Camelia, the Perl 6 bug

IRC log for #parrot, 2009-04-29

Parrot | source cross referenced

| Channels | #parrot index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:03 allison bacek: yes, but we don't have any deprecation items eligible in 1.4
00:04 allison bacek: only eligible in 1.5 (which can't be removed until after the 1.4 release)
00:05 allison bacek: if you're adding a new deprecation item, it has to be eligible in 1.5, because it didn't have a deprecation notice in 1.0, and so can't be removed until it has a deprecation notice shipped in 1.4
00:06 allison Whiteknight: I'm just translating the LongTermRoadmap to tickets, so I can delete the wiki page (and adding more details as I create the tickets)
00:09 AndyA joined #parrot
00:13 bacek_ allison: Parrot_MMD_method_name marked as DEPRECATED for a long time. Is it eligible to remove it now?
00:14 allison bacek_: what's the ticket #?
00:14 bacek_ allison: can't find... Looks like there is no ticket for it
00:14 allison bacek_: IIRC, there was some piece of code depending on it that kept us from removing it, even though it doesn't work anymore
00:14 allison bacek_: where's the deprecation notice?
00:14 purl the deprecation notice is probably committed now.
00:15 bacek_ src/oo.c
00:15 bacek_ line 478
00:15 allison bacek_: okay, it needs a deprecation notice in DEPRECATED.pod
00:15 bacek_ ok. For 1.5?
00:16 allison yes, marked as eligible in 1.5
00:16 bacek_ ok
00:16 bacek_ will do
00:21 allison bacek_: thanks!
00:22 bacek_ TT#625
00:22 dalek parrot: r38392 | bacek++ | trunk/DEPRECATED.pod:
00:22 dalek parrot: Add deprecation notice for Parrot_MMD_method_name
00:22 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38392/
00:22 * bacek_ returning to $work for managing own projects :)
00:32 * allison finished converting all remaining roadmap items to Trac tickets
00:33 darbelo left #parrot
00:34 Infinoid lots of new and exciting tickets
00:34 Infinoid allison++
00:35 Tene allison++
00:41 cotto (whoever implements them)++
00:42 Whiteknight Infinoid: I know right? Lots of exciting new things
00:44 Infinoid yep, lots of shinies
00:50 cotto karma for whoever implements them
00:50 purl whoever implements them has karma of 1
00:52 Whiteknight (whoever implements them)++
01:10 eternaleye joined #parrot
01:48 dalek parrot: r38393 | allison++ | trunk/MANIFEST.generated:
01:48 dalek parrot: [cage] Don't install both the "installable" and regular versions of
01:48 dalek parrot: executables, only install the ones intended for installation. Resolves
01:48 dalek parrot: TT #347.
01:48 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38393/
02:04 dmknopp joined #parrot
02:06 * Coke stares at quevlar's code.
02:09 Coke I will refrain from commenting on that ticket, and hope that whoever does is gentle.
02:24 cognominal joined #parrot
02:32 bacek joined #parrot
02:33 kid51 joined #parrot
02:35 janus joined #parrot
02:40 tetragon joined #parrot
03:02 dmknopp left #parrot
03:05 davidfetter joined #parrot
03:22 nopaste "tene" at 166.70.38.237 pasted "list() fix for pmichaud" (20 lines) at http://nopaste.snit.ch/16437
03:26 Coke tene; should the 'ns' checks go into the block?
03:26 Tene Can I declare ns in the if{} ?
03:26 Tene in C89?
03:27 Coke I think so.
03:27 Coke you remove code that already does that.
03:30 nopaste "pmichaud" at 72.181.176.220 pasted "refactored list() to move ns lookup inside of if()" (14 lines) at http://nopaste.snit.ch/16438
03:36 Tene pmichaud: committed.  Testing with rakudo/hll now...
03:36 pmichaud cool
03:36 pmichaud I have to take a break for a while
03:37 Tene kk
03:37 Tene Thanks. :)
03:37 pmichaud I would say that the easiest thing to start doing is backporting some of the changes from the branch into the trunk
03:37 pmichaud er, master
03:37 Tene Could not find non-existent sub !fixup_routine_type
03:37 Tene is the next error
03:37 Tene that shouldn't be too bad
03:37 pmichaud yes, that's what I was getting when bypassing the Capture issue.
03:38 dalek parrot: r38394 | tene++ | trunk/src/pmc/capture.pmc:
03:38 dalek parrot: Fix a class lookup issue with capture.pmc
03:38 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38394/
03:40 pmichaud afk
03:43 Tene ah, I know what it is.
03:47 Tene Yay, new segfault!
03:51 Tene Same loop as before.
03:53 ilia_ joined #parrot
03:55 Tene pmichaud: when debugging things like this, ulimit helps; 'ulimit -s 1' will limit the stack to 1k
03:56 Tene I'm getting a gdb segfault when I try to view the bottom of the stack...
03:58 bacek_ Tene: "bt 10' will help
04:00 Tene pmichaud: check this out, from pmcproxy.pmc:
04:00 Tene STRING *cname  = pclass->vtable->whoami;
04:00 Tene PMC *pproxy = Parrot_oo_get_class_str(interp, cname);
04:03 Coke urk?
04:03 purl urk is this a global past now?
04:03 Coke forget urk
04:03 purl Coke: I forgot urk
04:09 eternaleye_ joined #parrot
04:12 Tene It compiles!
04:12 purl Ship it!
04:12 Tene And it compiles Test.pm!
04:12 Tene spectesting...
04:12 Tene and testing Parrot...
04:15 Tene Okay, rakudo can compile things, but can't run them yet.  null pmc in invoke().
04:16 dalek parrot: r38395 | tene++ | trunk/src/pmc/pmcproxy.pmc:
04:16 dalek parrot: Fix an issue with PMCProxy using strings to look up classes.
04:16 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38395/
04:42 tetragon joined #parrot
04:51 Andy joined #parrot
04:52 cotto What's a really good perl module for profiling?  I want to know what the output from a good Parrot profiler should look like.
04:52 Tene Devel::nytProf
04:52 purl Devel::nytProf is, like, probably getting memory profiling soon
04:53 cotto good to know
04:54 cotto Thanks
04:55 cotto although I should see if partcl can be rehabilitated first
05:03 dalek parrot: r38396 | cotto++ | trunk/config/gen/crypto/digest_pmc.in:
05:03 dalek parrot: [PMC] add VTABLE decorations to functions in the geneated md* and sha* PMCs
05:03 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38396/
05:12 flh joined #parrot
05:29 masak joined #parrot
05:39 japhb joined #parrot
05:42 Tene Haha, this is hilarious.
05:42 Tene the last thing to get working in the hll branch (I hope) is exporting
05:43 Tene and it looks like rakudo is exporting to the perl6;whatever;E;X;P;O;R;T;D;E;F;A;U;L;T namespace
05:43 Tene enoclue why it's breaking it up like that.
05:45 Tene it's interpreting the string as an array of strings.
05:46 Tene and that's because the vtable->base_type isn't enum_class_String
05:51 Tene if i switch it to a call to VTABLE_isa... it works
05:51 Tene testing the parrot change...
05:53 Tene spectesting rakudo...
05:53 Tene tests are passing!
05:53 Tene It seems somehow slower than before... maybe.
05:53 Tene oops!  An error!
05:54 Tene more passing...
05:59 dalek parrot: r38397 | tene++ | trunk/src/global.c:
05:59 dalek parrot: Fix a type comparison to avoid breaking make_namespace for HLLs that hll_map parrot strings, I think.
05:59 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38397/
06:15 uniejo joined #parrot
06:15 cotto slow and correct is better than fast and broken
06:15 cotto fast and correct is even better
06:27 chromatic joined #parrot
06:30 Tene purl: msg jonathan ClassHOW really could use some updates... subclassing... I'm going to try to do it, but don't know if i'll succeed
06:30 purl Message for jonathan stored.
06:47 iblechbot joined #parrot
06:50 Tene purl: msg allison it looks like the consensus on the list was to make HLL pbcs with the same case... having said that, .HLL 'Perl6' puts stuff in the 'perl6' namespace.  note the case.  should this be changed?
06:50 purl Message for allison stored.
06:53 Tene purl: msg jonathan i got a workaround... it's not right, but it gets whatever working again.  Looks like just a few spectest failures left.
06:53 purl Message for jonathan stored.
06:56 Tene purl: msg jonathan segfault in t/spec/S03-junctions/autothreading.rakudo
06:56 purl Message for jonathan stored.
07:03 Gerd joined #parrot
07:11 japhb_ joined #parrot
07:14 bobke joined #parrot
07:29 bacek hi there
07:30 cotto back at you
07:36 bacek cotto: howisgoing?
07:38 cotto other than unemployedness, life is good
07:39 bacek it's... good.
07:40 bacek Any thought about DOTPLAN in pmc_pct?
07:40 cotto I've been distracted by trying to get partcl working again.
07:40 moritz cotto: if you're looking for something temporary, you could apply for some TPF or Hague grant for hacking on parrot or Rakudo
07:41 cotto moritz, that'd be pretty cool.
07:41 bacek moritz++ # good idea
07:41 cotto I'll look into that.
07:41 cotto moritz++
07:41 cotto (of course it's not a long-term solution, but I'd love to get paid to hack on Parrot)
07:41 moritz cotto: http://www.perlfoundation.org/ian​_hague_perl_6_development_grants
07:41 shorten moritz's url is at http://xrl.us/beqstu
07:41 cotto That was my next question.
07:43 szbalint I guess the upside is having free time for once
07:43 cotto I wonder if Parrot profiling would qualify, since it'd benefit rakudo.
07:44 cotto It's worth a shot to look into what'd be required to implement it.
07:44 moritz cotto: I guess you need something more specific to rakudo
07:45 cotto I'll ask rdice next time I see him around.
07:45 moritz cotto: talk to jonathan and pmichaud, maybe they can tell you what they need in Parrot for rakudo, to improve your chances
07:45 cotto moritz, you're full of good ideas today.
08:18 cotto man.  partcl's a mess
08:18 cotto It's not happy living out of the svn repo
08:24 cotto It's kinda nice working with an installed Parrot, now that the installation works.
08:34 cotto msg Coke Do you want partcl to eventually build outside of languages/tcl in Parrot's build dir?
08:34 purl Message for coke stored.
08:42 davidfetter i love the smell of purl in the morning
08:43 cotto Mmmm.  botsteak.
08:43 cotto probably with spam
08:58 HG` joined #parrot
09:45 dalek parrot: r38398 | bacek++ | branches/pmc_pct (7 files):
09:45 dalek parrot: Add handling of VTableInfo.
09:45 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38398/
09:48 dalek parrot: r38399 | bacek++ | branches/pmc_pct/compilers/pmc/t/04-header.t:
09:48 dalek parrot: Temporary disable emitting test. It's useless ATM.
09:48 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38399/
09:55 riffraff joined #parrot
09:56 amoc joined #parrot
10:41 cognominal joined #parrot
10:45 dalek parrot: r38400 | bacek++ | branches/pmc_pct (1 files):
10:45 dalek parrot: Rename PMC::Compiler to PMC::Emitter
10:45 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38400/
10:45 dalek parrot: r38401 | bacek++ | branches/pmc_pct/compilers​/pmc/src/vtable_info.pir:
10:45 dalek parrot: Fix init flags
10:45 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38401/
10:45 dalek parrot: r38402 | bacek++ | branches/pmc_pct (4 files):
10:45 dalek parrot: First cut for implementing PMC::Emitter
10:45 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38402/
11:12 Whiteknight joined #parrot
11:14 iblechbot joined #parrot
11:36 dalek parrot: r38403 | bacek++ | branches/pmc_pct/compilers/pmc/t/04-header.t:
11:36 dalek parrot: Refactor 04-header test for future extension.
11:36 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38403/
11:36 dalek parrot: r38404 | bacek++ | branches/pmc_pct/compilers/pmc/src/builtins.pir:
11:36 dalek parrot: Add 'uc' function
11:36 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38404/
11:36 dalek parrot: r38405 | bacek++ | branches/pmc_pct/compilers/pmc (2 files):
11:36 dalek parrot: Generate header guards
11:36 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38405/
11:42 ruoso joined #parrot
11:58 whoppix joined #parrot
12:05 amoc joined #parrot
12:15 FL2 joined #parrot
12:22 dalek parrot: r38406 | bacek++ | branches/pmc_pct/compilers​/pmc/src/vtable_info.pir:
12:22 dalek parrot: Fix param passing to vtable_info
12:22 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38406/
12:22 dalek parrot: r38407 | bacek++ | branches/pmc_pct/compilers/pmc (3 files):
12:22 dalek parrot: Build VTable hash
12:22 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38407/
12:22 dalek parrot: r38408 | bacek++ | branches/pmc_pct/compilers/pmc (2 files):
12:22 dalek parrot: Start generating C function definitions
12:22 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38408/
12:25 bacek purl: msg cotto pmc_pct branch is "playable" now. Just in case if you'll have time to play with it.
12:25 purl Message for cotto stored.
12:25 bacek afk #sleep
12:25 bacek o!
12:28 Coke msg cotto ,allison - partcl now linking against the right lib.
12:28 purl Message for cotto stored.
12:28 Coke msg allison ,cotto - partcl now linking against the right lib.
12:28 purl Message for allison stored.
12:28 Coke msg cotto the goal is to compile against an installed parrot only.
12:28 purl Message for cotto stored.
12:30 Coke msg cotto partcl has never worked against anything other than a build dir; but we're halfway to that, /AND/ there are deprecations that have gone, so it's hard to fix everything at once.
12:30 purl Message for cotto stored.
12:30 Coke msg cotto: biggest blocker for me right now is the struct macros.
12:30 purl Message for cotto stored.
12:37 amoc joined #parrot
12:37 rg1 joined #parrot
12:56 LylePerl_ joined #parrot
13:10 gryphon joined #parrot
13:25 Coke message bacek - your 'uc' sub is unnecessarily shimmering. just delcare a string parameter and the calling conventions will do the string conversion for you.
13:25 purl Message for bacek stored.
13:39 * Coke wonders how long ago morph started taking not an integer.
13:40 Coke message cotto I can't fix it running until you fix it building. =-)
13:40 purl Message for cotto stored.
14:02 Andy joined #parrot
14:25 amoc joined #parrot
14:26 iblechbot joined #parrot
14:28 fperrad joined #parrot
14:34 gaurav joined #parrot
14:37 Theory joined #parrot
14:40 Coke ~/ I've got a theory... It's a demon... A dancing demon. no, something isn't right there. /~
14:54 masak joined #parrot
15:06 nopaste "tene" at 166.70.38.237 pasted "Rakudo test summary report for pmichaud++" (75 lines) at http://nopaste.snit.ch/16445
15:07 particle joined #parrot
15:15 nopaste "tene" at 166.70.38.237 pasted "git diff from hll to master on rakudo for pmichaud++" (503 lines) at http://nopaste.snit.ch/16446
15:16 moritz karma pmichaud
15:16 purl pmichaud has karma of 2657
15:16 moritz karma moritz
15:16 purl moritz has karma of 461
15:20 Infinoid karma me
15:20 purl infinoid has karma of 604
15:20 Infinoid whoa, I finally broke 500?  huh.  when did that happen.
15:20 particle karma me
15:20 purl particle has karma of 1532
15:20 particle sheesh, for doing nothing?
15:20 particle i must have done a lot of it.
15:20 * Infinoid accelerates particle
15:24 Infinoid gah.  so many projects, so little time
15:24 Infinoid I'm adding a testsuite to dalek's feed parsers, probably gonna throw all of that up on github
15:24 Infinoid ...and then add a feedparser for that too, heh.
15:34 pmichaud karma leo
15:34 purl leo has karma of 1885
15:34 pmichaud karma jonathan
15:34 purl jonathan has karma of 1277
15:34 pmichaud karma Coke
15:34 purl coke has karma of 2733
15:34 NotFound karma purl
15:34 purl purl has karma of 8548
15:34 pmichaud purl-- # just because.
15:34 purl pmichaud: huh?
15:35 Infinoid karma purl
15:35 purl purl has karma of 8548
15:35 Infinoid cheater.
15:36 pmichaud foo purl-- # does this work?
15:36 pmichaud karma purl
15:36 purl purl has karma of 8547
15:36 Infinoid a-ha!
15:36 pmichaud okay.  But it's a pretty lousy scorekeeper that cannot win a game.  :-)
15:52 jhorwitz joined #parrot
15:53 * Theory dances like a demon for coke
15:59 gerd joined #parrot
16:00 gerd left #parrot
16:16 flh joined #parrot
16:22 davidfetter joined #parrot
16:24 amoc joined #parrot
16:31 dalek partcl: r334 | wcoleda++ | trunk/config/makefiles/ (2 files):
16:31 dalek partcl: link against an installed parrot!
16:31 dalek partcl: doughera++ for finding the option in parrot_config that I missed.
16:31 dalek partcl: review: http://code.google.com/p/p​artcl/source/detail?r=334
16:31 dalek partcl: r336 | wcoleda++ | trunk/ (2 files):
16:31 dalek partcl: We don't live in "languages/tcl" anymore.
16:31 dalek partcl: Patch courtesy allisonrandal++
16:31 dalek partcl: review: http://code.google.com/p/p​artcl/source/detail?r=336
16:31 dalek partcl: r338 | wcoleda++ | trunk/config/makefiles/root.in:
16:31 dalek partcl: Clean up new or relocated files.
16:31 dalek partcl: review: http://code.google.com/p/p​artcl/source/detail?r=338
16:32 AndyA joined #parrot
16:57 jhorwitz i think i heard mention of this yesterday but can't find the logs....how are we supposed to build dynpmcs using only an installed parrot?
16:58 gryphon joined #parrot
17:01 uniejo joined #parrot
17:09 fperrad coke?
17:09 purl i guess coke is Will Coleda <mailto:will@coleda.com> or perpetually annoyed.
17:11 fperrad ping coke
17:11 purl I can't find coke in the DNS.
17:13 pmichaud jhorwitz: it's supposed to be done using some specific 'make' targets, iirc
17:14 pmichaud see tools/dev/mk_language_shell.pl for an example
17:14 * jhorwitz roots around in mk_language_shell.pl
17:15 jhorwitz oh ok.  pmc2c.pl and friends are in LIB_DIR.
17:32 chromatic joined #parrot
17:51 HG` joined #parrot
17:52 amoc joined #parrot
17:56 barney joined #parrot
18:03 japhb joined #parrot
18:05 cotto so many messages
18:07 cotto Coke, ping
18:09 cotto msg Coke partcl now *builds* against an installed svn head.  It doesn't do much else, but building is progress.
18:09 purl Message for coke stored.
18:12 DietCoke joined #parrot
18:12 DietCoke .
18:12 DietCoke nick Coke
18:12 Coke whoops
18:14 cotto hi
18:14 Coke hio. saw your patch, investigating...
18:15 Coke ok. builds, now can't load dynops.
18:16 cotto I think it'll dtrt, since pmc_reuse takes care of destruction/init of the old/new PMC.
18:16 cotto one step closer
18:19 ruoso joined #parrot
18:21 iblechbot joined #parrot
18:23 amoc joined #parrot
18:24 Coke Parrot_load_lib doesn't seem to have a way to change the load path, does it.
18:28 Ademan_ joined #parrot
18:32 Coke cotto: http://code.google.com/p/p​artcl/issues/detail?id=75
18:37 cotto everyone has issues
18:38 pmichaud Once again, partcl leads the way :-)
18:39 pmichaud Coke++ cotto++
18:39 Coke I was about to ask on list if anyone was building against an installed parrot with dynops.
18:39 pmichaud About 60% of the time when Rakudo "solves" a problem, it turns out partcl solved it first :-)
18:39 pmichaud Coke: Rakudo needs to do that, but I haven't had the tuits to spend on it
18:40 pmichaud although supposedly rurban and fperrad have come up with ways to do it.  I just haven't had time to try them out.
18:40 Coke I'm happy to let someone else do it first. I'll ask on list for pointers. =-)
18:45 darbelo joined #parrot
18:53 * Coke ponders replacing Data::Dumper with JSOn.
18:54 davidfetter JSON ftw :)
18:55 Tene pmichaud: can you confirm that my commit in r38397 is good?  All tests passed, but I'm not sure of the semantics of vtable_isa
18:55 pmichaud Tene: checking.
18:55 purl well, checking is just different
18:57 LylePerl joined #parrot
19:00 pmichaud Tene: r38397 looks reasonable to me
19:00 Tene ok, thanks. :)
19:09 pmichaud Tene: how should we see about re-synchronizing the hll branch with master?  I'm not sure I want to "fix everything in hll and then do a massive merge"
19:10 Tene pmichaud: hll branch is currently synced with master.
19:11 Tene pmichaud: I merged master into the branch, and everything applied without problems.
19:11 pmichaud Tene: okay, but I'm in the mood to do some major refactors in master.
19:11 Tene Ah.
19:11 pmichaud that I think might cause problems with hll
19:12 Tene sure, strip all the .HLL lines and merge it back into master, then delete the branch.
19:12 Tene That's fine with me.
19:12 Tene actually... leave the branch there, for now.
19:12 pmichaud well, I suspect we're fairly close to getting .HLL to work now.
19:13 pmichaud and part of my refactor is (intended) to fix it so that we only need one .HLL directive at the top of perl6.pir
19:13 Tene Okay.
19:13 pmichaud so, what's the git command to merge hll into master?
19:13 pmichaud (having never done a git branch merge)
19:13 Tene git checkout master; git merge hll
19:13 PerlJam pm: git checkout master; git merge hll
19:14 moritz ;-)
19:14 pmichaud ah yes, once again we behold the incredible awesome power of git.
19:14 Tene but, we don't want to do that until we track down these segfaults.
19:14 Tene i can't work on that until tonight.
19:14 pmichaud ...but the segfaults are due to the .HLL directive, yes?
19:15 pmichaud if I take the .HLLs out, we shouldn't get segfaults.
19:15 Tene Right.
19:15 pmichaud okay, I have no problem with doing that.
19:15 Tene Okay.
19:15 pmichaud most of what has happened in the hll branch should be transparent to master  *if* the .HLLs aren't present.
19:15 Tene Just spectest first, after removing them, to make sure.
19:15 pmichaud oh yes, of course.
19:15 Tene It DID pass everything without .HLL a few revisions back.
19:16 pmichaud I nearly always spectest prior to any sort of a push.
19:16 Tene i occasionally don't, so I'm trying to make sure i'm in the habit of reminding myself. :)
19:18 pmichaud well, I figure if I break the build/spectests then I have much less authority to yell at others who do it :-)
19:19 pmichaud (not that I ever plan to do any yelling about it.  forgiveness better than permission, and all that :-)
19:19 Tene when do you expect to have your refactoring done?
19:20 pmichaud this afternoon
19:20 Tene pmichaud: also, make sure you deal with the specific mention of the 'perl6' hll namespace in ClassHOW.pir
19:20 pmichaud I'm waiting on a spectest now.  Then commit, then push.  Then merge.
19:20 pmichaud and yes, I'll fix up the ClassHOW.pir stuff.
19:20 pmichaud then commit, then spectest, then merge
19:20 pmichaud er s/merge/push/
19:21 jhorwitz pmichaud: switching to RT #65046 real quick.  Foo.new was resulting in Foo being parsed as a listop b/c it's an unknown identifier.  i was able to get around that using ::Foo.new() -- is that a bug or proper behavior?
19:21 pmichaud jhorwitz: proper.
19:21 pmichaud ::Foo says that Foo is a type.
19:21 jhorwitz excellent.  i like proper solutions.  :)
19:23 mikehh_ joined #parrot
19:25 jhorwitz pmichaud: want me to close that ticket out?
19:28 pmichaud jhorwitz: sure, if you're happy with the resolution for now.
19:29 jhorwitz yeah, i still don't think everyone understands the overarching problem, but this particular issue is solved.  i can open another ticket with a more general description.
19:35 pmichaud I don't think we need an additional ticket.
19:35 pmichaud so I'd just close this one.
19:35 pmichaud (and mark what your resolution was)
19:36 jhorwitz done and done.  :)
19:37 dalek parrot: r38409 | coke++ | trunk/docs/project/release_manager_guide.pod:
19:37 dalek parrot: slight update to trac procedures.
19:37 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38409/
19:37 pmichaud jhorwitz++
19:41 davidfetter pmichaud, how far do you estimate rakudo is away from being package-able in the sense of requiring some installed version of parrot, etc.?
19:42 pmichaud davidfetter: depends on the degree of package-ableness
19:42 davidfetter could you elucidate that a bit?
19:42 pmichaud I think fperrad is already packaging up instances of rakudo for Windows
19:43 pmichaud in theory, each rakudo monthly release should be buildable against the previous parrot release.
19:44 pmichaud we're not building against installed parrot yet because all of the pieces for that aren't in place yet.
19:44 pmichaud see, for example, http://lists.parrot.org/pipermail/​parrot-dev/2009-April/002154.html
19:44 shorten pmichaud's url is at http://xrl.us/beqvaa
19:44 Infinoid fperrad has a win32 build of Parrot-1.1.0-Rakudo-16 and various other things on http://parrotwin32.wiki.sourceforge.net/
19:45 davidfetter i guess i was thinking of "packageable" as including "builds against an installed parrot"
19:47 pmichaud Right.  For that, we need to update some makefiles.  But the problem mentioned in the parrot-dev post still exists.
19:47 pmichaud i.e., we can build but the dynops aren't available.
19:48 pmichaud (so it doesn't run)
19:48 darbelo joined #parrot
19:50 davidfetter k
19:50 davidfetter thanks :)
19:50 * davidfetter not much of a dev on this stuff, but has been trying to help package :)
19:55 darbelo joined #parrot
19:56 fperrad Coke, a fix for linking ParTcl on win32
19:56 fperrad see http://parrot.privatepaste.com/44A8s8axwS
20:00 Coke fperrad: the ops already link.
20:00 Coke it's actually loading them that seems to fail.
20:00 Coke also, see: https://trac.parrot.org/parrot/ticket/627
20:01 * Coke tries out that patch regardless.
20:02 fperrad Coke, I can't link on MinGW without my patch
20:02 fperrad tcl_ops_cg.o: In function `cg_tcl':
20:02 fperrad C:/fperrad/Parrot/trunk/langu​ages/tcl/src/ops/tcl.ops:40: undefined reference to `ParTcl_binary_scan'
20:02 Coke ah, ok.
20:03 jhorwitz davidfetter: going to YAPC?
20:03 purl See : going to yapc::na or yapc::whatever or cwest, uri, DrForr
20:03 Coke that didn't complain on darwin. would ../binary work for you?
20:03 Coke (the \ fails)
20:04 davidfetter jhorwitz, hope so. if i get a speaking gig, i go
20:04 jhorwitz cool
20:04 * davidfetter hasn't been to pittsburg in years, and would like to see it
20:04 Coke fperrad: committed with a ../binary$(O)
20:04 Coke danke.
20:04 davidfetter there's an off chance i might be at fisl then, though :P
20:04 dalek partcl: r339 | wcoleda++ | trunk/config/makefiles/ops.in:
20:04 dalek partcl: Update link to reflect deps.
20:04 davidfetter not that i dislike brazil or anything, but...
20:04 dalek partcl: Courtesy fperrad++
20:04 dalek partcl: review: http://code.google.com/p/p​artcl/source/detail?r=339
20:05 pmichaud phone
20:05 davidfetter is anybody going to linux malaysia end of may?
20:12 allison purl: msg Tene yes, we should be consistent in case everywhere, also change the directory and name of the main .pbc file to 'Perl6' (or change the .HLL and compiler name to 'perl6')
20:12 purl Message for tene stored.
20:13 Tene allison: when I say .HLL 'Perl6', it stores things in the 'perl6' HLL namespace.  Are you saying that we should change that behavior to preserve case?
20:14 allison Tene: ugh, yes
20:14 pmichaud did we decide that implementations should use the name of the language or the name of the implementation?
20:14 pmichaud i.e., should it perhaps be .HLL 'rakudo' instead?
20:14 allison pmichaud: we decided they should pick one name, we don't care what it is
20:14 pmichaud okay.
20:15 pmichaud fwiw, I prefer "perl6" to "Perl6"
20:15 allison pmichaud: then make it so
20:16 allison at some point we'll probably have to set up aliasing so "pynie" and "python" work "pipp" and "php"
20:17 pmichaud I think I'll be able to do that in HLLCompiler w/o too much trouble.
20:17 allison but, that's future
20:17 Coke allison: do you have access to darwin/x86?
20:17 Coke (regarding the partcl build)
20:17 allison pmichaud: well, it extends beyond HLLCompiler, since it affects namespaces, etc
20:17 pmichaud ah.
20:18 pmichaud I somewhat figure we get hold of the compiler first, and then let it map namespaces as appropriate.
20:18 allison pmichaud: aye, it may be possible to put in that restriction
20:19 pmichaud iow, I think that cross-language issues ought to be handled by the appropriate compiler object as much as possible.
20:19 allison Coke: not right here, but I'm headed over to the part of town where that laptop is this afternoon
20:19 allison pmichaud: agreed
20:20 Coke hokay. any pointers helpful.
20:20 Tene for example, we shouldn't be grabbing ['perl6'] explicitly, we should ask the cmpiler for the namespace.
20:20 Coke we're much further along than we were. =-)
20:20 Tene hllns = compiler.'get_namespace'() oslt
20:20 Coke thanks for the legwork you did.
20:20 pmichaud correct.  Or, even better, asking the compiler for whatever symbol we're wanting to look up int he namespace.
20:20 allison Coke: I suspect it's a makefile problem, that the ops aren't actually building (because I had a similar problem with the makefile already)
20:20 Coke they seem to be building.
20:20 allison Coke: I have a patch you can try
20:20 Coke happy to try it.
20:22 allison Coke: it may have no effect, but I'll attach it to the ticket
20:24 allison Coke: another question is which version of parrot you're building on? searching dynext/ in the current directory wouldn't have been supported before 0.9.x or so
20:26 Coke allison: no visible effect from your patch.
20:26 Coke allison: 1.1
20:27 Tene pmichaud: you still haven't committed the "remove .HLL directives and merge into master".  do you want me to do it?
20:28 pmichaud I'm on phone -- you can do it if you want :-)
20:32 allison Coke: ah, I haven't tried 1.1 since last week, will try again
20:56 bacek joined #parrot
20:57 pmichaud Tene: I'm off phone again -- are you working on the merge now or want me to pick it up?
21:03 Coke tene: does your nick rhyme with penne or ballpeen?
21:03 pmichaud Coke:  "penne"
21:03 Coke i was asking because it sounded wrong when you said it. =-)
21:03 pmichaud I asked him about it at PDS
21:03 Coke heh.
21:05 Tene Coke: pmichaud is right
21:05 Tene pmichaud: I was distracted by $job immediately after asking.
21:05 Tene :)
21:07 pmichaud Tene: okay, I already started on it.
21:07 Limbic_Region joined #parrot
21:29 fperrad left #parrot
21:54 HG` joined #parrot
22:05 bacek good morning
22:06 Infinoid hi bacek
22:06 bacek hi Infinoid
22:08 cotto hey bacek, is there an easy way to generate c code from a pmc file in pmc_pct?
22:08 dalek parrot: r38410 | bacek++ | branches/pmc_pct/compilers/pmc/src/builtins.pir:
22:08 dalek parrot: Simplify 'uc'. Coke++
22:08 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38410/
22:08 bacek cotto: yeah. Just implement generate_c_file in Emitter :)
22:09 cotto just a smop
22:09 bacek cotto: I'm going to finish all scaffolding and generating .h files tonight.
22:10 bacek Then it will be slightly easier to implement .c generating.
22:19 cotto bacek, what did you mean that pmc_pct is "playable"?
22:19 bacek cotto: It can partially generate header files. And ready for extending :)
22:21 cotto That's important.  Is there an easy way to generate a header, or does it require some glue pir?
22:21 bacek 04-header.t
22:30 Whiteknight joined #parrot
22:31 allison Coke: ping
22:33 allison purl: msg Coke I just realized I don't have an example Tcl script to run to test the dynops, do you have one that you're using?
22:33 purl Message for coke stored.
22:33 Theory_ joined #parrot
22:35 chromatic TT #627 has one
22:35 kid51 joined #parrot
22:38 allison chromatic: hmmm... was looking for one in Tcl, rather than in PIR
22:38 chromatic As far as I remember Partcl, any Partcl program should do then.
22:39 Whiteknight does TCL even compile/run right now?
22:39 Whiteknight last i heard, I didn't think it did
22:39 allison Whiteknight: yes, it does
22:39 bacek cotto: r38412 handles vtables almost properly during parsing/header emitting
22:39 allison Whiteknight: it even runs 'make test', though it fails a large percentage at the moment
22:39 Whiteknight oh, then I am sorry for being mistaken
22:39 bacek afk # $day_job
22:40 allison Whiteknight: (mostly problems to do with the testing infrastructure)
22:40 Whiteknight ah, okay
22:40 allison Whiteknight: well, it's a recent fix. only in the past week
22:40 Whiteknight thanks for the update
22:40 Whiteknight kid51: ping
22:41 * allison driving across town to deposit a carload of boxes in a storage container
22:41 dalek parrot: r38411 | bacek++ | branches/pmc_pct/compilers/pmc/src/builtins.pir:
22:41 dalek parrot: Add join.
22:41 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38411/
22:41 dalek parrot: r38412 | bacek++ | branches/pmc_pct/compilers/pmc (4 files):
22:41 dalek parrot: Handle parameters during parsing and emitting
22:41 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38412/
22:41 Whiteknight bacek++
22:43 Whiteknight purl msg kid51 thanks for the ticket, you can pass any others on to me that you need to find a home for.
22:43 purl Message for kid51 stored.
22:58 tetragon joined #parrot
23:13 rdice joined #parrot
23:15 kid51 back from dinner; Whiteknight:  thanks
23:27 TonyC joined #parrot
23:51 whiteknight joined #parrot
23:52 whiteknight any windows users around here?
23:53 kid51 Infinoid helped me out with a Win32 question the other day.
23:54 darbelo ping cotto
23:54 purl I can't find cotto in the DNS.
23:55 Infinoid What have I gotten myself into? :)
23:58 Infinoid whiteknight: Whatcha need?
23:58 whiteknight I'm trying to get a smolder report working on Win64, but I can't even get configure to run
23:59 whiteknight On the compiler config probe, I get the error "Compilation failed with 'cl'"
23:59 whiteknight and I dont know nearly enough about the config system to intelligently debug it

| Channels | #parrot index | Today | | Search | Google Search | Plain-Text | summary

Parrot | source cross referenced