Camelia, the Perl 6 bug

IRC log for #parrot, 2009-01-24

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 Hunger joined #parrot
00:08 kj joined #parrot
00:08 AndyA joined #parrot
00:09 Util joined #parrot
00:09 NotFound joined #parrot
00:17 dalek r35933 | cotto++ | trunk/config/gen/makefiles:
00:17 dalek : [languages] add up-x targets to update externally hosted languages
00:17 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35933
00:23 alvar joined #parrot
00:28 cotto That'll work fine until we have a Parrot-based language called "yours".
00:29 rg ?
00:40 Infinoid heh.
00:46 cognominal > sub a(Int $a) { say $a };  a("1")
00:46 cognominal Parameter type check failed for $a in call to a
00:46 cognominal > sub a(Int $a) { say $a };  a(+"1")
00:46 cognominal Parameter type check failed for $a in call to a
00:46 cognominal >
00:46 cognominal is this what is expected?
00:47 cognominal I can understand the first even if it surprise me but the second looks very suspiscious to me.
00:47 tetragon joined #parrot
00:49 Infinoid rakudo: say (+"1").WHAT
00:49 polyglotbot OUTPUT[Num␤]
00:50 Infinoid leaving aside whether numification should return an Int or a Num, I think some sort of coercion should be possible.  But I have no idea how that stuff works.
00:51 * cognominal forgot what a Num is
00:51 cognominal a floating point?
00:51 purl a floating point is scary, no matter how you look at it. Among other things, addition isn't associative in flaoting point arithmetic. So (x + y) + z is not necessarily equal to x + (y + z).
00:56 skv joined #parrot
01:02 Coke_afk purl, scooby snack.
01:02 purl Rank roo, Coke_afk!
01:04 particle joined #parrot
01:10 Coke hurm. "../../parrot tcl.pbc foo.tcl" : bus error. gdb ../../parrot ; run tcl.pbc foo.tcl : argument mismatch error.
01:10 skv joined #parrot
01:13 Tene nice
01:13 cotto Coke, ping
01:14 skv joined #parrot
01:14 Themeruta joined #parrot
01:15 Coke cotto: pong
01:16 cotto Does t/cmd_lsort have one failing test for you?
01:16 Coke no. all tests pass.
01:16 Coke but the harness complains.
01:16 Coke http://rt.perl.org/rt3//Publ​ic/Bug/Display.html?id=57088
01:17 cotto ok.  That's what I see too.
01:17 Coke want tcl commit bit?
01:18 Coke (if so, just need googlecode id)
01:18 Coke cotto++ for whatever is making him check tcl.
01:18 Coke so what do I do if a bus error goes away when run in gdb?
01:19 dalek r35934 | cotto++ | trunk (4 files):
01:19 dalek : [pmc] make ATTRs implicitly be passed down the inheritance chain
01:19 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35934
01:19 nopaste "cotto" at 96.26.202.243 pasted "patch to fix tcl for ATTR update" (49 lines) at http://nopaste.snit.ch/15410
01:19 Whiteknight joined #parrot
01:19 Coke cotto: ooc, why move set_number_native?
01:20 cotto Coke, that nopaste may be of interest to you.
01:20 cotto becase TclObject doesn't inherit the fv ATTR from Float
01:22 Coke ok. if that doesn't break any tests, we can probably just drop the vtable entirely.
01:22 Coke ... and I note that it's not a vtable.
01:22 Coke I hate pmc2c.
01:22 cotto I was curious about that.
01:23 Coke it /used/ to be a vtable. and then pmc2c changed, but tends to be very forgiving (and forgetful) when compiling
01:23 Whiteknight Coke, anybody else express interest in TT#216 yet?
01:23 Coke Whiteknight: not that I've seen.
01:26 * Coke forgets he has to update parrot for this to work.
01:27 cotto might help
01:28 Whiteknight Okay, I'm going to look into it and see what I find
01:28 Whiteknight I suspect that :slurpy argument are hardwired to be RPAs internally
01:30 Coke Whiteknight: no.
01:30 Coke it used to work with a PMC subclass. =-)
01:30 Coke it was only when I switched from PMC to a PIR subclass that it began to fail.
01:30 Coke (just to save you some time.)
01:31 Coke +    void set_number_native(FLOATVAL value) {
01:31 Coke +        SELF.morph(SELF->vtable->base_type);
01:31 Coke +        SET_ATTR_fv(INTERP, SELF, value);
01:31 Coke +    }
01:31 Coke whoops.
01:32 Coke "src/inter_call.c" line 1590 of 2897
01:32 Coke Whiteknight: that's where slurpy is getting the HLL variant.
01:32 Coke cotto: with your patch, I get a test failure in float division. checking...
01:35 cotto That's funny.  The only thing make test complained about was the test I mentioned.
01:35 cotto (for me)
01:41 bkuhn joined #parrot
01:45 particle left #parrot
01:46 Coke lemme realclean and try again.
01:47 skv joined #parrot
01:47 dalek r35935 | julianalbo++ | trunk (2 files):
01:47 dalek : pirric: add example of HLL interoperability
01:47 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35935
01:50 nopaste "coke" at 72.228.52.192 pasted "this test fails" (13 lines) at http://nopaste.snit.ch/15411
01:55 dalek r35936 | julianalbo++ | trunk (2 files):
01:55 dalek : codingstd fixes
01:55 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35936
01:56 skv joined #parrot
02:04 Coke cotto: that one test, I get one less 3 than I'm expecting. =-)
02:13 Coke cotto: should SELF.get_number() still work?
02:17 JC1 joined #parrot
02:22 dalek r35937 | allison++ | trunk/ports/debian (6 files):
02:22 dalek : [debian] Updated packaging for 0.9.0 release. Collapsed several testing
02:22 dalek : releases into one current release. Reverted parrot.docs list to include in
02:22 dalek : parrot package to original short list (full docs should only be installed in
02:22 dalek : parrot-doc package). Removed --optimize flag from configure in rules file, for
02:22 dalek : compatibility with 64-bit architectures. Updated copyright to 2009.
02:22 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35937
02:39 cotto Coke, it should
02:44 skv joined #parrot
02:47 cotto Coke, if you want, I can look into it later tonight
02:49 kid51 joined #parrot
02:50 skv joined #parrot
03:00 cotto Coke, I can repro the problem after updating tcl from svn.  Curiously the problem didn't exist before I updated.
03:22 kid51 cotto:  Can you take a look at https://trac.parrot.org/parrot/ticket/219 ? Thanks.
03:24 dalek r35938 | jkeenan++ | trunk/docs/book (9 files):
03:24 dalek : Applying patch submitted in https://trac.parrot.org/parrot/ticket/215 by
03:24 dalek : mikehh:  POD corrections in docs/book/.
03:24 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35938
03:54 skv joined #parrot
04:08 particle joined #parrot
04:12 skv joined #parrot
04:25 dalek r35939 | coke++ | trunk:
04:25 dalek : we broke the deprecation cycle rule for this one, it was removed before the 0.9.0 release.
04:25 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35939
04:28 * Coke wonders if "make -j3 TEST_JOBS=3 test" used to work.
04:28 Theory joined #parrot
04:40 Coke cotto: no clue why it's failing. looks like it's still a tclfloat and still calling the get_string override.
04:40 Coke (I think)
04:42 Coke chromatic: ping
04:49 * Coke prepares to remove intlist.
04:50 skv joined #parrot
04:51 chromatic pong
04:51 chromatic remove away
04:52 Coke also going to remove src/intlist.c
04:52 Coke seems to just be c-level helper stuff for intlist pmc.
04:52 chromatic Yes.
04:52 Coke just need to fixup a make depend issue and re-test.
04:53 Coke must remember that duplicating my local svn repository and doing a revert -R . is faster that getting a new checkout. :|
05:00 Coke chromatic: I very much do not like the code in src/pmc/iterator.pmc that is poking in specific register numbers.
05:06 dalek r35940 | allison++ | trunk/docs/resources (3 files):
05:06 dalek : [doc] Add and update resources for 'make html' target.
05:06 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35940
05:07 chromatic Looking now, Coke.
05:08 chromatic Oh dear.  That's horrible.
05:09 dalek r35941 | allison++ | trunk/lib/Parrot/Docs (2 files):
05:09 dalek : [doc] Update generated HTML docs to look reasonably similar to parrot.org site,
05:09 dalek : while still keeping simple template.
05:09 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35941
05:10 dalek r35942 | coke++ | trunk (15 files):
05:10 dalek : TT #187 - remove intlist PMC
05:10 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35942
05:10 bacek hi there
05:11 Coke bacek: hio
05:12 bacek Coke:
05:12 skv joined #parrot
05:12 * bacek running git svn rabase in ~/src/parrot first time for last few month...
05:15 Coke ... is there a but?
05:15 Coke or is that more of a "hi"? =-)
05:17 bacek Coke: I'm just trying to find some spare time to do something for parrot/rakudo
05:17 bacek And I'l  have couple of hours today.
05:18 Coke excellent.
05:19 bacek Any particular tickets to resolve?
05:19 * Coke removes another (easier) PMC and kicks off make test.
05:19 Coke I can't speak to rakudo.
05:20 Coke I have a particular fondness for the list here: http://code.google.com/p/partcl/wiki/ParrotIssues =-)
05:25 dalek r35943 | coke++ | trunk (6 files):
05:25 dalek : TT #187 - remove Enumerate PMC
05:25 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35943
05:29 bacek Coke: tt#68 (tcpstream.pir) I can probably fix it. Or you waiting for decision?
05:39 dalek r35944 | coke++ | trunk (8 files):
05:39 dalek : TT #187 - remove MultiArray PMC
05:39 dalek : Also some cruft in editor/pir-mode.el from other recent removals
05:39 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35944
05:42 Coke bacek: I was hoping for a decision; i wouldn't want you to waste your time on something destined for the trash.
05:42 Coke anyone care to disagree with pmichaud on keeping the pair PMC?
05:43 Coke (TT#187)
05:49 bacek rakudo: my Int $a=0; $a = 2 + 3.0; say $.a
05:49 polyglotbot OUTPUT[Type mismatch in assignment.␤current instr.: 'die' pc 16767 (src/builtins/control.pir:204)␤called from Sub 'infix:=' pc 15035 (src/builtins/assign.pir:35)␤called from Sub '_block14' pc 118 (EVAL_17:52)␤called from Sub '!UNIT_START' pc 18152 (src/builtins/guts.pir:321)␤called from Sub
05:49 polyglotbot ..'parrot;PCT;HLLCompiler;eval' pc 950 (src/PCT/HLLCompil...
05:54 Woody2143 joined #parrot
06:05 dalek r35945 | coke++ | trunk (6 files):
06:05 dalek : TT #187 - remove Pair PMC
06:05 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35945
06:06 dalek r35946 | allison++ | trunk/docs/project:
06:06 dalek : [doc] Updading Debian pakaging guide.
06:06 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35946
06:35 dalek r35947 | coke++ | trunk (17 files):
06:35 dalek : TT #187 - remove delegate and deleg_pmc PMCs; these had remnants in
06:35 dalek : several locations; ran 'make configure_tests test', no failures.
06:35 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35947
06:37 dalek r35948 | coke++ | trunk:
06:37 dalek : This item seems to have been resolved in r35928
06:37 dalek : (please add ticket numbers to deprecation items to facilitate tracking))
06:37 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35948
06:37 Coke purl, network?
06:37 purl somebody said network was used so programmers don't have to get up from their chairs in order to share ideas.  They're still working on a way to network their brains and just work as one big collective or viewed by non-programmers as another target for push media or irc.infobot.org
06:39 idemal joined #parrot
06:39 davidfetter joined #parrot
06:42 dalek r35949 | coke++ | trunk:
06:42 dalek : Add newly created ticket reference
06:42 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35949
07:10 Zaba_ joined #parrot
07:52 chromatic Coke is on fire.
08:21 dalek r35950 | simon++ | branches/strings/pseudocode (4 files):
08:21 dalek : First stab at appending graphemes; we can now convert strings from UTF8 to ParrotNative with NFG.
08:21 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35950
08:21 * davidfetter gets out the fire extinguisher
08:21 davidfetter coke_z, do you want to be put out?
08:56 iblechbot joined #parrot
08:58 dalek r35951 | fperrad++ | trunk (8 files):
08:58 dalek : [Markdown] has left the nest to http://github.com/fperrad/markdown/tree/master
08:58 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35951
09:11 dalek r35952 | fperrad++ | trunk:
09:11 dalek : Ignore language checkout 'markdown'
09:11 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35952
09:12 skv joined #parrot
09:14 mberends joined #parrot
09:37 alvar joined #parrot
09:49 dalek r35953 | bernhard++ | trunk/config/gen/makefiles:
09:49 dalek : [build] put back a tab in languages/Makefile
09:49 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35953
10:14 skv joined #parrot
10:15 masak joined #parrot
10:36 skv joined #parrot
11:01 dalek r35954 | julianalbo++ | trunk/examples/pir (2 files):
11:01 dalek : pirric: on error goto 0, improve error handling and upcase keywords and predefined functions in interlang.bas example
11:01 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35954
11:38 bacek joined #parrot
11:50 Gerd joined #parrot
12:15 dalek r35955 | fperrad++ | trunk/languages/WMLScript/pmc:
12:15 dalek : [WMLScript] fix build
12:15 dalek : - since r35934, ATTR are inherited
12:15 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35955
12:30 alvar joined #parrot
12:36 dalek r35956 | simon++ | branches/strings/pseudocode (4 files):
12:36 dalek : Some bug fixes, and now we have UTF8->NFG->UTF8 round-tripping.
12:36 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35956
12:42 lathos That's quite a milestone. I think I'll write to the mailing list about it.
12:50 NotFound lathos: good idea, some of us doesn't even know what strings/pseudocode is ;)
12:52 lathos It's the beginnings of a Perl 6 implementation of Parrot. ;)
12:52 mberends joined #parrot
13:07 dalek r35957 | simon++ | branches/strings/pseudocode (5 files):
13:07 dalek : Rearrange things a bit, having all the encodings in one file was a bit unwieldy
13:07 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35957
13:07 lathos OK, that's enough for one morning.
13:23 Whiteknight joined #parrot
13:29 hudnix joined #parrot
13:43 kid51 joined #parrot
13:48 Casan lathos: whats the plan, we can write everything in pseudocode, and have parrot parse it into any language we want? thought it was just a dream :)
13:53 NotFound If it's parseable is not pseudocode, is code.
13:58 Casan say everything is parseable, if the parser is sufficiently intelligent.
13:59 NotFound Casan: and if the parser exists.
14:00 NotFound If you have one and use it, is no more pseudocode.
14:41 DietCoke joined #parrot
14:43 Coke my internet is wonky; TT#222 looks safe and reasonable to apply.
14:44 cotto That's a very interesting use of the semicolon.
14:46 NotFound Coke: apply wath, the pacth for the "<html...>" thing?
14:49 Coke yes.
14:54 dalek r35958 | julianalbo++ | trunk/t/perl:
14:54 dalek : apply patch fot TT #222, mikehh++
14:54 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35958
15:04 eidolon joined #parrot
15:10 skv joined #parrot
15:14 tetragon joined #parrot
15:14 skv joined #parrot
15:40 Tene_ joined #parrot
15:48 dalek r35959 | julianalbo++ | trunk/examples/nci:
15:48 dalek : fix examples/nci/ls.pir
15:48 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35959
15:51 kj joined #parrot
15:51 Whiteknight Coke, this slurpy issue is more difficult then I anticipated
15:52 Whiteknight I was trying to look at it in the debugger, but that little example you posted calls Parrot_process_args about a hundred times
15:59 Coke_away Whiteknight: welcome to tcl.
15:59 dalek r35960 | julianalbo++ | trunk/examples/past:
15:59 dalek : half-fix examples/past/blocktype_inmediate.pir
15:59 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35960
16:00 * Coke decliens to comment on #219.
16:04 Coke do folks actually find the language ignores in languages/ useful?
16:04 Coke I thought I didn't like it, but I realize I'm ambivalent.
16:05 * Coke wonders if his reply to particle's send ever made it to list.
16:06 Coke (on built in testing)
16:06 kid51 Coke:  What's your problem with TT 219?
16:14 ChrisDavaz joined #parrot
16:19 masak rakudo: say "I'd like to notify the Parrot team about the following:"
16:19 polyglotbot OUTPUT[I'd like to notify the Parrot team about the following:␤]
16:19 masak eh? on #perl6 that gives an out-of-memory error.
16:24 Whiteknight Coke, your mapping for RPA doesn't appear to be registered in the hllmap hash
16:24 Whiteknight so somewhere that is TEH FAILZ
16:26 masak ah, p6eval and polyglotbot run on different machines. of course.
16:27 kj good afternoon all
16:31 eidolon left #parrot
16:32 skv joined #parrot
16:36 Whiteknight Okay, I'm starting to figure this shit out. It's an IMCC problem
16:41 nopaste "NotFound" at 213.96.228.50 pasted "Workaround for TT #216" (18 lines) at http://nopaste.snit.ch/15412
16:42 NotFound Does this help or clarify the thing?
16:43 gravity joined #parrot
16:45 masak I just noticed that RT tickets are always even-numbered.
16:46 NotFound Odd X-)
16:46 masak :)
16:46 masak yes, one wonders why.
17:01 kid51 That's not actually true.
17:01 * kid51 has old tickets that are odd-numbered!  http://rt.perl.org/rt3/Tic​ket/Display.html?id=47391
17:02 masak I sit corrected.
17:02 masak all the recent perl6 tickets are even-numbered, though.
17:05 mberends in old hardware, that would indicate that bit 0 of a data bus is shorted to ground
17:09 mberends for example, a parallel printer does the alphabet as @@BBDDFFHH etc
17:17 mberends masak, are you using linux 32 bit?
17:17 masak mberends: I'm not the one running RT.
17:18 masak perl.org is.
17:19 mberends no, it's not about that RT topic. even numbering there is most likely a software oddity (oops)!
17:21 mberends the reason for asking is, my 'Try it out:' commands run fine on linux-32 but segfault on 64, and I was wondering if you could also give it a try?
17:21 masak I'm on Mac OS X here.
17:21 masak which 'Try it out:' commands?
17:22 mberends some suggestions in http://github.com/eric256/perl6-exa​mples/blob/master/lib/Pod/Makefile
17:22 shorten mberends's url is at http://xrl.us/bedifh
17:22 NotFound Whiteknight: given that the type is created at runtime, I think there is no solution other than doing the map also at runtime.
17:22 * masak checks
17:23 * mberends eats &
17:23 masak um, I'm in the middle of a something or two, might I give it a try later today or tomorrow?
17:25 Coke (ticket numbers in rt are part, I think, of a larger pool of object numbers)
17:29 * Coke is apparently not getting through on the "why are we running this test" question.
17:29 NotFound mberends: Do you have a backtrace?
17:29 * Coke gives up and ignores the thread. *sigh*
17:31 NotFound Coke: Did you talk about TT #219 ?
17:40 mberends masak: no prob, I may do $bugs-- later
17:41 mberends NotFound: never done a backtrace before, could you walk me through it?
17:41 Coke NotFound: what?
17:42 Coke ah.
17:42 Coke NotFound, kid51: I don't wish to get into the problems I have with those tests because I don't feel I can present them in a constructive fashion.
17:43 Coke NotFound: I had said I /wasn't/ going to say anything, which is, of course, actually saying something.
17:43 Coke kid51: ping on t/steps/auto_pmc-01.t
17:44 NotFound mberends: do: ulilit -c unlimited to allow core dumps. Execute the thing that segfaults, then do: gdb binary_prog_executed core , and in gdb use the backtrace command
17:45 Coke kid51: nevermind, I'll open a ticket.
17:45 Whiteknight NotFound, I'm thinking about redoing the hll map registry to be a hash of STRINGs instead of INTVAL type ids
17:45 Whiteknight so that the registration can happen at compile time, but the lookup can happen at runtime
17:46 NotFound Whiteknight: Do you want to make parrot slower? ;)
17:47 Whiteknight Haha, I was also thinking about creating a hash of hll mappings at compile time, and resolve them to type IDs as the respective classes are created at runtime
17:47 Whiteknight there are several options, I need to think about it more
17:48 NotFound I don't see the need for all that infrastucture. If the class is created at runtime, make sense to also do the mapping at runtime.
17:55 mberends NotFound: thanks, got all that. backtrace is 53 lines of 16 digit hex numbers, meaningless to me. I should warn that the parrot segfaulting here is r35309, (the last pre-rvar2 because I cannot find every deprecated bit) probably not interesting for you now.
17:56 mberends s/bit/bit of my code/
17:59 NotFound mberends: uh, looks like your parrot is built without debug information
18:00 mberends yes, lots of question marks my that hex. it was built with perl Makefile.PL;make;make perl6
18:01 mberends do you also have a build-parrot-with-debug-information-howto ?
18:02 NotFound mberends: Configure.pl usually will do without any special option, don't know what the Makefile.PL does
18:03 Coke makefile invokes configure.
18:03 Coke it's just for CPAN compatability.
18:03 Coke or so I thought.
18:04 Coke yah. if you don't specify anything on the command line, it just invokes Configure.pl
18:04 mberends and do you get debug information with that?
18:05 NotFound mberends: take care to specify the correct binary in the gdb command line. It can be perl6 or parrot, depending on what you where doing.
18:06 * mberends begins again, extra carefully
18:07 NotFound The file command can be helpful, do: file core
18:07 NotFound It must says something like: core: ..... from './parrot ....'
18:10 mberends ha! there was an alias perl6=/tmp/parrot/languages/perl6/perl6 in the works!
18:12 NotFound Specify the full path, then
18:12 alvar joined #parrot
18:13 mberends yes, gdb looks more informative now. pasting into a text file...
18:16 mberends nopaste?
18:16 purl rumour has it nopaste is at http://nopaste.snit.ch/ (ask TonyC for new channels) or http://rafb.net/paste or http://poundperl.pastebin.com/ or http://paste.scsys.co.uk/ or App::Nopaste or tools/dev/nopaste.pl or at http://www.extpaste.com/ or http://paste.scsys.co.uk (for #catalyst, #dbix-class, #moose  and others)
18:16 clunker3 http://pasta.test-smoke.org/ or http://paste.husk.org/ or http://nopaste.snit.ch:8001/ or http://rafb.net/paste or http://poundperl.pastebin.com/ or http://paste.scsys.co.uk/
18:18 nopaste "mberends" at 83.161.42.49 pasted "NotFound: r35309 segfault backtrace" (70 lines) at http://nopaste.snit.ch/15413
18:19 mberends thanks for the guidance, I may find time to study the details soon
18:23 NotFound mberends: It will be better to able to check it in the current cvs version
18:28 mberends agreed. I try the lastest svn checkout regularly and cannot pinpoint my deprecated code. I was debugging with $*ERR.say: pre-rvar2, but that stopped working in r35311 (reported RT#62540). I'll run the current rakudo in a moment, to recite the current error.
18:30 skv joined #parrot
18:31 chromatic joined #parrot
18:36 alvar joined #parrot
18:38 mberends NotFound: do you actually have time for this muckraking? it's such an imposition...
18:41 NotFound mberends: I have time, but I don't have a 64 bit system, so I can't do very much.
18:43 barney joined #parrot
18:43 mberends let me fill your time... ;-) apart from segfaults, both cpu families give the same language errors.
18:44 NotFound mberends: uh, perl6 syntax is not one of my areas of expertise :D
18:45 mberends Null PMC access in get_string() current instr.: 'infix:ne' .... there is only one string 'ne' comparison in the source code, and when that is commented out, another error takes its place...
18:46 mberends are you willing to try a copy of the code on your system?
18:46 NotFound What code is?
18:46 purl code is, like, message mapping => usually pulled straight from the RFC
18:47 * mberends is not sure whether to censor purl on that opinion
18:47 NotFound mberends: git://github.com/eric256/perl6-examples.git ?
18:47 mberends exactly. then, cd lib/Pod; make clean test
18:48 NotFound Wait a little, I must install git first
18:49 mberends sudo apt-get install git-core # afair
18:52 mberends NotFound: the Perl 6 depth is limited. the saying is something like, "a good programmer can write Fortran in any language". It could have been written in Fortran.
18:53 ask_ joined #parrot
18:53 jhorwitz joined #parrot
18:54 NotFound I think is a *true* programmer, not just a good one ;)
18:55 mberends *nod*
19:00 mberends so, that combination will definitely throw errors. to see it run as intended, you could add: svn checkout --revision 35309 http://svn.perl.org/parrot/trunk /tmp/parrot;cd /tmp/parrot;perl Makefile.pl;make;make perl6
19:00 NotFound mberends: Do you know at what Parser.pm line corrspond the pir instruction?
19:02 mberends I suspect 440: if @!podblocks[$topindex].style ne 'DELIMITED' { # ABBREVIATED or POD_BLOCK
19:03 mberends meaning .style is undef
19:05 NotFound Mmmm... but an undef usage will give "Use of uninitialized value, isn' it?
19:06 mberends oh
19:07 kid51 Problem with SVN?  $ ping http://svn.perl.org/parrot/
19:07 kid51 PING http://svn.perl.org/parrot/ (63.251.179.13): 56 data bytes
19:07 kid51 ^C
19:07 kid51 --- http://svn.perl.org/parrot/ ping statistics ---
19:07 kid51 11 packets transmitted, 0 packets received, 100% packet loss
19:09 mberends NotFound: so far my approach has been to tweak the running code on r35309 for sanity, and then venture to svn HEAD periodically in the hope of a breakthrough. Admittedly not deterministic at all. can you suggest a more productive modus operandi?
19:09 dalek r35961 | jkeenan++ | trunk/t/steps:
19:09 dalek : In partial satisfaction of issues raised in
19:09 dalek : https://trac.parrot.org/parrot/ticket/223, reformat lists to match coding
19:09 dalek : standards.
19:09 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35961
19:09 kid51 That was strange:  I was asked to authenticate .. for first time in >2 years!
19:10 kid51 $ svn ci t/steps/auto_pmc-01.t
19:10 kid51 Authentication realm: <https://svn.perl.org:443> perl.org
19:10 kid51 Password for 'jkeenan':
19:10 NotFound mberends: better ask some rakudo guy, I don't know what are the good ways to diagnose rakudo problems.
19:11 mberends will do. thanks NotFound++
19:13 * mberends suggests to self: port code into rakudo latest revision piece by piece. it's obvious!
19:13 NotFound kid51: I don't had any problem today, other than some of the usual hangs when commiting
19:39 NotFound Now I have the same issue as kid51, it ask me for my pass
19:46 Zaba joined #parrot
19:46 chromatic joined #parrot
19:47 Whiteknight NotFound, so are you suggesting an hllmap opcode?
19:47 Whiteknight Because I like that idea too
19:47 skv joined #parrot
19:48 NotFound Whiteknight: no need for opcode, there is a method in the interpreter pmc
19:50 NotFound Whiteknight: http://nopaste.snit.ch/15412
19:51 Whiteknight That's not a bad idea either
19:52 * Whiteknight really needs to read over all fields and methods of the interpreter PMC object
19:54 NotFound It has very few methods, maybe it will benefit from some more.
19:54 skv joined #parrot
19:56 Whiteknight If runtime options are the way we want to go, then we should consider deprecating .HLL_map
19:56 NotFound Whiteknight: or document and enforce that is only for pmc, not for class
19:58 Whiteknight what do you mean by that?
19:58 Whiteknight oh, nevermind. I think I got it
19:58 Whiteknight But I don't agree, we shouldn't be treating C-based PMCs any differently from how we treat PIR-based classes
19:59 Whiteknight The differences should be transparent to the user
19:59 NotFound Then deprecation looks like a good way to me.
20:00 Whiteknight Let me post a note about it to the list then.
20:11 dalek r35962 | julianalbo++ | trunk/examples/pir:
20:11 dalek : pirric: activate deprecation warnings
20:11 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35962
20:12 skv joined #parrot
20:16 dalek r35963 | bernhard++ | trunk/docs/resources:
20:16 dalek : [codingstd] svn properties
20:16 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35963
20:16 skv joined #parrot
20:29 dalek r35964 | bernhard++ | trunk (4 files):
20:29 dalek : [unlambda] add a Configure.pl
20:29 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35964
20:40 dalek r35965 | fperrad++ | trunk/config/gen/makefiles:
20:40 dalek : [external languages]
20:40 dalek : checkout PJS in the directory 'pjs' instead of 'parrotjs-read-only'
20:40 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35965
21:02 NotFound Whiteknight: I've found that in t/op/box.t there is an alternative approach: creating the class in a :inmediate sub located before all .HLL things
21:04 Whiteknight Ah! That's what I wanted to try. I couldn't remember about :immediate, I tried using :init :load instead, and that didn't work correctly
21:07 NotFound Note that the classes here are not created inside the .HLL, though
21:10 skv joined #parrot
21:14 Whiteknight Yeah, but either way .HLL_map is limited and idiosyncratic
21:14 Whiteknight I would rather have a solution that Just Works instead of one that's limited
21:21 dalek r35966 | bernhard++ | trunk (3 files):
21:21 dalek : [unlambda] add example with the "k" function
21:21 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35966
21:29 Theory joined #parrot
21:30 skv joined #parrot
21:32 NotFound Whiteknight: I agree. And kill some code that only exists for that purpose.
21:52 dalek r35967 | bernhard++ | trunk/languages/pipp/src/pct:
21:52 dalek : [Pipp] name mangling for 'use' vars is not necessary
21:52 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35967
21:52 dalek r35968 | bernhard++ | trunk/languages/unlambda/t:
21:52 dalek : [unlambda] defined order of test cases
21:52 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35968
21:56 skv joined #parrot
22:01 lathos eval creates a new scope?
22:02 lathos I can't seem to use lexicals from an enclosing scope inside an eval in rakudo.
22:06 dalek r35969 | bernhard++ | trunk (4 files):
22:06 dalek : [unlambda] build unl.pbc and use it for testing
22:06 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35969
22:09 lathos Another fun one:
22:09 lathos % ~/svn/parrot/perl6 -e 'my $a = eval ""'
22:09 lathos Null PMC access in isa()
22:10 dalek r35970 | bernhard++ | trunk/languages/unlambda:
22:10 dalek : [unlambda] PIR formatting
22:10 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35970
22:21 redlemon joined #parrot
22:22 lathos That one's fixed in head though.
22:23 redlemon joined #parrot
22:24 redlemon left #parrot
22:33 lathos Yuck. Rakudo HEAD resolves class names at compile time, not run time.
22:33 chromatic The phrase "insufficiently dynamic" describes a lot of things.
22:34 lathos rakudo: class Foo {}; sub A { Foo.new(); }; say A()
22:34 polyglotbot OUTPUT[Foo<0xb6c021a8>␤]
22:34 lathos rakudo: sub A { Foo.new(); }; class Foo { }; say A()
22:34 polyglotbot OUTPUT[invoke() not implemented in class 'Foo'␤current instr.: 'A' pc 124 (EVAL_19:59)␤called from Sub '_block14' pc 64 (EVAL_19:40)␤called from Sub '!UNIT_START' pc 18152 (src/builtins/guts.pir:321)␤called from Sub 'parrot;PCT;HLLCompiler;eval' pc 950 (src/PCT/HLLCompiler.pir:527)␤called from Sub
22:34 polyglotbot ..'parrot;PCT;HLLCompiler;evalfiles' pc 1275 (src/...
22:34 lathos NO. WRONG. DON'T DO THAT.
22:36 dalek r35971 | simon++ | branches/strings/pseudocode:
22:36 dalek : Work around current Rakudo idiosyncrasies.
22:36 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35971
22:38 dalek r35972 | chromatic++ | trunk (2 files):
22:38 dalek : [src] Removed one use of deprecated SArray PMC.  All tests pass.
22:38 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35972
22:43 dalek r35973 | chromatic++ | trunk/src:
22:43 dalek : [src] Removed the other use of the deprecated SArray PMC.
22:43 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35973
22:53 alvar joined #parrot
22:54 dalek r35974 | julianalbo++ | trunk/src:
22:54 dalek : quick fix for TT #65
22:54 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35974
22:56 dalek r35975 | chromatic++ | trunk (15 files):
22:56 dalek : [PMC] Removed SArray PMC, per TT #191.  Use a FixedPMCArray now instead.
22:56 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35975
23:15 dalek r35976 | julianalbo++ | trunk/t/steps:
23:15 dalek : fix? for t/steps/auto_pmc-01.t
23:15 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35976
23:17 dalek r35977 | chromatic++ | trunk/lib/Parrot/Pmc2c (2 files):
23:17 dalek : [lib] Renamed package in Parrot::Pmc2c::Dumper to just that and exported
23:17 dalek : dump_pmc() so that the real Parrot::Pmc2c::Dumper can use it.  See TT #224,
23:17 dalek : reported by Jim Keenan.
23:17 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35977
23:17 samlh hello, the link on rakudo.org to the parrot wiki is broken
23:19 Limbic_Region joined #parrot
23:45 dalek r35978 | julianalbo++ | trunk/t/steps:
23:45 dalek : fix?? for t/steps/auto_pmc-01.t
23:45 dalek review: http://www.parrotvm.org/svn​/parrot/revision?rev=35978

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

Parrot | source cross referenced