Camelia, the Perl 6 bug

IRC log for #parrot, 2008-07-22

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 dngor_ joined #parrot
00:09 AndyA joined #parrot
00:19 timbunce joined #parrot
00:34 ruoso joined #parrot
00:34 Theory joined #parrot
00:48 kid51 joined #parrot
01:23 dalek r29660 | jkeenan++ | trunk:
01:23 dalek : Merge revisionpm branch into trunk.  Cf.:
01:23 dalek : http://rt.perl.org/rt3/Tic​ket/Display.html?id=56948.  Refactor
01:23 dalek : Parrot::Revision and add tests, including one new test file.
01:23 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29660
01:25 dalek r29661 | jkeenan++ | trunk:
01:25 dalek : Revert to previous version due to unintentional commit of several lines.
01:25 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29661
01:27 dalek r29662 | jkeenan++ | revisionpm:
01:27 dalek : Branch has been merged into trunk and is no longer needed at head.
01:27 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29662
01:27 dalek r29663 | jkeenan++ | revisionpm-29566:
01:27 dalek : Branch to which tag corresponded has been deleted, so tag may go as well.
01:27 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29663
01:38 cout joined #parrot
01:40 dalek r29664 | Whiteknight++ | gsoc_pdd09:
01:40 dalek : [gsoc_pdd09] Fixes to implementation:
01:40 dalek : * Added long-missing support for GC_trace_normal and GC_trace_stack_FLAG
01:40 dalek : * Re-added trace_system_areas() into Parrot_dod_trace_root()
01:40 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29664
01:48 Coleoid joined #parrot
01:51 dalek r29665 | Whiteknight++ | gsoc_pdd09:
01:51 dalek : [gsoc_pdd09] Make GC_finish_FLAG consistent with the other flags.
01:51 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29665
02:10 dalek r29666 | Whiteknight++ | gsoc_pdd09:
02:10 dalek : [gsoc_pdd09] Improve flag logic to account for flag interdependencies
02:10 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29666
02:11 gmansi joined #parrot
02:20 Andy joined #parrot
02:52 Psyche^ joined #parrot
04:33 Andy joined #parrot
04:37 timbunce joined #parrot
04:48 Theory joined #parrot
05:00 Tene purl: coke?
05:00 purl rumour has it coke is mailto:will@coleda.com or just a figurehead. or http://coke-floats.blogspot.com/ or DietCoke or a pest
05:15 Psyche^ joined #parrot
05:18 Tene I kinda fixed Coke's hll map problem.
05:19 japhb OK, Tene, drop the other shoe ...
05:36 dalek r29667 | japhb++ | trunk:
05:36 dalek : Merge branch 'benchmarks'
05:36 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29667
05:37 japhb tewk, chromatic: Not only does it seem to work (well at least, not explode) to treat a ManagedStruct with named members as an array, it seems to be the fastest way to fill at least a small ManagedStruct.  See examples/benchmarks/float4.pir
05:41 timbunce joined #parrot
06:02 timbunce joined #parrot
06:09 uniejo joined #parrot
06:20 Ademan joined #parrot
06:22 timbunce joined #parrot
06:29 tewk joined #parrot
06:55 barney joined #parrot
07:01 masak joined #parrot
07:03 apple-gunkies joined #parrot
07:43 Ademan joined #parrot
08:29 mj41 joined #parrot
08:38 iblechbot joined #parrot
08:44 Debolaz joined #parrot
09:37 donaldh joined #parrot
09:52 masak there's a significant chance Perl 6 will be much about cross-language communication (of data and calls)
09:53 masak I'm backlogging yesterday's conversation
09:55 masak Perl 6 could end up in a position where it makes such communication very easy
09:55 masak but it will take a lot of work
09:59 moritz Perl 6 is one of the languages that actually acknowledges the existance of other languages ;-)
09:59 Whiteknight joined #parrot
10:04 donaldh Yeah, that was an interesting discussion.
10:04 donaldh HLL interoperability is one of the things that makes Parrot really attractive
10:20 barney joined #parrot
10:47 kid51 joined #parrot
11:02 confound joined #parrot
11:18 ruoso joined #parrot
11:25 Whiteknight joined #parrot
12:00 ruoso joined #parrot
12:09 dalek r29668 | jkeenan++ | parallel:
12:09 dalek : Add some basic tests.
12:09 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29668
12:18 Tene japhb: there might be a timing problem.  I'm not sure.  I fell asleep before verifying it.
12:42 schmalbe joined #parrot
12:52 barney joined #parrot
13:05 japhb Tene: fair enough
13:09 japhb masak, moritz, donaldh: yeah, I'm with Coke on this one ... Parrot is selling itself not just on the capability to *host* multiple languages, but to get them to interoperate.  I want 1.0 as much as anyone, but 1.0 without interop is a non-starter in my mind.  If Perl 6 (or Rakudo and thus PCT) force that, so much the better.  :-)
13:10 donaldh japhb: hopefully I captured that in my notes http://www.perlfoundation.org/parro​t/index.cgi?intra_hll_mapping_notes
13:12 japhb donaldh: thanks, reading ...
13:13 moritz japhb: good to hear. I don't think the perl 6 specs actually enforce a particular level of interaction with other languages (except perl 5 perhaps), but it does describe the hooks
13:15 donaldh It sounds like we have a fair amount of work to do before it is a reality tho.
13:15 japhb moritz: nodnod.  I suspect that it will end up being the standard-bearer anyway, because outside people are going to want to test the interop capabilities as we enter the beta phase, and Perl 6 does seem to have more written in that vein than most other languages.
13:17 japhb donaldh: yes, but very important work.  "Sudden change in programming landscape" important.
13:18 donaldh agreed.
13:18 moritz should I create a ticket?
13:18 donaldh I had assumed that it was closer to a reality.
13:19 dalek r29669 | bernhard++ | trunk:
13:19 dalek : [Pipp] PHP 5.3 is the reference implementation.
13:19 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29669
13:19 donaldh I guess there should be a ticket. But that will not help raise the visibility of the subject.
13:19 donaldh Perhaps a blog would help.
13:20 moritz donaldh: but then we have a ticket ;-)
13:20 japhb donaldh: I knew that HLL_map was incomplete (and I was getting nervous because of how long the Rakudo folks have been working around it, seemingly avoiding the main problem) -- I didn't realize that it wasn't just the implementation, but the whole design that was not done.
13:20 donaldh s/blog/blog entry/
13:20 donaldh blogs don't help per-se ;-)
13:21 japhb moritz, donaldh: Yes, a ticket is in order.  Because then RT-watchers will keep pinging the list if the ticket languishes.
13:21 donaldh :D
13:21 moritz ok, I'll write one
13:21 donaldh japhb: Is that what you do?
13:21 japhb donaldh: And yes, blog entry in order.  If we can get more than one person doing so, that would be good ....
13:22 donaldh How long do patches normally sit in the new state for?
13:22 * donaldh has submitted a couple.
13:22 japhb donaldh: Nope.  There are a couple people who do, though.  kid51 comes to mind, but there are definitely others.
13:23 donaldh japhb: did you have a go at making the opengl driver accessible from Rakudo ?
13:23 moritz donaldh: depends on the subsystem, how big the changes are, how easy or hard they are to verify
13:23 japhb donaldh: Oh, that's a weak point.  I've definitely been unhappy about tickets left in new state, but unfortunately I don't know enough of the core yet to be a patch manager.  Sigh.
13:24 gryphon__ joined #parrot
13:26 * donaldh thinks that a quick hack of 'use' is in order so that .pir and .pbc libs can be 'use'd
13:26 japhb donaldh: I keep bugging jonathan and pmichaud to make it possible, because I have a pile of Perl 5+i OpenGL code that I really want to port to Perl 6.
13:27 japhb Both of them keep saying basically "Yes, we know it's really important, we know a lot of people are asking, we're not ready for that yet.  But soon, we promise."
13:27 japhb "Soon" is a getting to be a four letter word on this topic ....
13:28 donaldh Yes, I'd like to start a wee project in Perl 6 but this is a gating feature.
13:28 Infinoid donaldh: (from new ticket discussion above) you're talking about RT #56824 and #56828?
13:28 japhb I didn't just write the Parrot OpenGL bindings out of the goodness of my heart.  ;-)
13:28 Infinoid japhb: to be fair, they've been getting closer and closer to that for quite some time now
13:29 * donaldh wishes ChatZilla wouldn't make rt numbers clickable (as channel names).
13:29 apple-gunkies closer to "soon"? :)
13:30 Infinoid for smaller and smaller values of "soon" :)
13:31 japhb Infinoid: I know, and I'm not mad, or complaining that they're not doing useful stuff -- I'm going to want all of the stuff in Perl 6 that they're working on now, when I port the main corpus -- but I'm frustrated that I can't even start *testing*, or finding out what improvements I need to make to OpenGL.pir, and getting started on that.
13:32 donaldh japhb: agreed. I am astounded at the rate of feature development as it is. But I'd like to be able to contribute to the things I need so much and I don't see a way to contribute.
13:32 japhb Which is why I'm happy to be talking to fperrad, who is starting to use OpenGL.pir in a HLL ... he's just writing a ton of PIR to manually re-wrap OpenGL.pir into Lua, which is OK as a stop-gap, but not where we want to be.
13:32 japhb donaldh: exactly.
13:34 donaldh Infinoid: yes, those two tickets.
13:35 Infinoid ok, great.  #56824 looks relatively harmless for everyone except cygwin, though I don't know why it didn't work on cygwin.  (unless your arguments to memalign are totally different from mine on linux.)
13:35 Infinoid anyway, I can commit that after tests finish
13:35 donaldh okay, cool.
13:36 Infinoid I'm going to stare at #56828 a little longer :)
13:36 japhb donaldh: Looking at your HLL_map notes, it's worth including the pros and cons of changing the M*N direct mapping problem to the M+N map-source-to-base-then-base-to-target method.
13:37 donaldh Infinoid: It appears that cygwin memalign expects both the start _and_ size to be aligned. And it expects to be aligned on 64k boundaries 8-(
13:37 moritz RT #57190 for HLL Interop
13:39 japhb moritz: thank you
13:40 donaldh moritz++
13:40 moritz japhb: no problem
13:40 Infinoid yeah, that's bizarre.  but fair enough, donaldh++ for diagnosing it
13:42 japhb moritz++ # Good "ticket is closable when" section
13:42 Infinoid donaldh: regarding #56828: any chance you can write a test for it?  I believe those formats are tested in t/pmc/nci.t.
13:42 moritz we should make that a requirement for non-bug RT's
13:42 donaldh Infinoid: I was surprised that #56828 wasn't already implemented. I wondered if one of the other signature chars was meant to do void**
13:43 japhb donaldh: Would you mind including a link to this morning's discussion in your HLL_map notes, since moritz's ticket points interested parties to your notes?
13:43 donaldh Infinoid: yes, I'll try amd write a test.
13:43 donaldh japhb: yes, I'll do that.
13:43 japhb donaldh: thansk
13:43 japhb er thanks
13:45 donaldh japhb: just reading the backlog. I think Tene was referring to a specific problem Coke was having with HLL_map not working for ResizablePMCArray.
13:47 japhb nod. But "I kinda fixed Coke's hll map problem." demands *some* form of follow up.  :-)
13:47 dalek r29670 | infinoid++ | trunk:
13:47 dalek : [config] Apply patch to test_exec_cygwin.in from RT#56824.  donaldh++
13:47 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29670
13:53 dalek r29671 | bernhard++ | trunk:
13:53 dalek : [Pipp] Regenerate MANIFEST, after inadvertedly adding a file.
13:53 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29671
13:56 dalek donald@sealgair.com | Intra-HLL Mapping Notes:
13:56 dalek link: http://www.perlfoundation.org/parro​t/index.cgi?intra_hll_mapping_notes
13:58 japhb donaldh++
13:58 dalek donald@sealgair.com | Intra-HLL Mapping Notes:
13:58 dalek link: http://www.perlfoundation.org/parro​t/index.cgi?intra_hll_mapping_notes
14:00 donaldh I suspect that TIMTOWTDI will come into force for HLL interoperability.
14:01 donaldh Because the community will desire all the pros without any of the cons.
14:08 DietCoke joined #parrot
14:08 DietCoke (hll_map) I didn't see a commit, but I saw a comment from tene.
14:09 DietCoke my original test case was malformed in that it wouldn't work even if the original bug was fixed, but there's still an underlying bug.
14:13 DietCoke japhb, tene: you stil lhere?
14:15 paco joined #parrot
14:18 japhb DietCoke: yes
14:19 japhb donaldh: Oh, and shouldn't it be 'Inter-HLL' rather than 'Intra-HLL'?
14:20 donaldh yes, probably.
14:20 purl Really? Probably? Are you Certain it's not certain? Are you sure it's unsure? I think you need to look harder.
14:20 donaldh lol
14:20 donaldh thanks purl.
14:23 donaldh japhb: any other pros and cons?
14:23 dalek r29672 | bernhard++ | trunk:
14:23 dalek : [Pipp PCT] Add test for closures, or anonymous subs, that works with php5.3-200807221230
14:23 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29672
14:26 japhb donaldh: M+N might force the issue of moving more HLL container types into Parrot, when any two languages share a significant semantic or collection type that the current PMCs don't support.  That could be both a pro and a con ....
14:27 donaldh Yesterday I was advocating shared PMC implementations with HLL namespaced methods. This is quite similar.
14:27 donaldh Agreed, it could be a pro and a con.
14:28 japhb M+N might not just dumb things down, but be really slow.
14:28 donaldh Also true.
14:28 donaldh And be forced to copy.
14:28 japhb Yes.
14:29 japhb re: freeze/thaw -- this makes sense to me, but only as a fallback; I don't think it should be the generic method, since it's practically guaranteed to be slow.
14:30 japhb But it sure is nice to know that if all else fails, or there is no existing mapping (for user-generated types, for instance), there's always something that will minimally work.
14:31 DietCoke wow. nothing like subscribing to the smolder failed RSS feed to make you want to TODO failing tests!
14:31 japhb DietCoke: heh
14:31 DietCoke I highly recommend it for someone else so I don't have to.
14:31 * japhb looks pointedly at a cloud
14:31 * donaldh hides behind his coffee.
14:32 DietCoke looks like netbsd is failing one 'atan' test.
14:32 DietCoke http://smolder.plusthree.com/app/public_p​rojects/report_details/2395#first_failure
14:33 japhb For M+N, it might be nice to be able to apply a hint, or even manually run the proper conversion, when the default conversion path loses too much, but another path might work better.
14:36 japhb Actually, from my OpenGL experience, it's clear that it's nice to be able to generate a foreign language's container, and manipulate it within the source language as an opaque object with lots of methods.  The "at worst, everything is an object" design.
14:37 donaldh japhb: re moving HLL PMCs into parrot, both approaches benefit from that.
14:37 japhb PDL, OpenGL::Array, ctypes I'm guessing, and so on
14:37 donaldh Yes. I'm guessing that is what would be required by a PIR implementation of DBI.
14:38 donaldh An opaque PMC that has db drivers behind it and HLL OO methods in front of it.
14:38 japhb donaldh: yes, but whereas M*N can get away with *not* moving the HLL PMCs to Parrot, by just doing the direct conversion from HLL to HLL, the M+N method will probably force the issue.
14:39 japhb donaldh: nodnod
14:40 japhb Speaking of which, It's been a while since I've seen timbunce talking about DBI 2 around here ... though I've missed a few days, so he could have snuck in.  :-)
14:40 donaldh I haven't ever seen timbunce talking about DBI.
14:40 japhb donaldh: Did I get the wrong name?
14:40 * japhb checks ....
14:40 donaldh I was thinking of asking if the project is alive or not.
14:41 japhb Nope, that's right.
14:41 DietCoke no, you got the right name.
14:42 japhb He was here quite a while ago saying he wanted to take everything he'd learned from years of Perl DBI, fold in good stuff from other HLL DBI implementations, and create an uber-DBI for Parrot.  At the time IIRC he was calling it DBI 2.
14:42 donaldh But then I'm in the UK so might not have tz overlap.
14:42 donaldh I googled the subject and found a proposal from 2004 called DBDI.
14:43 Tene DietCoke: no, there was no commit.  I can't get wireless to connect at the hotel.
14:43 Tene DietCoke: also, I'm much more likely to get work done when I stay offline, so I didn't try very hard.  I was reporting from my phone.
14:45 * donaldh hates wikis. They are the victim of browser shortcuts that lose edits.
14:46 japhb donaldh: either he had a newer proposal, or it's been a lot longer than I thought.
14:46 japhb donaldh: s/\. They/ that/
14:46 moritz karma socialtext
14:46 purl socialtext has karma of -6
14:46 moritz socialtext-- # i can confirm it sucks
14:47 japhb Someone forwarded me a job posting recently from socialtext.  I had a bad feeling about it, even though I don't use it myself, so didn't pursue.
14:49 donaldh I use confluence at work. At least it auto saves.
14:49 nopaste "tene" at 63.131.52.132 pasted "fixes coke's problem" (14 lines) at http://nopaste.snit.ch/13631
14:50 Tene DietCoke: ^ is the patch that fixes it.  It passes tests, but I haven't committed because I'm unsure if I also need to patch parrot_fetch_arg_nci() in the same file.
14:50 Tene I'm reading it right now.
14:50 Tene I'm not sure what it does, and the "not yet documented" doesn't help. :)
14:52 Tene Eh, I'll patch it and see if the tests pass.
14:53 Tene japhb: ping
14:53 japhb pong
14:53 Tene japhb: do you know what parrot_fetch_arg_nci() does?
14:53 DietCoke br
14:53 DietCoke b
14:53 DietCoke joined #parrot
14:54 Tene Eh, looks good, I'll commit after tests pass.
14:54 japhb Tene: not off the top of my head, but I'm happy to try being a second pair of (really tired) eyes ...
14:55 Tene japhb: I'll just use grep. :)
14:55 japhb way ahead of ya.  :-)
14:55 dalek donald@sealgair.com | Intra-HLL Mapping Notes:
14:55 dalek link: http://www.perlfoundation.org/parro​t/index.cgi?intra_hll_mapping_notes
14:55 DietCoke tene: that patch seems to fix my issue with Tcl.
14:56 dalek donald@sealgair.com | Intra-HLL Mapping Notes:
14:56 dalek link: http://www.perlfoundation.org/parro​t/index.cgi?intra_hll_mapping_notes
14:56 dalek donald@sealgair.com | Inter-HLL Mapping Notes:
14:56 dalek link: http://www.perlfoundation.org/parro​t/index.cgi?inter_hll_mapping_notes
14:56 dalek donald@sealgair.com | Parrot:
14:56 dalek link: http://www.perlfoundation.o​rg/parrot/index.cgi?parrot
14:57 DietCoke Tene: RT for that is http://rt.perl.org/rt3/Tic​ket/Display.html?id=56958 if you would like to claim and resolve.
15:00 DietCoke tene++
15:00 dalek r29673 | tene++ | trunk:
15:00 dalek : Fetch the proper type from the HLL mapping for ResizablePMCArray
15:00 dalek : when processing slurpy params.
15:00 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29673
15:00 dalek r29674 | coke++ | trunk:
15:00 dalek : [tcl] remove faux-executable on clean
15:00 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29674
15:02 Tene DietCoke: you said your example was flawed?  TCL can work fine now?
15:04 Tene Coke++ # Gives me easy problems to fix.
15:35 * japhb gets done cleaning up multiple projectile vomits from sick kid
15:35 japhb Tene, are you still looking for help, or all done now?
15:36 timbunce joined #parrot
15:37 _timbunce joined #parrot
15:39 Tene japhb: I'm not sure I'm correct about what that function is for, so I can' tbe certain I didn't break something, but all tests pass the same, so I'm not looking for help anymore.
15:40 japhb gotcha
15:40 Tene I just asked you 'cos of all your NCI stuff.
15:42 dalek r29675 | bernhard++ | trunk:
15:42 dalek : [Pipp] Move some comment tests from hello.t into comments.t.
15:42 dalek : Add a TODO test WRT to end of line comments.
15:42 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29675
15:49 dalek r29676 | coke++ | trunk:
15:49 dalek : [tcl] remove workaround for RT#56958, now resolved
15:49 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29676
15:54 dalek r29677 | coke++ | trunk:
15:54 dalek : [tcl] http://code.google.com/p/p​artcl/issues/detail?id=58
15:54 dalek : Rename more __ subs.
15:54 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29677
15:54 dalek r29678 | coke++ | trunk:
15:54 dalek : [tcl] skip more spec tests we can't pass.
15:54 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29678
16:03 japhb _timbunce: Your DBI-on-Parrot project came up earlier today -- is that still in the works?
16:04 dalek r29679 | bernhard++ | trunk:
16:04 dalek : [Pipp] More TODO test cases with three end of line comments
16:04 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29679
16:05 _timbunce Whoever just spoke to me, please repeat. My irc transcript is empty for some reason.
16:05 moritz 18:03 <@japhb> _timbunce: Your DBI-on-Parrot project came up earlier today --  is that still in the works?
16:05 Infinoid _timbunce: [09:03] <@japhb> _timbunce: Your DBI-on-Parrot project came up earlier today -- is that still in the works?
16:05 Infinoid heh.
16:05 _timbunce testing
16:05 moritz he, two echo bots ;-)
16:05 * Infinoid high-fives moritz
16:06 japhb And I didn't even have to repeat myself.  :-)
16:06 timbunce joined #parrot
16:06 japhb welcome back
16:06 purl my mind is.. not going.
16:06 DietCoke Tene: I've already removed the workaround in tcl based on that ticket. Thanks.
16:07 timbunce Thanks japhp. Colloquy is acting odd for me. No transcript and I couldn't post anything.
16:07 timbunce You at OSCON?
16:08 japhb timbunce: I wish ... but no.
16:09 nopaste "barney" at 84.154.57.31 pasted "Confused why three comments are parsed differently from two comments" (82 lines) at http://nopaste.snit.ch/13632
16:11 timbunce japhb: the dbi on parrot project hasn't made much progress lately. My immediate goal is to use GlueGen and the parrot embed api to embed parrot in Java. I got stuck on some OS X JNI linking issues and didn't get back to it.
16:13 japhb timbunce: What stalled the DBI on Parrot project?  Too big a goal all at once?
16:13 DietCoke ... didn't he just answer that? =-)
16:14 timbunce japhb: I got stuck on some OS X JNI linking issues and didn't get back to it.
16:14 donaldh Yet somehow parrot embedded in Java is a more achievable goal.
16:14 DietCoke timbunce: I think jeff horwitz has stolen your killer app thunder, btw., with mod_parrot.
16:14 japhb DietCoke: perhaps I misunderstood.  I thought he said DBI-on-parrot got stuck for reason unstated, and SEPARATELY, Parrot-embedded-in-Java got stuck on JNI problems.
16:15 timbunce DietCoke: I'm not in need of killer app thunder, I've got NYTProf for that :)
16:15 pmichaud barney: in the grammar I'm looking at, a comment can be either:  (1) whitespace, (2) a single line of '//'  (but not more than one), or (3)  a /* ... */ sequence.  It's not possible to combine these together.
16:16 pmichaud s/comment/<?ws>
16:16 japhb timbunce: And that looks damn skippy.  Someone convince Debian and DAG to package it, so I can use it at work ....  ;-)
16:16 pmichaud I suspect it's the "single line of //" that is causing comments to parse funny
16:17 iblechbot joined #parrot
16:17 timbunce japhb: I'm hoping to find someone at OSCON to help with the JNI linker issues. I should also write up a presentation about the strategic value of the approach I'm trying to take.
16:17 donaldh timbunce: is there anything worth sharing from the DBI-on-parrot project?
16:19 timbunce donaldh: not yet. Effectively dbi-on-parrot is blocked waiting on embed-parrot-in-java. Anyone want to help with that?
16:20 pmichaud afk # travel to oscon
16:21 japhb timbunce: why are you tying dbi-on-parrot to embed-parrot-in-java.  Planning to use jdbc as a backend?
16:22 japhb Oh yeah, excellent punctuation, me.
16:23 Andy joined #parrot
16:24 timbunce japhb: JDBC as the API definition, JDBC test suite as the test suite. Plus allow java developers access to all parrot code - which is a big deal given that there must be several orders of magnitude more Java developers than parrot developers right now. If just a tiny percentage cross that bridge to help with parrot it would be great.
16:24 japhb NODNOD
16:25 donaldh hence the "strategic value" comment earler.
16:25 japhb What does JDBC do right that DBI 1.x did not?  (Feel free to point me to a paper or slideware, if this is a FAQ)
16:25 timbunce japhb: trouble is I'm trying to embed one VM I don't know (parrot) into another VM I don't know (Java) on a platform with unusual dynamic linking (OSX).
16:26 japhb Oh, then you should have no problems then!
16:26 donaldh I'd have thought that libparrot.so could be build with no surprising dependencies.
16:27 donaldh I have embedded libperl.so in Java before so may be able to help.
16:27 japhb And as I recall we (Parrot) use a magic linking flag on OS X that papers over a lot of our ... issues.
16:28 donaldh But I probably wouldn't try it on Cygwin (my current platform)
16:28 japhb I think it might have been tetragon who explained some of that to me, but I've forgotten the details
16:28 timbunce japhb: JDBC is mature, feature rich, stable, documented, and has a good test suite. The *DBI-to-DBD* interface has *none* of those.
16:28 DietCoke timbunce: (NYTProf) I saw that. looks verrrry nifty.
16:29 DietCoke I wish I got paid to perl so i'd have more reason to use it. =-)
16:29 particle 25/18.6
16:29 purl 1.34408602150538
16:29 timbunce DietCoke: Thanks. It's been a blast to work on.
16:29 japhb timbunce: ah.  I only have to use the top DBI API, so I see none of the warts, I suppose.
16:30 japhb As soon as I read the posts on NYTProf, my first reaction was: I want *that* for Parrot ....
16:30 barney pmichaud: Thanks for the eyeopener. Somehow there was an implicit '+' on <ws> in my mind.
16:30 * donaldh is just reading timbunce's  NYTProf blog entry.
16:30 japhb It's the DWIM/.
16:33 particle 13/8
16:33 purl 1.625
16:35 dalek r29680 | bernhard++ | trunk:
16:35 dalek : [Pipp] A saner <ws> token. pmichaud++.
16:35 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29680
16:38 particle i can get a processor that's 35% faster, and uses 35% less power that works with my mobo? for a few hundred dollars? sign me up.
16:39 japhb whiteknight's journal entries are suddenly (as in the last couple hours) missing from http://www.brokd.com/psoc08/
16:39 japhb particle: in reference to?
16:39 particle OT
16:40 particle oh, replacing my 1.86GHz 80W clovertown with a 2.5GHz 50W Harpertown
16:41 PerlPilot joined #parrot
16:41 moritz japhb: complain to diment at gmail.com (Kieren Diment), he runs the aggregator
16:42 japhb particle: ah
16:42 MagNET joined #parrot
16:44 paco joined #parrot
16:44 DietCoke joined #parrot
16:44 jonathan joined #parrot
16:44 Psyche^ joined #parrot
16:44 jq- joined #parrot
16:44 wolv joined #parrot
16:44 jan joined #parrot
16:44 dalek joined #parrot
16:44 skv joined #parrot
16:44 schmalbe joined #parrot
16:44 mj41 joined #parrot
16:44 iblechbot joined #parrot
16:44 Patterner joined #parrot
16:44 AndyA joined #parrot
16:44 cognominal joined #parrot
16:44 nopaste joined #parrot
16:44 jq joined #parrot
16:44 Maddingue joined #parrot
16:45 japhb moritz: done, thanks for address
16:46 leo joined #parrot
16:46 mj41 joined #parrot
16:47 dalek joined #parrot
16:47 barney joined #parrot
16:52 d4l3k_ joined #parrot
16:53 Psyche^ joined #parrot
16:53 gmansi joined #parrot
16:53 jan joined #parrot
16:54 jq joined #parrot
16:54 wolverian joined #parrot
16:54 paco joined #parrot
16:54 mj41_ joined #parrot
16:54 mj41 joined #parrot
16:54 skv joined #parrot
16:54 Maddingue joined #parrot
16:55 nopaste joined #parrot
16:59 jonathan joined #parrot
16:59 DietCoke joined #parrot
16:59 AndyA joined #parrot
17:01 timbunce joined #parrot
17:18 Theory joined #parrot
17:19 ruoso joined #parrot
17:21 cognominal joined #parrot
17:24 slavorg joined #parrot
17:24 paco joined #parrot
17:26 Limbic_Region joined #parrot
17:27 Limbic_Region salutations all
17:28 * moritz greets Limbic_Region
17:29 * Limbic_Region has lost his motivation and asks that if anyone finds it that they return it ASAP
17:31 * moritz sold it on ebay already, together with his own motivation
17:31 * Infinoid bought it on ebay and is never giving it back!
17:31 japhb Limbic_Region: Loud music with a strong, fast beat.  It's like absorbing motivation sonically.
17:31 japhb sonicly?
17:31 japhb hmmm
17:32 Infinoid aurally?
17:32 moritz acustically?
17:32 particle Limbic_Region: http://www.richardsimmons.com/
17:32 Limbic_Region japhb - I think I would need one of those fancy directional speakers so that only I could hear it - I live in a cube farm for the government
17:33 moritz they are called "head phones" I believe ;-)
17:33 japhb headset -- with a boom mic, it looks like you're on an important call while your eardrums are being pleasantly assaulted.
17:35 Limbic_Region no, head phones (unless really expensive) bleed sound
17:37 japhb Limbic_Region: how much is your motivation worth, when you figure your hourly rate?
17:37 Infinoid I have a pair of studio cans which totally rock.  people always glance at me in surprise when I start pulling them off and suddenly they can hear how loud my music is.
17:38 japhb (Like the old joke about how expensive a helmet to buy: "How much is your head worth?")
17:38 Limbic_Region japhb - I work for the government so I get paid regardless.  OTOH, I have personal projects I would like to finish - so quite a bit
17:38 japhb there you go -- consider it an investment in your personal happiness.  :-)
17:55 Ivatar joined #parrot
17:56 Andy joined #parrot
17:58 timbunce joined #parrot
18:07 Andy joined #parrot
18:15 peepsalot joined #parrot
18:16 timbunce joined #parrot
18:32 Ron joined #parrot
18:44 DietCoke joined #parrot
18:45 DietCoke Tene: ping.
18:46 nopaste "DietCoke" at 193.200.132.135 pasted "slurpy hll test" (38 lines) at http://nopaste.snit.ch/13633
18:47 DietCoke eh. check your email. =-)
18:54 timbunce joined #parrot
18:58 iblechbot joined #parrot
19:02 dalek r29682 | coke++ | trunk:
19:02 dalek : [HLL] RT#56958 - add a test to avoid regression.
19:02 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29682
19:10 * japhb chuckles at donaldh's cons for using YAML for Inter-HLL mappings
19:20 pac1 joined #parrot
19:20 pac1 left #parrot
19:24 timbunce joined #parrot
19:25 Theory joined #parrot
19:28 Coleoid joined #parrot
19:35 dalek r29683 | coke++ | trunk:
19:35 dalek : [codingstd] Update to version of Perl::Critic that includes an exception we had to add manually before.
19:35 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29683
19:39 dalek r29684 | coke++ | trunk:
19:39 dalek : [HLL] Objects also work fine for .slurpy pmc HLL mappings now.
19:39 dalek : but you have to make sure they're actually defined when you try to do the
19:39 dalek : mapping, so use an :immediate sub to make sure of it. (tene++)
19:39 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29684
19:39 Tene karma tene
19:39 purl tene has karma of 108
19:39 Tene Hmm.
19:40 Tene 2*2*3*3*3
19:40 purl 108
19:40 Tene Nice number.
19:40 moritz 2**2*3**3
19:40 purl 108
19:43 cotto_work If someone with c-fu could update the patch in #46681 to fix the missing } in STRERROR_R_EXCEPTION and simplify that macro to be POSIXy if _GNU_SOURCE is undefined, I'd be much obliged.
19:43 Tene Exactly.
19:44 cotto_work I can do it from home later, but having the change here for testing would shorten the cycle.
19:45 Andy joined #parrot
19:47 cotto_work If somebody's up for that, ping me when it's in RT so I can test it.
19:59 Tene I wonder if we shouldn't throw an exception on an attempt to run .HLL_map on a type ID that hasn't been defined yet.
20:03 Theory joined #parrot
20:36 Theory joined #parrot
20:47 Whiteknight joined #parrot
20:51 jan joined #parrot
20:55 pmurias joined #parrot
21:00 purl joined #parrot
21:25 purl joined #parrot
21:29 cotto_work Whiteknight++ #eating cookies in chromatic's honor
21:32 dalek r29685 | Whiteknight++ | gsoc_pdd09:
21:32 dalek : [gsoc_pdd09] update to trunk r29684
21:32 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29685
22:18 pmichaud ...use an :immediate sub?
22:19 pmichaud could we have a way instead to dynamically control HLL mapings at runtime?
22:19 pmichaud :immediate subs are Evil.
22:20 pmichaud in Rakudo's case, using an :immediate sub won't help or work.
22:20 pmichaud (or if it does, it'll be because we're doing everything in an :immediate sub.)
22:30 dalek r29686 | jkeenan++ | parallel:
22:30 dalek : Refactor steps to pull hard-coded lists of files into the step object's data
22:30 dalek : structure.  Add some basic tests.  Some tests in gen_parrot_include-01.t need
22:30 dalek : fixing yet.
22:30 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29686
22:34 Theory joined #parrot
22:34 Andy joined #parrot
22:39 dalek chromatic@wgz.org | Inter-HLL Mapping Notes:
22:39 dalek link: http://www.perlfoundation.org/parro​t/index.cgi?inter_hll_mapping_notes
22:41 dalek chromatic@wgz.org | Inter-HLL Mapping Notes:
22:41 dalek link: http://www.perlfoundation.org/parro​t/index.cgi?inter_hll_mapping_notes
22:41 Limbic_Region joined #parrot
22:46 kid51 joined #parrot
23:08 Auzon joined #parrot
23:20 Andy coke: Some day I will remember to drop the "s" in URLs.
23:22 Whiteknight "s" is so bourgeois
23:22 Whiteknight The working class doesn't need this letter of oppression
23:24 Infinoid or any form of literacy whatsoever
23:25 Infinoid all that stuff is an impediment to writing syntactically correct lolcode
23:27 Whiteknight IM IN UR ALFABAT KILLIN UR LETTERZ
23:27 Infinoid 5 out of 5 grammatically challenged felines prefer "Z".  Down with "S"!
23:27 Patterner Hooray for the working cla!
23:31 Whiteknight WE R TEH WORKIN CLAZZ
23:35 Infinoid I don't know what's worse... that I've just wasted the last 10 minutes catching up on ICHC, or that it only took me 10 minutes.
23:44 dalek joined #parrot
23:47 cotto_work ichc++
23:47 Whiteknight ichc?
23:47 cotto_work purl--
23:47 purl cotto_work: excuse me?
23:48 cotto_work ichc is is http://icanhascheezburger.com
23:48 cotto_work purl, ichc is is http://icanhascheezburger.com
23:48 purl OK, cotto_work.
23:48 cotto_work ichc?
23:48 cotto_work purl, ichc?
23:48 purl cotto_work: bugger all, i dunno
23:48 cotto_work purl's not getting smarter
23:49 cotto_work purl, ichc is http://icanhascheezburger.com
23:49 purl OK, cotto_work.
23:49 cotto_work purl, ichc?
23:49 purl cotto_work: i haven't a clue
23:52 Infinoid purl--
23:52 purl Infinoid: sorry...

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

Parrot | source cross referenced