud: you can close TT #267 now if you care 21:17
Coke rurban: given that languages/perl6 is still in the repository... 21:23
pmichaud It's about to disappear (<24h)
dalek kudo: 159a3a5 | pmichaud++ | (2 files):
Move ROADMAP file to docs/ .
 21:30
shorten dalek's url is at xrl.us/beexp2
dalek kudo: e45ea51 | pmichaud++ | CREDITS:
Add CREDITS (distilled from Parrot repository CREDITS file).

were specific to Rakudo. If your entry was removed and you want to remain in CREDITS, just send a patch and we'll put you back in.
shorten dalek's url is at xrl.us/beexp4
dalek kudo: 1ab9b07 | pmichaud++ | .gitignore:
Update .gitignore to ignore parrot/ subdirectory.
shorten dalek's url is at xrl.us/beexp6
21:33 ask_ joined 21:37 ask- joined
dalek rrot: r36362 | rurban++ | trunk/src/packfile.c:
[core] Revert the r36294 change for 64-bit to fix packfile errors.

no alignment at all.		 21:38
21:41 kj joined
rurban Should parrot -h omit -j --jit-core when no JIT is available? same for HAVE_COMPUTED_GOTO 21:48
users will ask which core does my parrot have. We can just say parrot -h
chromatic That's an intriguing thought. Can you ask the list? 21:49
rurban parrot -h | grep -- -core
NotFound parrot_config jitcapable 21:51
rurban I'll do. maybe we also need a switch to report the cores, Because -R CORE begs also for an explanation
But I just broke the build, oh no.
rg1 you did? you fixed the tests for me :) 21:52
rurban src/packfile.c:4495: undefined reference to `_Parrot_str_not_equal'
rg1 is your complete source current? 21:53
rurban That's the question...
Coke that was recently touched; possible it is missing a PARROT_MUMBLE attribution.
chromatic PARROT_EXPORT? 21:54
That shouldn't be, as it's in *the same shared library*.
... with the caveat that I have no idea what perversions Apple has made to its linker, or what chthonic horrors lurk in MSVC. 21:55
dalek kudo: c108072 | (Moritz Lenz)++ | build/gen_metaop_pir.pl:
XopX is now spelled Xop
 21:56
shorten dalek's url is at xrl.us/beextj
22:01 davidfetter joined 22:07 Whiteknight joined
rurban svn up was the trick :) 22:16
22:26 szabgab joined 22:28 iblechbot joined 22:30 gryphon joined
dalek rrot: r36363 | kjs++ | trunk/languages/pod:
[pod] heavily fix languages/pod, based on fperrad++ 's Markdown implementation (not in repository).
 22:36
rrot: r36364 | kjs++ | trunk/languages/pod:
[pod] restore some Pod::DocTree nodes and remove last bits of markdown code.
 22:41
22:42 rurban_ joined 22:47 Aisling joined
dalek rrot: r36365 | kjs++ | trunk/languages/pod/src:
[pod] last revision doesn't build for pod; fixed this. + some more action.
 22:48
rrot: r36366 | kjs++ | trunk/languages/pod/src/parser:
[pod] paragraph, formatted text. all of this needs to be looked into more carefully, especially the whole thing with blocks and lists: these can be nested, so they must be stored on stacks, and they must have "current block" variables, so to add =items and other stuff to the current list/block.
 22:55
rrot: r36367 | kjs++ | trunk/languages/pod/src/parser:
[pod] =over, =item, =back. Again, needs more work.
 23:02
rrot: r36368 | kjs++ | trunk/languages/pod/src/parser:
[pod] last bits for today. =encoding; + I think you can use up to 4 '<' as in '<<<<' for format_code, so add the 4-<-version.
 23:08
23:12 Limbic_Region joined
rurban gdb ./miniparrot => gdb: error whole loading shared libraries: libexpat.so.0 ...? 23:12
okay, just a gdb problem, fixed it. 23:18
23:18 Casan joined
rurban my icc 10 build crashes at imcc/parser_util.c: is_op("get_params") with config_lib.pasm. That's the very first op. Strange. 23:37
23:39 bacek_ joined
Infinoid What's the crash? 23:42
rurban I just cannot paste from that box. But I assume it's because I linked with gnu ld, not with icc 23:43
Nope, linking with icc is the same. interp->op_lib is random 23:45
Infinoid which platform is this? 23:48
icc worked last week on x86-64
23:49 TiMBuS joined
rurban gentoo-i686 with icc 10. all interp->op_* pointers are wrong 23:49
op_count is 324608
particle sizeof int == sizeof void * ? 23:50
i expect not, which is why parrot fails to build with msvc-x64
rurban plain i686, intsize 4 ptrsize 4
NotFound particle: not granted by C std
rurban I'll continue tomorrow, have to go to sleep now. bye 23:51
particle ? there is no c standard for that, that i'm aware of
there's ILP, LLP, and LP iirc
rg1 shouldn't that be sizeof long == sizeof void * ? 23:52
NotFound particle: yes, there is no standard for that, then it does not grant it
Infinoid particle: do you mean... yeah, what rg1 said
particle yes, that's what i mean
rurban int is always 4, long is 8 on 64bit
particle brain's a little fried from an interview
TiMBuS sizeof uintptr_t == sizeof void*
Whiteknight particle: where did you interview? 23:53
any place fun?
particle valvesoftware.com
Infinoid rurban: the problem with win64 is, long is still 4
rurban aah, very special again!
rg1 infinoid: for real?
particle yes, long is 4, void* is 8 23:54
Whiteknight long long is 8
Infinoid yep, hence calling it LLP
whereas linux/x86-64 is LP
particle long long isn't c89, though
so, we kinda make an exception there
TiMBuS waves arms
Infinoid is waiting for 128 bit and long long long 23:55
NotFound uintptr_t can be used where stdint.h is available and defines it
particle far far far *
we've almost got the star wars opening
Infinoid way far
heh.
23:57 szabgab joined
particle manyeyes.alphaworks.ibm.com/manyeye...0255111976 23:59
shorten particle's url is at xrl.us/beeyca
particle please tell me you don't need a login to see that

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!