Camelia, the Perl 6 bug

IRC log for #parrot, 2009-09-21

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:00 dalek TT #1025 created by coke++: PASM Registers limited to 2 digits.
00:05 Whiteknight joined #parrot
00:05 Austin Hello, Andrew.
00:06 dalek parrot: r41384 | coke++ | trunk/t/compilers/imcc/syn/regressions.t:
00:06 dalek parrot: Migrate ticket from RT -> TT
00:06 dalek parrot: review: https://trac.parrot.org/parrot/changeset/41384/
00:12 dalek TT #1026 created by coke++: Can't use } in TGE transform definitions.
00:12 dalek tracwiki: v4 | coke++ | DebuggerTaskList
00:12 dalek tracwiki: https://trac.parrot.org/parrot/wiki/Deb​uggerTaskList?version=4&action=diff
00:15 dalek parrot: r41385 | coke++ | trunk/src (2 files):
00:15 dalek parrot: Moved this issue into the debugger wishlist.
00:15 dalek parrot: review: https://trac.parrot.org/parrot/changeset/41385/
00:16 dalek parrot: r41386 | coke++ | trunk/src/debug.c:
00:16 dalek parrot: Remove speculative comment.
00:16 dalek parrot: review: https://trac.parrot.org/parrot/changeset/41386/
00:16 dalek tracwiki: v5 | coke++ | DebuggerTaskList
00:16 dalek tracwiki: https://trac.parrot.org/parrot/wiki/Deb​uggerTaskList?version=5&action=diff
00:19 kid51 joined #parrot
00:20 quek joined #parrot
00:23 nathanmccauley joined #parrot
00:25 Whiteknight hello Austin
00:26 Austin How is your weekend so far?
00:29 Whiteknight lousy, and very busy
00:29 Austin Well, I'm sorry about the first.
00:32 dalek TT #1027 created by coke++: PIR-only segfault
00:34 Whiteknight how is your weekend?
00:35 dalek parrot: r41387 | coke++ | trunk/t/compilers/imcc/syn/regressions.t:
00:35 dalek parrot: Add failing TODO test for TT #1027
00:35 dalek parrot: (warning, segf)
00:35 dalek parrot: review: https://trac.parrot.org/parrot/changeset/41387/
00:36 Coke Whiteknight: have a segfault.
00:36 Coke that'll cheer ya up!
00:37 Coke alrightee. down to 231 parrot tickets.
00:37 Coke s/parrot/RT
00:37 Coke if everyone with a CLA moved one ticket over to trac, we'd be under 200.
00:38 Whiteknight Once the pcc_arg_unify branch lands, there are about two dozen tickets that I think we can close as either fixed or obsolete
00:38 Coke ... right after christmas! =-)
00:41 TiMBuS joined #parrot
00:42 Austin And if everyone moved 5, you'd almost be done.
00:42 Austin What's a cla?
00:42 purl it has been said that a cla is Contributor License Agreement or http://www.perlfoundation.org/​contributor_license_agreement or http://www.parrot.org/foundation/legal or http://www.parrot.org/files/parrot_cla.pdf
00:42 Austin Yikes.
00:44 Whiteknight you have to sign your life away and mail in a few nude pictures of yourself that they can use to blackmail you into fixing Coke's segfaults later
00:45 Whiteknight standard operating procedure
00:45 Whiteknight or, at least that's what they told me
00:46 Whiteknight Coke: in TT #1027, are we supposed to be able to retrieve two arguments from an exception handler?
00:47 Whiteknight I've only ever seen .get_results() called with one argument
00:49 dalek TT #1028 created by coke++: can't iterate a dynlexpad
00:51 Whiteknight what do dynlexpads do that lexpads dont?
00:51 Coke Whiteknight: that's old school.
00:52 Coke I wouldn't expect that code to work, I just expect it to not segf.
00:52 Coke dynlexpads don't require you to predeclare with .lex
00:52 Coke (they're dynamic. =-)
00:53 Coke dynlexpads basically let you add variables later. (which is crucial for tcl, especially with things like [upvar])
00:53 Coke (when you are 3 sub calls down the chain from teh original lexpad)
00:54 Coke ok. down to 230 RTs.
00:54 Coke enough cleanup for today. later.
00:54 Coke ->
00:55 Whiteknight I guess I don't understand why we have LexInfo/LexPad then.
00:55 dalek parrot: r41388 | coke++ | trunk/t/dynpmc/dynlexpad.t:
00:55 dalek parrot: Fix TODO test; Migrate ticket from RT to TT
00:55 dalek parrot: review: https://trac.parrot.org/parrot/changeset/41388/
00:58 * Whiteknight goes to bed. Goodnight!
01:02 kyle_l5l joined #parrot
01:02 nopaste "kid51" at 70.85.31.226 pasted "kill_jit branch: preconfig test needs more fixing" (28 lines) at http://nopaste.snit.ch/18011
01:07 Andy joined #parrot
01:10 mikehh Coke - t/compilers/imcc/syn/regressions.t hangs on me
01:34 mikehh I'm too tired to look at it now - bbl
02:00 tokuhirom____ joined #parrot
02:26 mokurai joined #parrot
02:48 janus joined #parrot
02:54 rg1 joined #parrot
03:04 chromatic joined #parrot
03:11 Coke chromatic: hio
03:19 chromatic howdy
03:20 Coke hurm. does 'make -j' work with rakudo/
03:20 chromatic Yes.
03:28 dalek TT #1029 created by coke++: taillcall into PIR compiler broken
03:38 dalek TT #1030 created by coke++: long sub invocation with named parameters
03:43 particle1 joined #parrot
03:43 Andy joined #parrot
03:44 dalek parrot: r41389 | coke++ | trunk/t/compilers/imcc/syn/pcc.t:
03:44 dalek parrot: Add TODO test for TT#1030
03:44 dalek parrot: review: https://trac.parrot.org/parrot/changeset/41389/
03:48 dalek TT #1031 created by coke++: Free()ing of SymReg variables
03:55 dalek parrot: r41390 | coke++ | trunk/compilers/imcc (4 files):
03:55 dalek parrot: Remove speculative comment.
03:55 dalek parrot: review: https://trac.parrot.org/parrot/changeset/41390/
04:00 dalek TT #1032 created by coke++: Document available HLL Mappings
04:04 jrtaylor joined #parrot
04:08 dalek parrot: r41391 | coke++ | trunk/t/op/calling.t:
04:08 dalek parrot: Migrate ticket from RT to TT
04:08 dalek parrot: review: https://trac.parrot.org/parrot/changeset/41391/
04:10 dalek TT #1033 created by coke++: no args checking for a sub without params
04:25 * Coke chuckles at http://search.cpan.org/dist/AnyE​vent-Kanye/lib/AnyEvent/Kanye.pm
04:28 Coke YARGIBTS: doing a bisect is all local, and you don't need to wait forever for the server to download.
04:33 petdance joined #parrot
04:35 chromatic RGIBTS?
04:36 petdance joined #parrot
04:37 Tene chromatic: reason git is better than svn
04:40 Coke that update is STILL running.
04:40 Coke it's been 7 minutes.
04:40 Coke (plus. and I had to restart it after moving a file out of the way)
04:41 Coke now, granted, the initial git clone would have taken a while, but (one time cost)
04:41 Tene and not actually a very significant cost if it was just git.
04:42 Tene It's just the git-svn clone that's absurdly long.
04:42 Coke ah.
04:56 fperrad joined #parrot
04:57 Bzek joined #parrot
04:57 Bzek left #parrot
05:00 bacek_at_work Coke: and you can download prepackaged git-svn repo for parrot. And just update it.
05:01 desertm4x joined #parrot
05:05 Coke (9.944+8.913+8.688)/3
05:05 purl 9.18166666666667
05:05 Coke (9.944+8.913+8.688)/3/6
05:05 purl 1.53027777777778
05:06 Coke 220/8
05:06 purl 27.5
05:07 * Coke closes one of bacek's old tickets.
05:11 * Coke wonders how much time he has lost on the RT->TT conversion.
05:11 Coke s/lost/spent/
05:13 Andy joined #parrot
05:20 dalek TT #1034 created by coke++: attempt to access code outside of current code segment
05:27 Andy joined #parrot
05:29 Coke Andy: ping.
05:29 Andy yessir
05:29 Coke Andy: is http://rt.perl.org/rt3/Tic​ket/Display.html?id=44041 covered sufficiently by 'make cover' ?
05:29 Coke (I know it's not quite the same thing)
05:29 Andy Don't know
05:29 Coke 44041 is a ticket you opened.
05:29 Andy have never looked at macke cover
05:29 Andy yes, I know
05:29 Coke ah
05:30 Andy but I don't know if make cover is sufficient
05:30 Coke are you familiar with the general premise?
05:30 Andy but if you think it is ,then go ahead
05:30 Andy I'm sure I put that there just as a general "this is a good idea"
05:30 Coke I think it is, I'll explain in the ticket.
05:31 Coke cover?
05:31 purl i guess cover is not at link?
05:31 Coke make cover?
05:31 purl hmmm... make cover is running for me, but it fails some tests.
05:31 Coke bah.
05:31 Coke coverage?
05:31 purl hmmm... coverage is http://cv.perl6.cz
05:34 Wolong joined #parrot
05:55 dalek TT #1035 created by coke++: PIR should allow comments between .param lines
05:56 dalek parrot: r41392 | coke++ | trunk/t/compilers/imcc/syn/regressions.t:
05:56 dalek parrot: Migrate ticket from RT to TT.
05:56 dalek parrot: review: https://trac.parrot.org/parrot/changeset/41392/
06:04 uniejo joined #parrot
06:05 Coke chromatic: ping.
06:05 dalek TT #1036 created by coke++: readonly should be shallow
06:05 chromatic pong
06:05 Coke is TT #50894 still an issue?
06:06 chromatic Yes.
06:06 dalek parrot: r41393 | coke++ | trunk/t/pmc/ro.t:
06:06 dalek parrot: Migrate ticket from RT to TT
06:06 dalek parrot: review: https://trac.parrot.org/parrot/changeset/41393/
06:07 chromatic I can put it on my list to test NotFound's patch.
06:08 Coke feel free to steal it from me.
06:08 dalek TT #1037 created by coke++: throw useful exception on non-quoted non-word characters in PGE
06:11 bacek_at_work Coke: which my ticket you've closed?
06:13 Coke ... one of the last 50 or so.
06:13 Coke (MMD slowdowns, as I recall.)
06:13 Coke you complained of 27x slowdowns. it's now only 1.5x slower.
06:13 dalek tracwiki: v36 | cotto++ | ParrotQuotes
06:13 dalek tracwiki: the CLA is serious business
06:13 dalek tracwiki: https://trac.parrot.org/parrot/wiki/Pa​rrotQuotes?version=36&action=diff
06:14 Coke (funny story about nude pics preserved for posterity on the wiki)--
06:17 cotto Coke, feel free to delete it.  Now that I think about it, it may be best left off the wiki.
06:21 bacek_at_work Coke: it wasn't me. I just bisected it. (And "fixed" in TT#452 :)
06:21 Coke I am downvoting. do as you will.
06:21 Coke but left the ticket open, so I'm still blaming you. =-)
06:22 bacek_at_work Hey! It's The Architect's ticket!
06:35 dalek TT #1038 created by coke++: Convert Digest::MD5 to object-based implementation
06:43 dalek parrot: r41394 | chromatic++ | trunk/compilers/imcc/main.c:
06:43 dalek parrot: [IMCC] Fixed runcore flag setting with IMCC optimization options (-t, in
06:43 dalek parrot: specific).
06:43 dalek parrot: review: https://trac.parrot.org/parrot/changeset/41394/
06:44 JimmyZ joined #parrot
06:45 dalek TT #1039 created by coke++: MMD bug in FixedPMCArray.sort
06:47 dukeleto joined #parrot
06:48 dalek TT #1040 created by coke++: inplace math on subclasses failing (MMD/vtable)
06:49 barney joined #parrot
06:49 Coke if someone would like to push us under 200 RTs, you have not much work to do so.
06:49 chromatic There's a weekly goal.
06:49 Coke well, there's only 4 left.
06:50 Coke so I'd aim a little higher than that.
06:50 chromatic Okay.  150.
06:50 Coke doable if everyone would simply resolve or transfer the tickets they already have assigned to them.
06:50 Coke transfer to TT, not to another owner on RT.
06:53 Coke 66 of those tickets were opened by ptc. I bet 80% of those can be rejected after a quick comment deletion.
06:54 chromatic Agreed.
06:54 Coke 66*.8
06:54 purl 52.8
06:54 Coke 53/66
06:54 purl 0.803030303030303
06:56 dalek tracwiki: v37 | cotto++ | ParrotQuotes
06:56 dalek tracwiki: delete previous quote: It seemed like a good idea at the time.  Now it doesn't.
06:56 dalek tracwiki: https://trac.parrot.org/parrot/wiki/Pa​rrotQuotes?version=37&action=diff
07:04 dalek TT #1041 created by coke++: pmc2c silently ignores bad code
07:09 dalek parrot: r41395 | coke++ | trunk/docs/pdds/pdd03_calling_conventions.pod:
07:09 dalek parrot: This already throws a proper exception:
07:09 dalek parrot: too many arguments passed (3) - 1 param expected
07:09 dalek parrot: review: https://trac.parrot.org/parrot/changeset/41395/
07:12 cotto_w0rk joined #parrot
07:14 dalek TT #1042 created by coke++: PGE doesn't support angle quotes inside of regexes.
07:17 Coke bacek: is TT#46687 closable?
07:18 dalek TT #1043 created by coke++: create YAML::Emitter::Syck that does what Data::Dumper does.
07:20 bacek_at_work Coke: nope
07:21 Coke ok. you get to transfer it over to TT, then. =-)
07:22 Coke RT now at 199 tickets.
07:23 Coke msg cotto You own RT #36407, 48439; can you either resolve them or reject them and transfer them to TT?
07:23 purl Message for cotto stored.
07:23 mikehh i still get t/compilers/imcc/syn/regressions.t hanging on me
07:23 Coke mikehh: for me it was a segfault. if it hangs for you, go ahead and skip it for now.
07:24 mikehh make test TEST_JOBS=5 run through all the other tests - then just sits there
07:24 Coke msg whiteknight: You own RT #37578, 38432, 48278, and 59628. can you either resolve them or reject them and transfer them to TT?
07:24 purl Message for whiteknight stored.
07:25 Coke msg kid51 you own 8 RT's - can you either resolve them in place, or reject them and transfer them to TT?
07:25 purl Message for kid51 stored.
07:26 Coke msg rurban - I stole your RT's as I expect you're not going to close them out. just FYI.
07:26 purl Message for rurban stored.
07:26 mikehh if I run it as perl t/harness t/compilers/imcc/syn/regressions.t it PASSes
07:27 Coke mikehh: ok. I'm not going to be able to diagnose it at this hour. =-)
07:28 Coke msg jonathan you own RT #41583 and 46687 - can you either resolve them , or reject them in place and transfer them over to TT?
07:28 purl Message for jonathan stored.
07:28 mikehh Coke: I will have a look after I take my grandsons to school
07:29 Coke msg pmichaud you own RT's #40002, 44979, 48445, 56186 - can you either resolve them, or reject them in RT and transfer them over to TT?
07:29 purl Message for pmichaud stored.
07:29 iblechbot joined #parrot
07:30 bacek_at_work Coke: no way!
07:30 bacek_at_work Coke is magical ticket robot
07:30 bacek_at_work purl: Coke?
07:30 purl hmmm... Coke is Will Coleda <mailto:will@coleda.com> or perpetually annoyed.
07:31 bacek_at_work purl: Coke is also magical ticket robot
07:31 purl okay, bacek_at_work.
07:31 Coke msg allison you own abut 9 RT's:  can you either resolve them, or reject them in RT and transfer them over to TT?
07:31 purl Message for allison stored.
07:31 Coke msg coke go to bed.
07:31 purl Message for coke stored.
07:31 Coke scooby snack
07:31 purl Rank roo, Coke!
07:32 Coke messages erase
07:33 bacek_at_work Coke: TT#1042 looks duplicate of TT#837
07:34 Coke If you think so, go ahead and merge them. Thanks.
07:34 Coke (duplicates in TT are better than one in TT and one in RT)
07:36 bacek_at_work Actually no. They are related anyway.
07:37 chromatic TT #1025 is because IMCC thinks the long register name is a STRINGCONSTANT.
07:39 chromatic Fixing the lexer to allow arbitrary PASM registers... segfaults elsewhere.
07:39 chromatic Guess: direct addressing.
07:39 Coke chromatic: I thought I mentioned that.
07:39 Coke P99 maps to ... P99.
07:39 Coke so if you allow P9999999 ... boom.
07:40 chromatic Ah.
07:40 chromatic I see that in your message now, but I didn't read that into the ticket.
07:40 chromatic either way: sleep!
07:44 Coke chromatic: hokay.
07:44 Coke ... or, it's 4am, I could just stay up
07:44 Coke (email about rts sent)
07:51 Austin Coke, I thought you were on the west coast
08:01 Tene Coke: great job on those tickets. :)
08:03 dalek rakudo: adddb73 | moritz++ | build/PARROT_REVISION:
08:03 dalek rakudo: bump PARROT_REVISION to a post-1.6.0 revision to get some testing
08:03 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/a​dddb73fea2ee04aa239a8bcdd8e5aed6c2443f0
08:03 dalek rakudo: 836c8c8 | moritz++ |  (2 files):
08:03 dalek rakudo: mention spectest_smolder in README and ChangeLog
08:03 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/8​36c8c8012205cccad66b497461b3c769be7315e
08:12 moritz jdv79: regarding your recent patch to add spectest_smolder to rakudo... sometimes when I run it nothing seems to happen... what do you think about making the generation of the .tar.gz and the uploading a single step in the Makefile?
08:15 mikehh Coke: I left that make tst running when I went out - it is parrot is now using 3.1GB
08:15 mikehh test
08:33 bacek joined #parrot
08:34 bacek o hai
08:40 dalek lua: 16f2b05 | fperrad++ |  (2 files):
08:40 dalek lua: add dump_paths which introspects the interpreter (dev only)
08:40 dalek lua: review: http://github.com/fperrad/lua/commit/16​f2b056098451496cc42293ea5600a1453e1dd5
08:40 dalek lua: 0fd29ff | fperrad++ |  (2 files):
08:40 dalek lua: add ../languages/lua/library/?.pbc as first path.
08:40 dalek lua: This allows to :
08:40 dalek lua:  - load languages/lua/library/uuid.pbc which loads library/uuid.pbc
08:40 dalek lua: review: http://github.com/fperrad/lua/commit/0f​d29ff3ec13db4d12d9e52e467231fa90154c14
08:40 dalek lua: 84bc11c | fperrad++ | src/lib/luapackage.pir:
08:40 dalek lua: try .pbc before .lua
08:40 dalek lua: tests run faster because Test.More is already compiled.
08:40 dalek lua: Now, with previous commit, require 'Test.More' loads
08:40 dalek lua:  languages/lua/library/Test/More.pbc
08:40 dalek lua: instead of
08:40 dalek lua:  library/Test/More.pbc
08:40 dalek lua: review: http://github.com/fperrad/lua/commit/84​bc11cb9c23edf5d58deb1f439673201847b3d8
08:41 masak joined #parrot
08:50 payload joined #parrot
08:54 gaz joined #parrot
09:13 Whiteknight joined #parrot
09:14 Austin Hello again, whiteknight
09:14 Austin Feeling rested and ready to rock the week?
09:16 Austin Man, NQP would suck a lot less if it occasionally said "symbol not found" instead of "Null PMC access in invoke()"
09:16 MoC joined #parrot
09:18 Whiteknight hello again, Austin
09:19 TiMBuS joined #parrot
09:24 mokurai left #parrot
09:42 dalek parrot: r41396 | bacek++ | trunk/t/compilers/imcc/syn/regressions.t:
09:42 dalek parrot: [t] Mark regression with SKIP because it hangs on Linux/i386.
09:42 dalek parrot: review: https://trac.parrot.org/parrot/changeset/41396/
09:42 dalek parrot: r41397 | bacek++ | trunk (3 files):
09:42 dalek parrot: [cage][core] Emit additional noop after C<exit> if it's last in sub.
09:42 dalek parrot: Closes TT#1009.
09:42 dalek parrot: review: https://trac.parrot.org/parrot/changeset/41397/
09:44 bacek ==2728== ERROR SUMMARY: 10000000 errors from 120 contexts (suppressed: 41 from 1)
09:45 bacek Yay! TEN MILLIONS DOLL^W ERRORS!
09:45 Patterner Always remember: Error - Bad. Dollars - Good.  €€€ - Best.
10:08 dalek parrot: r41398 | mikehh++ | trunk/t/compilers/imcc/syn/regressions.t:
10:11 dalek parrot: fix skip and add amd64
10:11 dalek parrot: review: https://trac.parrot.org/parrot/changeset/41398/
10:13 bacek And here I declare: exception handling in parrot suck big time...
10:13 bacek sigh...
10:13 mikehh it's all about stack frames and continuations
10:14 bacek Parrot_cx_find_handler_local is just... bad.
10:17 dalek parrot: r41399 | mikehh++ | trunk/compilers/imcc/pcc.c:
10:17 dalek parrot: codetest faillure - at least one space between keyword and subsequent open parenthesis
10:17 dalek parrot: review: https://trac.parrot.org/parrot/changeset/41399/
10:34 dalek parrot: r41400 | fperrad++ | trunk/MANIFEST.generated:
10:34 dalek parrot: [install] add libpaths.pasm
10:34 dalek parrot: review: https://trac.parrot.org/parrot/changeset/41400/
10:37 dalek TT #811 closed by bacek++: [RFC] Deprecate "new Iterator" form for creating Iterators.
10:43 payload joined #parrot
10:44 toyam joined #parrot
10:45 mikehh messages
10:51 MoC joined #parrot
10:52 MoC When I first opened my Email client today and saw that there were 48 messages being downloaded, my first thought was: "oh noes, the spambots got me again.". Then I realised it was Coke++'s moving of the RT tickets.
10:54 moritz Coke++ the ticketbot
10:57 bacek Coke?
10:57 purl i heard Coke was Will Coleda <mailto:will@coleda.com> or perpetually annoyed. or magical ticket robot or (if Diet), something that turns into formaldehyde in my blood.
10:58 bacek "magical ticket bot" :)
11:12 fperrad joined #parrot
11:44 bacek joined #parrot
11:44 iblechbot joined #parrot
11:58 whiteknight joined #parrot
12:07 tetragon joined #parrot
12:16 mikehh All tests PASS (pre/post-config, smoke, nqp_test, fulltest) at r41400 - Ubuntu 9.04 amd64
12:16 mikehh partcl r742 builds on parrot r41400 - make test PASS - Ubuntu 9.04 amd64
12:16 quek joined #parrot
12:25 mikehh rakudo (836c8c8) builds on parrot r41400 - make test / make spectest (up to 28334) PASS - Ubuntu 9.04 amd64
12:35 jdv79 moritz: did you alter the patch at all?  i just noticed it myself.  i'll look into it.
12:35 moritz jdv79: no, I applied as-is
12:36 jdv79 its odd cause i had the changes just local and it was solid for something like 2 weeks
12:38 JimmyZ joined #parrot
12:50 Coke msg austin I am basically in Albany, NY
12:50 purl Message for austin stored.
12:58 dalek TT #1044 created by doughera++: Configure.pl relies too heavily on perl5 ccflags settings.
13:02 AndyA joined #parrot
13:04 Coke doughera++
13:06 jdv79 moritz: :)
13:10 jdv79 moritz: uh, the leading tabs are missing from those 2 targets.  probably my fault.  Could you push that change through:)
13:11 donaldh joined #parrot
13:12 moritz jdv79: yes, will
13:12 moritz do
13:12 jdv79 thanks
13:30 szabgab joined #parrot
13:34 dalek rakudo: 0eaf628 | moritz++ | build/Makefile.in:
13:34 dalek rakudo: unbreak the whitespaces in Makefile.in (spaces -> tabs)
13:34 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/0​eaf628acff7d8b1758f2df8c80d92a2f7e83acc
13:34 dalek TT #1045 created by doughera++: #53494: [BUG] --parrot_is_shared=0 IS shared?
13:45 dalek TT #1046 created by doughera++: Configure.pl allows sizeof(INTVAL) != sizeof(void *), but parrot requires ...
14:00 bluescreen joined #parrot
14:00 Coke is the exec core gone?
14:05 Andy joined #parrot
14:14 jsut|work joined #parrot
14:16 PacoLinux joined #parrot
14:22 jrtaylor Coke, If you mean "has the code been removed?", no. If you mean "Is it working?", I don't know.
14:25 whiteknight the exec core really needs to be removed, I think
14:25 mikehh joined #parrot
14:37 jrtayloriv whiteknight, Why? (just out of curiousity)
14:37 whiteknight As I understand it, it is related to the JIT core
14:37 Psyche^ joined #parrot
14:46 dukeleto joined #parrot
14:53 JimmyZ refactor++
14:54 JimmyZ karma refactor
14:54 purl refactor has karma of 1
15:29 quek left #parrot
15:30 AndyA joined #parrot
15:32 cotto_w0rk git++
15:32 pmichaud good morning, #parrot
15:33 whiteknight good morning pmichaud
15:33 moritz good morning #pm
15:49 mikehh All tests PASS (pre/post-config, smoke, nqp_test, fulltest) at r41400, Ubuntu 9.04 i386
15:52 chromatic joined #parrot
15:55 * Coke wonders if there is an svn-git.
16:00 darbelo joined #parrot
16:07 payload joined #parrot
16:08 theory joined #parrot
16:09 theory joined #parrot
16:11 dukelet0 Coke: svn-git should be possible, but I don't think anybody has been insane enough to write it
16:11 Infinoid There's tailor, for the sufficiently motivated.
16:12 mikehh rakudo (0eaf628) builds on parrot r41400 - make test / make spectest (up to 28337) PASS - Ubuntu 9.04 i386
16:13 Infinoid dukelet0: Got your message.  If you add it to https://trac.parrot.org/parrot/wiki/Languages, dalek will magically start tracking it
16:14 mikehh partcl r742 builds on parrot r41400 - make test PASS - Ubuntu 9.04 i386
16:14 Infinoid (just a link to the project page is all dalek needs)
16:16 iblechbot joined #parrot
16:20 payload1 joined #parrot
16:34 dalek tracwiki: v18 | coke++ | GitObjections
16:34 dalek tracwiki: Feel free to format my comments to fit the style of the page.
16:34 dalek tracwiki: https://trac.parrot.org/parrot/wiki/Gi​tObjections?version=18&amp;action=diff
16:35 pmichaud coke:  if I understand correctly, HLL *development* is supposed to track parrot releases also
16:36 pmichaud (I agree that this doesn't work for most of us, but that's the way I've been led to understand things, on multiple occasions)
16:37 pmichaud in fact, as allison proposed to me for Rakudo Star:
16:37 pmichaud "Rakudo releases after Parrot 2.0 are developed against Parrot 2.0
16:37 pmichaud instead of Parrot trunk/release-of-the-month.
16:38 Coke pmichaud: that's the theory. it just doesn't work.
16:39 pmichaud Coke: I agree entirely that it doesn't work.
16:39 Coke so I'm just noting that it doesn't work, for the record. =-)
16:39 pmichaud fair enough.
16:40 pmichaud however, it bugs me that we'll say "we can't switch to git because HLL's are doing something we don't support".  It seems contradictory to so many of our other policies.
16:40 Coke I have no idea who said that.
16:40 Coke this wiki page is probably full of strawmen on both sides.
16:40 pmichaud that's objection #7, essentially
16:40 Coke (which is why I signed my comments)
16:41 pmichaud also, I'm wondering why there are separate #6 and #7.  TO me they're the same issue.
16:42 Coke slightly different.
16:42 pmichaud yes, I see the difference now.
16:42 pmichaud one is the fact of switching the repository itself, the other is switching the linear nature of tracking trunk.
16:42 Coke for # 7, I'm willing to just say "meh"
16:42 pmichaud same here.
16:42 Coke for # 6, that's just a documentation update for my users.
16:43 Coke (and maybe a small changes to a helper make target or config script.)
16:43 pmichaud I'll add my comments/updates after lunch.
16:43 Coke oh, I remember an oft-heard objection.
16:44 Coke (windows)
16:45 dalek tracwiki: v19 | coke++ | GitObjections
16:45 dalek tracwiki: add 'windows' objection.
16:45 dalek tracwiki: https://trac.parrot.org/parrot/wiki/Gi​tObjections?version=19&amp;action=diff
16:47 Coke rebooting to test tortoisegit...
16:47 jsut|work search adamprime
16:47 jsut|work bloops :o
16:47 szbalint http://en.wikipedia.org/wiki/Git_(software)
16:48 dukelet0 Infinoid: thanks!
16:48 szbalint someone added Parrot to the list of git using software
16:48 szbalint :)
16:48 szbalint (projects)
16:48 jsut joined #parrot
16:49 dalek tracwiki: v99 | dukeleto++ | Languages
16:49 dalek tracwiki: https://trac.parrot.org/parrot/wiki/​Languages?version=99&amp;action=diff
16:50 dukelet0 coke: are you spreading FUD about windows+git ? ;)
16:58 Tene whiteknight: Is there already or can you make a ticket describing the IO refactors that need to happen?
16:59 whiteknight Tene: yes, I've been meaning to do that. Haven't had time
16:59 whiteknight I'll try to get the wiki page updated today
17:00 Tene whiteknight: Thank you.  Can you ping me when you've done it?
17:00 Coke dukelet0: ... I suppose technically, yes.
17:00 whiteknight Tene: sure thing
17:01 dalek tracwiki: v20 | dukeleto++ | GitObjections
17:01 dalek tracwiki: https://trac.parrot.org/parrot/wiki/Gi​tObjections?version=20&amp;action=diff
17:02 Coke dukelet0: have you used tortoisegit?
17:02 dukelet0 Coke: no. I don't have any windows dev boxes.
17:03 dukelet0 Coke: is it decent? i never used tortoiseSVN either
17:03 Coke tortoisesvn has some nice qualities.
17:03 whiteknight I use tortoisesvn at work and I like it
17:04 whiteknight it integrates very nicely with the windows shell, so you right-click on things and select commands ("commit", "update", etc)
17:06 Coke tortoisegit requires msysgit.
17:07 Coke and I can't get to github directly. from this box. will try again from home later.
17:07 Coke tortoisegit?
17:07 Coke tortoisegit is http://code.google.com/p/tortoisegit/
17:09 Eevee joined #parrot
17:09 cotto_work The tricky thing about that page is that it'll be tricky to mark objections as suffuciently addressed without knowing who raised which point.
17:09 cotto_work (referring to GitObjections)
17:10 dalek tracwiki: v21 | cotto++ | GitObjections
17:10 dalek tracwiki: mark objection #6</a> as addressed
17:10 dalek tracwiki: https://trac.parrot.org/parrot/wiki/Gi​tObjections?version=21&amp;action=diff
17:10 dukelet0 cotto_work: we can attempt to color code
17:11 donaldh joined #parrot
17:11 donaldh left #parrot
17:11 cotto_work But what colors will we use?
17:11 cotto_work ;)
17:12 Coke cotto, how are your RT tickets coming? =-)
17:14 dukelet0 cotto_work: we can bikeshed about which colors to paint the brushes that we use to paint the bikeshed
17:15 cotto_work Coke, thanks for the poke.  I'll look at them later today.
17:15 Coke wow, leo is still on the "most opened tickets" list on RT.
17:16 Coke http://rt.perl.org/rt3/NoAuth/parrot/Overview.html for various ways of getting lists of tickets that are 10 or less and therefore make good blocks to work on. =-)
17:24 mj41_ joined #parrot
17:24 Coke 194 RT
17:25 dalek lua: 3867b6a | fperrad++ | src/lib/luapackage.pir:
17:25 dalek lua: Revert "try .pbc before .lua"
17:25 dalek lua: This reverts commit 84bc11cb9c23edf5d58deb1f439673201847b3d8.
17:25 dalek lua: review: http://github.com/fperrad/lua/commit/38​67b6ac4d74739a0b22cebe1584b9d06ea5e693
17:25 dalek lua: 9ac0782 | fperrad++ |  (7 files):
17:25 dalek lua: remove some :init modifiers
17:25 dalek lua: review: http://github.com/fperrad/lua/commit/9a​c078294c19a8fb90be9a4dda1dc197f765d18e
17:25 dalek lua: 497ce1b | fperrad++ |  (4 files):
17:25 dalek lua: try to load .pbc before an existing .lua
17:25 dalek lua: .pbc must be located in the same place than .lua
17:25 dalek lua: review: http://github.com/fperrad/lua/commit/49​7ce1b4a44347a61f2ad1e26d04a2a1df4ce310
17:25 dalek lua: c22b9ed | fperrad++ | test_lex.pir:
17:25 dalek lua: remove languages/lua prefix in some paths
17:25 dalek lua: review: http://github.com/fperrad/lua/commit/c2​2b9ed00fa3d83043975acdfb2b715810eed83f
17:25 dalek lua: cf55054 | fperrad++ | src/ (2 files):
17:25 dalek lua: now generate load_language
17:25 dalek lua: review: http://github.com/fperrad/lua/commit/cf​550545d7f485083e91d2526cad5166b6e24cec
17:25 dalek lua: 174bdc3 | fperrad++ | src/lib/luaaux.pir:
17:25 dalek lua: load_bytecode with an absolute path.
17:25 dalek lua: Avoid conflict between :
17:25 dalek lua:   Test/More.pbc (compiler from Lua)
17:25 dalek lua: and
17:25 dalek lua:   parrot/library/Test/More.pbc
17:25 dalek lua: review: http://github.com/fperrad/lua/commit/17​4bdc37b9194a3a76f720c15cf7c6a2e486a7ac
17:27 ruoso joined #parrot
17:28 dalek TT #1047 created by doughera++: [CAGE] pf_items assumes sizeof(INTVAL) == sizeof(opcode_t)
17:34 zak_ joined #parrot
17:36 cotto_work #ps in 1494
17:37 mberends joined #parrot
17:37 cotto_work you're welcome
17:37 dalek lua: eea640a | fperrad++ | config/makefiles/root.in:
17:37 dalek lua: fix target clean
17:37 dalek lua: review: http://github.com/fperrad/lua/commit/ee​a640a1fce923b49a4435be890b67ddac982e66
17:39 dalek TT #1048 created by doughera++: [TODO] Configure.pl needs a has_function() function.
17:52 * whiteknight probably won't be at #ps tomorrow
17:52 whiteknight I may have to post my report around 5am EST
17:52 dalek TT #1049 created by doughera++: [TODO] replace the value for $(MAKE) with the actual path ...
17:56 Coke doughera++
18:00 dalek TT #1050 created by doughera++: [TODO] [C] Use strerror_r instead of strerror
18:06 japhb joined #parrot
18:09 toyam_ joined #parrot
18:21 joeri joined #parrot
18:45 fperrad joined #parrot
18:49 mberends joined #parrot
18:51 dalek TT #1051 created by pmichaud++: [TODO]  NQP should report use of undeclared variable
18:52 pmichaud okay, that closes out all of the RT/parrot tickets I owned.  :)
19:02 * dukelet0 stands back and watches the RTpocalypse
19:03 dalek TT #1052 created by pmichaud++: [TODO] Add --target=pbc to HLLCompiler
19:18 kjeldahl joined #parrot
19:22 dalek rakudo: 798ea1c | pmichaud++ | docs/spectest-progress.csv:
19:22 dalek rakudo: spectest-progress.csv update: 436 files, 15498 (71.4% of 21695) pass, 0 fail
19:22 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/7​98ea1c449ad762e08bc3f9481b1d621a8d3c0d3
19:38 Zak joined #parrot
19:42 Coke RT: 184
19:47 bacek joined #parrot
19:53 Tene So, can anyone tell me what the purpose/policy is for the examples/ directory?  It seems like there's some useful stuff in there that doesn't get installed at all.  Can I just move it to runtime/library/ ?
19:54 Coke if it's actually usable as a library, I don't see a problem with that, except: then we have to support it. =-)
19:55 Tene SQLite3 was the earlier example.
19:55 Tene Mysql is the one I'm currently looking at.
19:55 Tene I guess this should really go in the "extra libs" thing that japhb was working on, whatever we ended up naming it...
19:56 NotFound Tene: I asked at #ps months ago, rejected because it was not enough mature nor complete... and still isn't.
19:57 Coke tene (extra libs) +1 here
19:57 Tene NotFound: so what's the plan for libraries like this?
19:57 Tene There was that big thread about aviary, and then japhb did some work, but didn't finish what he was working on yet, I think...
19:58 NotFound Tene: the plan is to use it, yes.
19:59 NotFound And move out of core some non core libs, like sqlite
19:59 Tene So if I want to make SQL libraries available, I should work on figuring out and helping to complete the aviary thing.
20:00 Coke last I heard on aviary, japhb got sucked into the #toolchain discussion.
20:01 darbelo You make it sound like he got hit by a bus.
20:01 NotFound For simple modules like MySql, maybe a short pir or nqp utility can be enough. Take the pir file, put it in the library directory and compile it to pbc.
20:02 geoff joined #parrot
20:03 geoff left #parrot
20:07 Tene purl: msg japhb Can I get an update on the aviary and toolchain stuff, and what I can help with?
20:07 purl Message for japhb stored.
20:09 mberends Tene for aviary do you mean http://gitorious.org/parrot-plumage ?
20:10 Tene mberends: Maybe.  I havne't been paying as much attention as I should have.
20:11 Tene That looks right, though.
20:13 Tene ... eww
20:13 Tene I run 'make' in that directory, and it tries to run: /home/sweeks/git/rakudo/parrot_install/bin/parrot
20:22 NotFound Tene: is a prototype. looks like you need to edit the Makefile manually to configure it.
20:22 bacek joined #parrot
20:22 japhb joined #parrot
20:23 nopaste "tene" at 97.117.70.208 pasted "I'm not so good with make... I assume this is unacceptable for some reason?" (16 lines) at http://nopaste.snit.ch/18017
20:23 Tene hi japhb
20:24 Tene japhb: I just pasted a patch for plumage's Makefile.  Can you tell me why that won't work?
20:26 darbelo Tene: that doesn't work if parrot_config isn't reachable in $PATH
20:26 darbelo And i'm not sure `` works on windows.
20:26 Tene Ah.
20:27 darbelo The standard solution seems to be /tools/dev/gen_makefile.pl
20:29 NotFound darbelo: And that is supposed to be in PATH?
20:29 Tene hehe
20:31 bacek Good morning
20:31 purl And good moroning to you, bacek.
20:31 darbelo That let's you reference parrot config itms in your makefile automagically, without caring about paths.
20:31 Tene I guess make doesn't have a way of doing that, so we need a configure script. :P
20:31 japhb *rez*
20:31 japhb reading back
20:31 darbelo Tene: Yes, that is what I was implying with gen_makefile.pl
20:32 Tene darbelo: but how can you use gen_makefile.pl without first locating the parrot directories?
20:32 * japhb is now caught up ... I think
20:33 darbelo You ask nicely to be told where they are?
20:34 japhb Tene: I had planned to do a Plumage hack day yesterday, but a $day_job deadline intervened
20:34 japhb General state of Plumage (higher level than in TODO):
20:35 kyle_l5l joined #parrot
20:35 japhb * Getting to the actual hacking.  Trying to make it go through a fetch-through-install cycle for Blizkost.
20:36 japhb * Lots of stuff is hackish right now, because I have been getting bogged down with problems, and don't want to delay the project any further to try to get "perfect" when instead I could get it "working".
20:36 japhb * Help de-hackifying greatly appreciated.
20:36 dukelet0 japhb: what is the relation between plumage and blizkost ?
20:37 darbelo Tene: right now I'm rewriting the decnum-dynpmcs Configure in PIR, I get the various parrot paths with 'load_bytecode "config.pbc"'
20:37 Coke (getting info from parrot config) check out partcl or rakudo's Configure.pl's parrot-config options.
20:38 japhb dukelet0: Blizkost is: a) cool, b) potentially useful, c) requires non-trivial work to fetch, build, and install, d) has external dependencies, so I can think about that problem with a real task in front of me, rather than "in theory"
20:38 dalek TT #1053 created by bacek++: [TODO] Allow Parrot Hashes to have PMC keys
20:38 dukelet0 japhb: i hack on blizkost, so I know that it is cool ;) Let me know if I can be of service
20:38 shockwave joined #parrot
20:38 japhb dukelet0, will do, thank you.
20:39 Tene japhb: would it be feasible to get SQLite3.pir and Mysql.pir somewhere where plumage could download and install them to the right place?
20:39 shockwave At docs.parrot.org, some paragraphs refer the user to see "table x-y". Where are these tables?
20:39 shockwave For instance, this page: http://docs.parrot.org/parrot/latest/ht​ml/docs/book/pir/ch07_objects.pod.html
20:39 shockwave at the very bottom; the last line.
20:40 japhb Tene: yes, that would be an excellent idea.  And trying to create the correct metadata JSON for those would help me think about places the metadata spec is lacking.
20:40 Tene japhb: Great.  I'd love to help you with this.
20:40 japhb Tene, Also, my makefile skills are limited to "just slightly more than cargo culting".  Brains applied there greatly appreciated.
20:40 Tene OK. :)
20:41 japhb Oh and ... writing a configure script for it, in NQP?  Pretty much rockin' help right there.
20:41 japhb ;-)
20:41 NotFound Please don't put MySql.pir in a git repository
20:41 japhb NotFound, I was thinking RCS ... ;-P
20:42 japhb NotFound, but seriously, no fear.  Plumage *must* be able to deal with a lot of different ways to fetch the code.  From tarball to git, and everything in between.
20:42 Tene NotFound: as it is, in examples/, it's not installed, can't be used, etc.  Your'e objecting to me putting a copy somewhere where people can start using it?
20:42 Tene NotFound: or are you saying something about git specifically?
20:42 Tene Or objecting to it being out of parrot's repo at all?
20:43 NotFound Tene: I just like to have it in a place I can work easily.
20:43 japhb I think the "out of parrot's repo" ship has just about left port.  Allison seems adamant about stripping the libs to the metal.  Just enough to bootstrap plumage, actually.
20:44 pmichaud I agree with that, fwiw
20:44 NotFound japhb: And just in a local directory?
20:45 japhb NotFound, you mean, as a place to fetch the code from?  Sure, why not?  It's just another url scheme.
20:45 Tene NotFound: So you're just saying that wherever I put it, you don't really care, just please don't use git?
20:45 NotFound Tene: yeah
20:45 * japhb chuckles
20:46 Tene NotFound: Okay.  The main Parrot repo isn't really an option, afaict, so where would you like to host DB libraries for Parrot?
20:46 Tene Do you expect to care to work on any db libs besides mysql?
20:46 mberends japhb: please remind me where you blog about Plumage
20:46 Tene There's a blog, too?
20:47 japhb u.p.o.  Except the journals fracking bit it last week.
20:47 japhb But there were a number of them, yes.
20:47 Tene japhb: is there a reason that your blog isn't on planet.parrotcode.org ?
20:47 Coke shockwave: in the raw PIR, I see:
20:47 Coke Table 7-1 shows the introspection characteristics supported by
20:47 Coke C<inspect>.
20:47 Coke =begin table Class Introspection
20:47 Tene japhb: upo username?
20:48 japhb I asked for it to be added in here, but unfortunately I think whoever I talked to (I forget who) forgot about it.  Ah, Irony.
20:48 Coke ... so I'm guessing that the pod2html translator has lost the tables.
20:48 japhb Tene: geoffrey
20:48 Coke if you could open a ticket at https://trac.parrot.org/ about that, I'd appreciate it.
20:48 NotFound Tene: I wrote the MySql because I have some familiarity with MySql API, and like to keep working on it just because I started. Other than that, no special interest in DB.
20:48 japhb .oO( I think my afternoon just got hijacked for a good cause .... )
20:48 Coke japhb: you were added.
20:49 Coke it just never worked, as i recall.
20:49 Tene NotFound: OK. Please tell me where I can put a repository such that you'll continue to work on it?
20:49 Coke please piggyback on dukeleto's trac ticket.
20:49 * japhb wonders why one's own thoughts actually often start with "I think", as redundant as that is ...
20:49 Coke (or open a new one)
20:49 japhb Coke, Ah.  Sigh.
20:49 Coke japhb: you need to self-actualize.
20:49 Tene japhb: disclaiming certainty.
20:49 Coke ->
20:49 japhb heh
20:49 * mberends thinks it makes sense
20:50 rdice joined #parrot
20:50 NotFound Tene: the quick solution may be svn.parrot.org/modules or something like that.
20:50 japhb Anyone know A) a way to get my post data out of u.p.o right now, and B) a Really Fine Place to put it?
20:51 japhb I don't really have the spare cycles to write a blogging engine and set it up somewhere, which is what I would have done 10 years ago ....
20:52 Tene Isn't there some blogging thing associated with parrot.org?
20:52 japhb I know there's a planet (clearly), but I didn't know of an actual place to post.
20:53 Tene Do I have privileges to make svn.parrot.org/modules/ ?
20:54 NotFound Tene: I don't know
20:54 Tene :)
20:55 japhb Oh, Tene: A quick hybrid of our ideas for the makefile would be useful.  I think I'll still define PARROT_BIN manually, but from there find parrot_config, and find the rest of the locations from that, if I can.  That at least reduces the surface area of the config problem.
20:55 * japhb goes to hack that in ....
20:55 japhb oh wait, someone pointed out `` is not portable.
20:55 japhb Damn.  Doesn't make have something portable for this?
20:56 japhb Some interpolation hell?  (Why does every tool have to define it's own interpolation syntax?!)
20:56 Tene japhb: I don't think it does... hence the persistent need for configure scripts.
20:57 darbelo japhb: You need a configure script to take care of that for you.
20:58 japhb Sure, to find the PARROT_BIN.  But really, there's no `` equivalent?  That's ... well, dumb.
20:58 Tene Yeah.
20:58 japhb Not to put too fine a point on it.
20:59 darbelo Dumb like the tabs and spaces thing? Yeah, sounds like make.
20:59 * japhb cringes
21:03 dukelet0 Tene: yes, if you have an account on parrot.org, you should be able make posts
21:03 Tene Looks like nqp doesn't have an 'open' function...
21:03 japhb Does parrot require a particular make (or a certain limited set) to build?
21:03 Tene dukelet0: japhb is looking for a place to host his blog posts.
21:03 Tene japhb: looks like NQP doesn't have 'open' and friends... inline PIR!
21:04 japhb Tene, Any time I need a "builtin", I've just been copying the relevant code from Rakudo and editing to fit, or writing my own with raw PIR ... in either case, that's what Glue.pir is for.
21:04 Tene nodnod
21:05 japhb There are a couple places I tried to clean things up a bit since the original  code was clearly hacked on a bit, and a couple other places where I drop the niceties.
21:05 Tene japhb: are we avoiding depending on Perl 5?
21:05 dukelet0 Tene: i am sure there are other options, but I can give japhb a login to post to http://leto.net/perl/ or make another blog if he wants a whole blog to himself
21:06 Tene dukelet0: perhaps you should direct that to japhb. ;)
21:06 dukelet0 Tene: indeed
21:07 japhb dukelet0: Nobody here should have to do any work for this.  I had only asked if someone had a recommendation for a good, solid "actually stays up" blogging host.  (And if anyone knew of a way to recover the u.p.o posts, or if they were just gone.)
21:07 japhb Goodness knows all the people here have enough tasks to deal with, without me adding to the pile.  :-)
21:07 dukelet0 japhb: i run my own instance of movable type that can host any number of blogs. It mostly stays out of my way.
21:09 japhb Well, if it's pretty reliably up all the time, then sure, make me a fresh new one, and I'll start there next time I do a Plumage post.
21:12 darbelo japhb: Parrot builds fine on most makes, but it uses makfile templates that the gen_makefile turns into full makfiles with data gathered in the configure steps.
21:13 japhb pmichaud, I vaguely recall you and Allison disagreeing about whether gen_makefile was kosher to rely on ... whatever happened with that discussion?
21:14 darbelo It's installed with install-dev, and parrot itself depends on it.
21:14 pmichaud I decided that Rakudo didn't want to rely on gen_makefile, so it doesn't.
21:14 pmichaud we'll do our own makefile generation
21:15 pmichaud in particular, I'm not sure that I'd trust that Parrot's gen_makefile api won't change :)
21:15 japhb ah
21:15 darbelo pmichaud: What do you use for makefile generation?
21:16 pmichaud darbelo: we do it in our Configure.pl.... but it's much simpler than Parrot's generation
21:16 darbelo So, perl 5?
21:16 purl rumour has it So, perl 5 is fucked
21:16 pmichaud yes, perl 5
21:16 pmichaud it might be entirely possible for me to rewrite it in nqp, though :)
21:17 japhb pmichaud, I think I may end up effectively prototyping that for you.
21:18 * japhb reading the GNU Make manual and just getting the heebie-jeebies.  ("Seriously, you guys are looking *down* on other make systems?")
21:20 darbelo japhb: "nmake" and that's all I'm sayin'
21:21 japhb oooh!  Oooh!  There's a 'shell' function:  contents := $(shell cat foo)
21:21 japhb Anyone who speaks nmake, gmake, et al.  -- is that function supported?
21:22 NotFound Oh, nice. mysql library gets confused by parrot list_delete function.
21:22 darbelo japhb: Portably, I seriously doubt it.
21:23 * japhb decides that Google, may, in fact, be his friend today.
21:24 shockwave @Coke, thanks for the info about the table issue. I can't create a ticket because I don't have a trac login.
21:25 darbelo make was created to manage dependencies in the "Oh, you touched this, you'll need to rebuild that as well" way. Everything else is non-portable, a horrible hack or both.
21:26 japhb OK, BSD make has an equivalent concept, using the != assignment operator
21:26 pmichaud anyway, the issue I had with parrot's gen_makefile is that the code is _really_ hard to understand and modify
21:26 pmichaud also, I can imagine that someone will someday need to do something in makefiles that parrot's gen_makefile doesn't quite support
21:26 japhb nodnod
21:27 pmichaud so I wanted a simpler version that can generate makefiles without having all of those parrot dependencies
21:27 Tene japhb: you could also have a generator that loaded config.pbc itself.
21:28 pmichaud (and that someone could understand relatively quickly that didn't require knowledge of parrot's entire makefile config subsystem, which is pretty complex)
21:28 japhb Tene: Plumage does exactly that.  But then we have a bootstrapping issue.
21:28 japhb pmichaud, makes sense
21:28 Tene japhb: make a .pir and assume that the user knows how to invoke parrot on it?
21:28 Tene and the .pir spits out a makefile.
21:29 Tene or we go back to the parrot --whatevermagic=nqp configure.nqp
21:29 Tene which is NYI and NYS etc.
21:29 pmichaud I'm still planning ot make nqp into a fakecutable and installing it as parrot_nqp
21:30 pmichaud then it would just be   parrot_nqp configure.nqp
21:30 japhb pmichaud, yes.  Please.
21:31 mberends japhb: found http://use.perl.org/~geoffrey/journal/39622
21:31 japhb mberends, Oooh.  I CAN HAZ UPO NOW?
21:32 NotFound What must be the subsystem prefix for list.c functions? Parrot_lst ?
21:32 pmichaud anyway, Rakudo's makefile generation is Configure.pl lines 164-186
21:33 * japhb pulls a current Rakudo
21:34 pmichaud oh, one can also look at the source directly on github :-)
21:34 pmichaud github.com/rakudo/rakudo
21:34 pmichaud github.com/rakudo/rakudo/tree/master/Configure.pl
21:34 japhb yes, but in a couple minutes I will have a current perl6 binary.  This is *so* worth the wait.  ;-)
21:35 dalek TT #1054 created by shockwave++: Some documents at docs.parrot.org are missing data.
21:35 shockwave @Coke, I didn't realize registration for trac was open.
21:46 japhb pmichaud, OK, I call shenanicgans on lines 174-175.  What is going on there?
21:47 japhb Or rather, why are you doing it (and doing it that way)?
21:48 japhb Why would an http: URI contain any backslashes at all?
21:48 pmichaud because the previous line converted allof the slashes into backslashes
21:48 pmichaud fwiw, this is the way that Parrot's gen_makefile handles the situation as well :)
21:48 pmichaud i.e., we stole that logic from Parrot :)
21:49 japhb oh.  duh.
21:51 japhb OK, as soon as I add music to this office, I'll hack up an NQP of create_makefile() for Plumage, and mayhaps some of it will be worth using for Rakudo.
21:52 japhb Anyone know offhand where to get the running parrot and/or fakecutable's path from PIR?
21:53 darbelo japhb: Put it somwhere I can steal it for decnum-dynpmcs.
21:54 Tene japhb: ask it to load config.pbc
21:54 darbelo load_bytecode "config.pbc"; conf = _config(); conf['bindir']; would be my guess.
21:54 japhb .oO( Ah, headphones are *so* much better than the tinny speakers in the wife's laptop .... )
21:55 japhb darbelo, no problem
21:56 darbelo japhb: another thing, you'll want to use @cp@, @mv@, @rm@, @rm_r@ and friends in your makefile, that way it will expand into whatever the current platform uses for copying, moving, deleting, etc without you having to think about it.
21:57 japhb nod, nod.
21:57 japhb In fact, I'm just going to combine your advice with Tene's and directly use the Parrot config items, if those are listed therein.
21:58 pmichaud fwiw, Rakudo will probably stick with the Perl 5 version of its build tools
21:58 pmichaud We don't carry the "get rid of Perl 5 at all costs" sentiment that Parrot sometimes seems to exhibit
21:59 darbelo japhb: Exactly. Under the current system @whatever@ get's replaced with conf['whatever']
22:00 japhb pmichaud, Plumage is a prisoner of its purpose.  :-)
22:01 Coke shockwave++ #more tickets.
22:01 darbelo FWIW, I don't mind that parrot needs perl 5 to bootstrap itself, but I would prefer that my parrot-based project only depend on parrot.
22:02 pmichaud I'll be very happy if nqp can fill that niche.  Indeed, it's somewhat intended to do exactly that :)
22:02 pmichaud (and I'm working to increase that intent )
22:09 darbelo left #parrot
22:12 japhb pmichaud, excellent.
22:13 darbelo joined #parrot
22:19 dalek parrot: r41401 | NotFound++ | trunk/examples/nci/Mysql.pir:
22:19 dalek parrot: [examples] minimal fix in MySql;Statement prepare - Still not working because of symbol name conflicts between parrot and mysql libs
22:19 dalek parrot: review: https://trac.parrot.org/parrot/changeset/41401/
22:29 japhb pmichaud, it looks like I can modify Rakudo's regex subst implementation (line 723 and following in src/builtins/any-str.pir) for my purposes.  Where is the canonical code for instantiating the regex in the first place?  src/parser/quote_expression.pir seems a bit ... self-referential ... on this topic.
22:30 pmichaud regexes are just quoted strings
22:30 pmichaud if you're planning to pull Rakudo's regex parsing into NQP, I should warn that all of it is about to be tossed aside in the next two weeks
22:32 japhb pmichaud, no, I just wanted to know how I could instantiate the *Regex PMC needed to do subst(string: sub, string)
22:33 pmichaud I don't understan *Regex PMC
22:33 japhb Figured there might be some special stuff I had to do to set it up.  I was looking at the Rakudo guts for inspiration
22:33 pmichaud are you just asking how you copmile a regex?
22:33 japhb pmichaud, the thing you can do .'!invoke'(string) on.
22:33 japhb Yes, I think os.
22:33 japhb er so
22:34 pmichaud from PIR:
22:34 pmichaud $P0 = compreg 'PGE::Perl6Regex'
22:34 pmichaud $P1 = $P0.compile('regex');   # I think
22:35 pmichaud then $P1 is a sub that you can invoke on a string to be matched
22:35 kid51 joined #parrot
22:35 japhb Ah!  That's way less of a pain than I expected.
22:35 pmichaud I'm not sure if it's $P0.'compile'('regex')  or just $P0('regex')
22:35 pmichaud eventually it'll be the former, but I don't know if we've made that switch yet.
22:35 pmichaud (given that PGE comes from before we had compiler objects)
22:36 japhb OK.
22:36 darbelo japhb: I think you might have some competition. http://github.com/flh/genfile/
22:36 japhb Yeah, that's why I expected craziness.
22:36 pmichaud well, the craziness is more "simplification"
22:36 pmichaud i.e., PGE provides a less sophisticated interface
22:37 japhb nodnod.  (I didn't mean that in a negative sense, actually.  Just meant that I expected to have to do a few contortions to get the sub I wanted.)
22:38 pmichaud right, I didn't take it negative
22:38 japhb darbelo: from a day ago?  *chuckle*
22:38 japhb pmichaud, ah, good.
22:39 * darbelo just saw it in the backlog.
22:40 * japhb chuckles that http://github.com/flh/genfile/blob/master/Makefile does a bunch of $(shell $(PARROT_CONFIG) ...) calls.
22:40 japhb I think I still have a purpose.  :-)
22:40 darbelo The thing starts at http://irclog.perlgeek.de/p​arrot/2009-09-20#i_1518892
22:41 japhb I think what I'm writing is what he would use to bootstrap his code.
22:43 darbelo I haven't really looked at the code, I was just noticing the similarity of intent.
22:45 payload joined #parrot
22:51 patspam joined #parrot
22:59 darbelo NotFound: ping
22:59 Austin Speaking of configuration data... How do I access configuration stuff from withing an installed parrot?
22:59 darbelo Austin: in PIR?
22:59 Austin Darbelo, yes, pir or nqp
23:00 Austin I want to get things like the installed directory, where my language files are, etc.
23:00 darbelo PIR: load_bytecode "config.pbc"; conf = _config(); conf['whatever'];
23:00 Austin Okay. Is that info the same as %PConfig in perl?
23:00 darbelo where conf is a string.
23:01 darbelo Yes, and the same that the parrot_config binary outputs.
23:01 Austin Perfect. And is there some mechanism for adding / installing to it?
23:02 Austin In other words, the "load_language" opcode pretends that it knows about the special language-specific dirs. Is it just using config['lib_dir']/$language_name, or is there something more sinister at work?
23:03 darbelo 'tis all pretend.
23:03 Austin :(
23:04 darbelo I think.
23:04 Austin It's cruel to lie to children...
23:05 Austin FWIW, japhb, it was kind of interesting to see what flh was willing to do, in terms of manhandling the innards of the compiler, to get his stuff set up. Sub-classing grammars, a new compiler phase, etc.
23:05 japhb What's the reason for the root_new opcode?  Is it a semantic disambiguation, or a performance optimization, or ...?
23:05 Austin Root_new because otherwise crossing hll's is too hard.
23:06 japhb "too hard"?
23:06 darbelo A workarround for some bug people were relying on. IIRC
23:06 japhb darbelo, oh.  ew.
23:06 japhb I guess I'll use it then.
23:06 Austin Too many instructions, too much checking.
23:06 japhb Austin, that sounds like an optimization
23:07 Austin Sure. Just like the PPSU opcodes, and a host of others.
23:08 Austin But more like "supporting a common case" than prematurely optimizing. A lot of HLLs will want to use Parrot stuff, which is an xhll operation - root_new.
23:09 dalek TT #1055 created by jkeenan++: t/pmc/complex.t:  sinh_of_complex_numbers:  some tests fail on Win32
23:09 japhb Oh, I didn't claim it was premature.  Just trying to understand its *purpose*.
23:10 Austin I think it falls out of the problem where PMCs are globally visible - that is, {{{   $P0 = new 'Hash'   }}} always creates a parrot::Hash
23:11 Austin Which meant that PMC names weren't namespace-ified, so there could be only one 'Hash', etc.
23:11 Austin (PerlHash, TclInteger, yech)
23:11 Tene but if I have a Hash class in the current non-parrot HLL...
23:11 Tene exactly
23:11 Austin So now it's {{{ $P0 = root_new [ 'parrot' ; 'Hash' ] }}}
23:12 darbelo msg NotFound Tried your patch for TT#1009, it still segfaults.
23:12 purl Message for notfound stored.
23:12 japhb ah so
23:12 Austin pmichaud added it back around YAPC time-frame
23:15 Austin Man, I wish I was smarter.
23:15 ruoso joined #parrot
23:18 bacek_at_work darbelo: (TT#1009) is it segfaulting in trunk?
23:18 mikehh joined #parrot
23:19 darbelo bacek_at_work: nope, works like a charm.
23:20 darbelo Just closed the ticket too.
23:21 dalek TT #1009 closed by darbelo++: examples/shootout/spectralnorm.pir shows a segfault in the continuation ...
23:21 darbelo There, see?
23:22 bacek_at_work darbelo: :)
23:22 darbelo bacek++
23:23 darbelo Thank to you I might just see what a succesful make fulltest looks like :)
23:23 bacek_at_work darbelo: Yay!
23:24 Austin Wow! I just discovered something new about NotePad++ - draggable separate windows.
23:24 bacek_at_work But NotFound's solution was better. I'll try to check why it doesn't work.
23:24 cotto_work karma notepad
23:24 purl notepad has karma of 32
23:25 cotto_work karma notepad++
23:25 purl notepad++ has karma of 2
23:25 bacek_at_work Austin: there is only 2 editors :)
23:25 Austin Bacek: There is only one editor. All others are impostors, or maybe just notepad programs.
23:26 darbelo bacek_at_work: Segfault is the same with NotFound's patch. Which it shouldn't be, but it is.
23:27 jsut_ joined #parrot
23:28 dalek TT #1056 created by jkeenan++: inter::progs ought to be smarter about finding linker
23:28 dalek TT #1009 reopened by chromatic++: examples/shootout/spectralnorm.pir shows a segfault in the continuation ...
23:56 quek joined #parrot

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

Parrot | source cross referenced