Camelia, the Perl 6 bug

IRC log for #parrot, 2008-08-26

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 bacek joined #parrot
00:29 Tene tewk++
01:09 Lorn joined #parrot
01:17 rurban_ joined #parrot
01:26 paco joined #parrot
01:32 dalek r30553 | chromatic++ | trunk:
01:32 dalek : [t] Improved test file such that its diagnostics are useful for debugging.
01:32 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30553
02:07 kid51 joined #parrot
02:09 kid51 Undefined command: "������".  Try "help".   -- Doesn't that come from the Spanish version of "Monty Python and the Holy Grail"?
02:12 tetragon I think they omitted the pilcrows that I'm seeing here
02:15 * kid51 knows not pilcrows
02:15 kid51 purl, pilcrow is a paragraph mark
02:15 purl OK, kid51.
02:16 kid51 Notwithstanding more than a decade working as a typographer, I never encountered the term 'pilcrow' before.
02:16 tetragon I spent half of my work day fighting with UTF-8, CP1252, and macroman encoded text
02:16 kid51 Typographer's nightmare!
02:16 tetragon The pilcrow was one of the symbols that appeared in the misinterpreted text
02:17 tetragon The codepoints just so happened to match
02:27 dalek r30554 | tene++ | trunk:
02:27 dalek : [cardinal]
02:27 dalek : * Many new tests for Array, thanks to Damian Knopp.
02:27 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30554
02:48 Tene Wait, # is a non-portable character in filenames?
02:48 Tene Huh.
02:54 dalek r30555 | tene++ | trunk:
02:54 dalek : [cardinal]
02:54 dalek : * Clean up some mess from adding new test files.
02:54 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30555
02:55 dalek r30556 | tene++ | trunk:
02:55 dalek : Oops, one more file with non-portable characters still.
02:55 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30556
03:24 petdance joined #parrot
03:52 grim_fandango joined #parrot
04:12 tetragon joined #parrot
05:24 Zaba_ joined #parrot
05:27 Psyche^ joined #parrot
05:31 Ademan joined #parrot
05:42 dalek r30557 | chromatic++ | trunk:
05:42 dalek : [PDB] Fixed debugger segfault when reaching the end of PIR read from the
05:42 dalek : command line; the program counter would walk off the end of the packfile,
05:42 dalek : making the next opcode number very invalid.  Now there's bounds checking at the
05:42 dalek : end of all logical lines of code.
05:42 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30557
05:49 cotto_home very invalid is much worse than a little invalid
05:58 Copas joined #parrot
06:06 dalek r30558 | julianalbo++ | trunk:
06:06 dalek : fix the fix for RT#58358 by chromatic's sugestion
06:06 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30558
06:13 cotto_w0rk joined #parrot
06:25 uniejo joined #parrot
06:25 dalek r30559 | julianalbo++ | trunk:
06:25 dalek : [PDB] allow debugger commands without delimiting whitespace
06:25 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30559
06:32 masak joined #parrot
06:54 masak the nifty {*} thingies, used all over the place to tie actions.pm to grammar.pg...
06:54 masak are they avaliable through Rakudo yet?
06:54 masak I'm writing a parser, and they would help immensely
06:57 moritz I don't think rakudo does any kind of closure in regexes yet
06:58 masak ah, true
06:58 masak moritz: thanks for the HTML::Template grammar yesterday
06:58 masak I had great fun turning it into a test suite and a module
06:58 moritz ;)
06:59 masak many small things changed on the way, but the essence is still there
06:59 masak hold on, let me paste the result-so-far for you
06:59 moritz with closures iin regexes the HTML::Template thing would be almost too trivial ;)
07:00 masak that's what I'm after
07:00 purl after is fine
07:00 masak purl: forget after
07:00 purl masak: I forgot after
07:00 masak I want Perl 6 to do the heavy lifting for me
07:03 masak http://sial.org/pbot/31991 and http://sial.org/pbot/31992
07:04 moritz so you renamed TMPL_LOOP to TMPL_FOR
07:04 moritz very good idea ;)
07:04 masak thank you :)
07:05 masak it's clearly wrongly named, now that Perl 6 has a keyword 'loop'
07:05 masak (...with semantics that wouldn't make much sense for HTML::Template)
07:06 moritz and the nice thing is that you can get the old semantics by subclassing the grammar ;)
07:06 masak ooh
07:06 masak yes :)
07:07 moritz I think some day '</TMPL_IF>'  should turn into [ '</TMPL_IF>'  || <panic: 'mssing </TMPL_IF>'> ]
07:07 masak yes, good idea
07:07 purl masak: Good Idea: Going trick-or-treating on Halloween. Bad Idea: Going trick-or-treating on St Patrick's Day.
07:08 masak purl: :)
07:08 purl masak: Unbalanced parentheses (too many ')')
07:08 masak eh.
07:09 moritz lol
07:11 masak moritz: I'm curious about what the '.' means in <.tag_start>
07:11 masak I haven't hit the synopsis yet to find out :)
07:11 Tene It means non-capturing.
07:12 masak ah, nice
07:15 masak also, I'm thinking changing `token attributes { \s+ 'NAME='? <name> [\s+ 'ESCAPE=' <escape> ]? };`
07:15 masak to `rule attributes { 'NAME='? <name> ['ESCAPE=' <escape> ]? };`
07:16 masak s/thinking/thinking of/
07:17 dalek r30560 | cotto++ | trunk:
07:17 dalek : [pipp] more freed and PIR'd phparray tests
07:17 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30560
07:17 moritz masak: if it works that's a good idea, IMHO
07:17 * masak tries
07:18 masak at present I have unforeseen trouble with test #3
07:18 masak but tests++
07:19 moritz we need tests for nested stuff at some point
07:19 moritz like nested loops
07:19 masak yes
07:19 masak one of the comments in HTML::Template is about that
07:20 masak I'll just get these tests to work first, then I'll add more :)
07:20 masak the current design is already much more pleasant to work with than the last one
07:21 moritz perhaps becase it's a design at all? ;-)
07:21 masak most probably
07:22 masak hm, I should really send tree fragments when I recurse, not strings. that way, I wond't have to match again.
07:24 * moritz has to go the dentist :(, TTFN
07:28 Zaba joined #parrot
07:28 gmansi joined #parrot
07:48 iblechbot joined #parrot
07:53 dalek r30561 | allison++ | trunk:
07:53 dalek : [pdd] Removing credits that are no longer accurate.
07:53 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30561
07:55 dalek r30562 | julianalbo++ | trunk:
07:55 dalek : [PDB] notnull condition for string and pmc registers
07:55 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30562
08:50 Debolaz joined #parrot
09:01 dalek r30563 | julianalbo++ | trunk:
09:01 dalek : [PDB] some more refactoring
09:01 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30563
09:14 Whiteknight joined #parrot
09:17 rurban_ joined #parrot
09:18 dalek r30564 | kjs++ | trunk:
09:18 dalek : [pirc/new] update documentation.
09:18 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30564
09:18 kj joined #parrot
09:45 dalek r30565 | kjs++ | trunk:
09:45 dalek : [pirc/new] improve allocation of PASM registers: only give a symbol a PASM register if it's actually used, not when merely declaring it. This prevents assigning (valuable) PASM registers to unused (but declared) symbols.
09:45 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30565
09:49 ejs joined #parrot
10:16 rba joined #parrot
10:18 dalek r30566 | kjs++ | trunk:
10:18 dalek : [pirc/new] use symbol->color for encoding usage of a symbol. Removing symbol->used flag (saving sizeof(int) bytes per symbol node).
10:18 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30566
10:27 gaz joined #parrot
10:34 Whiteknight joined #parrot
10:39 pako joined #parrot
10:39 pako left #parrot
11:09 ejs joined #parrot
11:11 ejs joined #parrot
11:14 dalek r30567 | kjs++ | trunk:
11:14 dalek : [pirc/new] fixes to allow special-handled ops (add, sub, set, etc.) as parameter name, constant name, subroutine name, local name.
11:14 dalek : + refactor duplicate code in lexer.
11:14 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30567
11:25 dalek r30568 | kjs++ | trunk:
11:25 dalek : [pdd19] add a comment that PIR keywords cannot be used as PIR identifiers (as opposed to opcode names, which are allowed).
11:25 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30568
11:29 dalek r30569 | kjs++ | trunk:
11:29 dalek : [pdd19] be a bit more precise in opcode names vs local var names vs sub names.
11:29 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30569
11:33 kid51 joined #parrot
11:34 bacek joined #parrot
11:34 ejs joined #parrot
11:39 kid51 joined #parrot
11:53 dalek r30570 | julianalbo++ | trunk:
11:53 dalek : [PDB] print line number and close file on script errors, RT#46117
11:53 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30570
12:06 dalek r30571 | kjs++ | trunk:
12:06 dalek : [pirc/new] add comments to pir.y and allocate a new PASM register for sub calls (instead of using P99, which is nonsense).
12:06 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30571
12:45 tetragon joined #parrot
13:14 dalek r30572 | kjs++ | trunk:
13:14 dalek : [pirc/new] add an assertion and add documentation.
13:14 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30572
13:22 gryphon joined #parrot
13:36 AndyA joined #parrot
13:41 AndyA joined #parrot
13:59 AndyA joined #parrot
14:03 cjfields joined #parrot
14:04 dalek r30573 | julianalbo++ | trunk:
14:04 dalek : [PDB] fix pmc notnull condition
14:04 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30573
14:08 cheflog__ joined #parrot
14:53 dalek r30574 | julianalbo++ | trunk:
14:53 dalek : [PDB] check register is in range in conditions, RT#46125
14:53 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30574
15:10 particle (11:54:00 PM) masak: the nifty {*} thingies, used all over the place to tie actions.pm to grammar.pg...
15:10 particle (11:54:11 PM) masak: are they avaliable through Rakudo yet?
15:11 masak yes, are they?
15:11 particle you mean just {*} or closures in general?
15:11 masak just {*}
15:11 masak so I can tie grammar things to methods
15:12 moritz at least they parse
15:12 particle looking now at how it's implemented in nqp
15:12 particle yes, they should parse.
15:12 particle but i don't know if they parse correctly :)
15:13 masak heh :)
15:15 barney joined #parrot
15:15 moritz [3] => PMC 'PGE::Exp::Action' => "{*}" @ 32
15:15 particle seems to be parsing as a block with get_global Whatever
15:15 moritz <type> => "term:{*}"
15:16 masak particle: how does that translate into whether Rakudo can access it?
15:17 particle still looking :)
15:18 masak particle++
15:20 particle pge's perl6regex parser is where term:{*} is defined
15:21 particle which means that the perl 6 grammar parser understands it natively
15:21 masak good news, I guess
15:22 particle the thing is, {*} is sort of a trick in STD
15:22 particle it's just a block with whatever inside, in real perl 6
15:22 Wknight8111 joined #parrot
15:23 moritz maybe it's a trick, but I quite like it ;)
15:23 particle yes, it's a neat trick :)
15:23 moritz otherwise we'd mix up closures and regexes a lot
15:23 masak I'd say it's a trick-become-de-facto-standard
15:23 particle sure, like #?
15:23 moritz which goes against the grain of separating form (syntax) from logic (closures)
15:24 masak the standard semantics of {*} is not much fun anyway :)
15:24 particle there's no spec for {*}, so it's not defined in STD
15:25 Andy joined #parrot
15:25 moritz we should bug TimToady over that
15:27 particle yep, seems so
15:27 * moritz bugs TimToady ;)
15:29 cognominal_ {*} does not appear in the synopses as far I can tell.
15:29 particle correct
15:43 ejs_ joined #parrot
15:44 cjfields re: {*}, I think pmichaud indicated these would be Closures and aren't implemented in Rakudo (yet)
15:45 cjfields been a while since I've been on list, though, so maybe things have changed
15:46 pmichaud TimToady originally envisioned {*} as being "empty closures"
15:46 pmichaud PGE goes a step further and recognizes them as action tokens that tie them to an action grammar
15:47 pmichaud I don't think there's been an official decision as to whether or not PGE's approach is valid :-)
15:47 pmichaud and STD.pm has since eliminated a lot of its {*} markers
15:47 moritz that's because it defeats LTM, iirc
15:48 pmichaud yes, it will do that.  It makes sense ot me to eliminate the {*} markers where they weren't really needed anyway
15:48 pmichaud STD.pm seems to eliminate quite a bit more than that
15:48 pmichaud so, we'll see what happens when I start doing LTM in PGE
15:48 cjfields makes sense to wait on it, then
15:49 moritz cjfields: we just discussed them because they'd be very useful for the Perl 6 implementation of HTML::Template ;)
15:49 pmichaud the {*} marker is likely to remain in PGE for some time
15:49 pmichaud if there's no action grammar present, then it acts exactly like {*} in Perl 6 :-)
15:50 moritz but currently masak has a recursive version for filling out the template variables, so no hurry
15:50 iblechbot joined #parrot
15:50 particle so, the question is, can it be made to work in rakudo?
15:51 cjfields {*} would be nice for us bioinformatics types (sequence file parsing)
15:51 pmichaud probably.  the trick is to be able to invoke the regex/grammar with 'action' as a named argument
15:52 pmichaud (and 'action' points to an instance of the action grammar)
15:52 pmichaud I'd have to think on the syntax a bit
15:52 particle use parseaction;
15:52 pmichaud no
15:52 pmichaud more like
15:53 pmichaud Grammar.TOP($target, :action(ActionGrammar))
15:53 pmichaud or probably
15:53 particle hrmm
15:53 pmichaud eventually it's likely to be
15:53 moritz $str ~~ Grammar.TOP :action(ActionGrammar)
15:53 pmichaud Grammar.new.TOP($target, :action(ActionGrammar))
15:53 particle yeah
15:54 moritz btw in rakudo you need Grammar::TOP
15:54 particle how do you add an attribute to a class at runtime?
15:54 cotto_work joined #parrot
15:54 pmichaud in Perl 6?
15:54 particle yes
15:54 particle seems it's the same as what we're trying here
15:54 pmichaud class Foo is also { has $.attribute; }
15:54 purl okay, pmichaud.
15:54 pmichaud ...but that's not really what I'd want for this.
15:55 cjfields mixin?
15:55 purl i heard mixin was multiple inheritance without the inheritance. or the secret superset of functionality that inheritance sub-types
15:55 pmichaud no, I think that passing the ActionGrammar as a named parameter is probably about right
15:55 particle so would you like to have a class method on Grammar?
15:55 particle at construction time, sure, that makes sense
15:55 pmichaud it doesn't need to attach to the Grammar.  It ought to be possible to use the same grammar with different actions
15:55 pmichaud oh, we could do it as
15:56 pmichaud Grammar.new($target, :action(ActionGrammar)).TOP()
15:56 pmichaud I'd have to go look and see what syntax STD.pm and friends are using these days
15:57 gryphon joined #parrot
15:58 moritz what's $target? the string against which you match?
15:58 particle yep
15:58 pmichaud yes.
16:04 rurban May I ask about plans for parrrot_config? Should parrot_config read config the hash at runtime from some of the source files, or use the linked hash?
16:05 rurban source files: config_lib.pasm (not installed currently), or include/config.fpmc
16:06 rurban Because I just added a patch for installable_parrot_config to #58354
16:15 cjfields moritz: did you see my comment on #56970?
16:15 barney rurban: I'm just running 'make install' in cygwin070patches/languages
16:16 rurban and then rename your build_dir and test the installables. pipp, tcl, forth and m4 will fail
16:16 moritz cjfields: right, I saw that bug later on
16:17 cjfields ok
16:17 rurban But I have top go to basketball now, sorry. will be back at 22:00 CEST (in 3 hrs + 30min)
16:17 moritz so should we I commit yoru patch, and hope that the fix for #57094 (when somebody writes it) also fixes your tests?
16:18 cjfields we should probably wait for pmichaud to have a look
16:19 barney Setting the environment var LANG is most likely a bad idea
16:20 davidfetter joined #parrot
16:20 rurban barney: pheme will also fail with searching for TGE.pbc
16:20 rurban oops, your are right. but this just a Makefile var, not to ENV
16:21 rurban this will not change libintl
16:21 cjfields moritz: not sure how to fix that myself w/o making a special-case, which I would like to avoid (jonathan also brought this up)
16:22 rurban barney: pdd30_parrot_install.pod which I posted recently explains the summary
16:22 moritz cjfields: anyway, I fudged the ok $str.match(/h/)~~Match test for now, and I still get some failures (in S05-substitution/match.t)
16:22 rurban not really: draft/pdd30_install.pod  it is
16:22 moritz cjfields: Null PMC access in can()
16:22 cjfields maybe have Match return another Match when item is called?
16:23 moritz cjfields: do you know if that's that same bug?
16:23 rurban Ok, I have to leave now.
16:23 moritz cjfields: so the bug is in the contextualizer?
16:24 barney rurban: I get warnings about   LANG = "tcl" and so on. Don't know where it is set.
16:25 cjfields That should be the same bug, yes (.match is returning a Str)
16:25 rurban LANG=tcl in tcl/config/makefiles/root.in
16:25 rurban first line
16:25 purl first line is probably leaking through.
16:28 Wknight8111 purl forget first line
16:28 purl Wknight8111: I forgot first line
16:29 cjfields moritz: if pmichaud indicates the patch is fine, we could commit it, fudge that test, and try fixing it later
16:29 cjfields the bug is already reported in RT.
16:30 pmichaud I have to run an errand now -- will look when I get back in about 60 mins
16:30 moritz cjfields: agreed
16:32 barney rurban: Should I add cardinal.install to https://svn.perl.org/parrot/branches/cygwin0​70patches/config/gen/makefiles/languages.in ?
16:33 cjfields lunch &
16:39 particle in the past hour, i've gotten two calls about new births
16:40 * particle steps away from the phone
16:40 peepsalot joined #parrot
16:42 * Wknight8111 has three babies, puts phone on autodial
16:45 cognominal_ waiting the call for the birth of Rakudo 1.0 :)
16:48 Wknight8111 wouldn't that be a nice phone message?
16:49 moritz indeed ;)
16:57 Tene purl: parrotsketch?
16:57 purl parrotsketch is a status meeting for parrot core committers held every Tuesday at 18:30 UTC in #parrotsketch
16:57 NotFound barney: Why do you think that setting LANG is bad?
16:58 dalek r30575 | bernhard++ | trunk:
16:58 dalek : [Eclectus]
16:58 dalek : Remove unused Perl 5 module Parrot::Test::Eclectus
16:58 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30575
16:59 particle because LANG is too generic an env var name to use for something as specific as a parrot hll
16:59 NotFound Ah, setting it to the HLL, I see now.
16:59 NotFound Yes, that way is very bad.
17:00 nopaste "barney" at 84.154.11.188 pasted "Warning about LANG when doing 'make install' in languages" (10 lines) at http://nopaste.snit.ch/13881
17:01 NotFound perl warns when LANG is set to a locale that it has not info about.
17:02 NotFound There is a way to avoid that warning, but is better not to play with the LANG env var,
17:03 barney Isn't there a way to have vars in Makefiles, without setting the env ?
17:11 clunker9 joined #parrot
17:15 barney Ah, when an env var called LANG already exists, then it is passed on to child processes
17:15 Khisanth joined #parrot
17:15 barney Otherwise it stays local to the Makefile.
17:16 NotFound barney: the problem is that LANG almost always exists.
17:17 rurban_ joined #parrot
17:18 barney but apparently not in rurbans setup
17:38 Zaba_ joined #parrot
17:49 particle call it PARROT_HLL and that goes away
17:49 moritz cygwin != normal ;)
17:53 dalek r30576 | kjs++ | trunk:
17:53 dalek : [pirc/new] add infrastructure for reentrancy testing using pthreads. Disabled by default. Must be ripped out when we're convinced it's thread-safe. (I'm convinced :-)
17:53 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30576
18:08 jhorwitz joined #parrot
18:20 dalek r30577 | bernhard++ | cygwin070patches:
18:20 dalek : [build] add some targets for 'make install'
18:20 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30577
18:22 dalek r30578 | kjs++ | trunk:
18:22 dalek : [pirc/new] change output of generated so, that it can be written to a file. Now we can properly check thread-safety, by starting several threads parsing the same file (we could add possibility for different files later), each thread writes output to its own output file; all output files should output the expected PASM (if not, then it's not thread-safe).
18:22 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30578
18:26 moritz #ps in $soon
18:26 moritz aka 4 mintues
18:27 allison joined #parrot
18:27 chromatic joined #parrot
18:27 chromatic #ps in 3
18:28 particle looks like kid51 got banned from #ps due to flooding
18:28 particle who can fix that?
18:28 particle i don't know any ircops
18:30 chromatic No one has ops in channel.
18:30 particle nope
18:30 dalek r30579 | bernhard++ | trunk:
18:30 dalek : [Pipp] Add single quotes around PMC names, chromatic++
18:30 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30579
18:32 moritz particle: looks like kid51 wasn't banned after all
18:32 particle hrmm, okies, thanks for checking
18:43 rdice joined #parrot
18:44 cotto_w0rk joined #parrot
18:48 cotto_w3rk joined #parrot
18:49 kj pmichaud: ping
18:50 pmichaud pong
18:50 kj hi, some days ago I sent you an email about a parrotblog draft. Did you get that?
18:52 pmichaud hmm.... not sure I saw that
18:52 kj ok, sent it to your pobox address
18:52 kj august 20 it was
19:04 pmichaud kj: okay, I found it
19:05 kj there's no hurry, just want to have it checked for sanity
19:19 dalek r30580 | kjs++ | trunk:
19:19 dalek : [pirc/new] disable re-entrancy testing on windows by default (assuming ms visual studio as compiler), and remove a comment that is no longer applicable.
19:19 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30580
19:20 jdv79 joined #parrot
19:39 rurban barney: I'm back. Did I miss cardinal.install? You're right. I missed that. Yes please
19:51 rurban I just sent some cygwin070patches to parrot-porters: LANG=>TMPL, cardinal, root makefile stuff
19:51 rurban purl: irclog?
19:51 purl rurban: bugger all, i dunno
19:51 cotto_work irc log?
19:51 purl irc log is http://irclog.perlgeek.de/parrot/
19:52 rurban Because I always miss the meeting for my basketball session.
19:53 moritz http://irclog.perlgeek.de/parrotsketch/2008-08-26
19:57 rurban chromatic: I've sent the signed CLA last week monday per snail mail. That's obviously very snail. I'm used to DHL speed.
19:57 chromatic It should be here today then.
19:58 particle rurban: it goes to a po box, which isn't checked every day
19:58 rurban can I ask about my _config bootstrap problem?
19:59 Auzon joined #parrot
20:02 dalek r30581 | smash++ | pdd27mmd:
20:02 dalek : [branches/pdd27mmd]
20:02 dalek :  * convert VTABLE with MMD_ blocks to MULTI definitions in src/dynpmc
20:02 dalek :  * allison++ for fixing pmc2c parser
20:02 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30581
20:09 NotFound I think I found the problem in the ATTR of perl6multisub: a pointer to a pointer is not recognized.
20:22 rurban Anyone knows why #CONDITIONED_LINE(win32): is used? 'win32' is no platform name and no config key.
20:24 Whiteknight probably a typo. What is the platform name for 32-bit windows?
20:24 NotFound ping cotto_work
20:24 purl I can't find cotto_work in the DNS.
20:25 NotFound cotto_work: ping
20:28 particle platform name for 32-bit windows is win32
20:28 NotFound <allison> it'll parse anything of the form 'ATTR <type> *<name>' or 'ATTR <type> <name>' --> but candidates_sorted has the form: ATTR type **name
20:28 particle there's no typo
20:28 rurban interesting but better then p5 $^O. I'll have to fix my logic now.
20:29 NotFound And other problem is that struct candidate_info is defined inside the pmc file, and the .h generted has no access to it,
20:29 rurban s/then/than/
20:29 particle it's stored in parrot::config
20:30 particle gvim config/gen/platform.pm +39
20:32 iblechbot_ joined #parrot
20:32 cotto_work NotFound: yes
20:33 cotto_work NotFound: pong #gotta do it right
20:33 NotFound cotto_work: see my last lines, about the ATTR problem you asked for in #ps
20:33 ejs joined #parrot
20:35 cotto_work NotFound, that looks similar to what I was running into
20:36 cotto_work all my structs were also defined in the same file
20:36 NotFound I'm trying a quick fix: allowing 'struct' before <type> and more than one *, perl6 coretest pass
20:37 jan joined #parrot
20:37 NotFound I'll post the patch to the list to ask for comments.
20:37 Zaba joined #parrot
20:37 chromatic I prefer #typedef struct Foo Foo
20:38 NotFound chromatic: but the problem is that the typedef is not visible in the generated .h
20:41 chromatic Ah, C.  The Principle of Dumbest Astonishment strikes again.
20:42 particle isn't that then a problem with the generator?
20:42 particle shouldn't the typedef appear there
20:43 NotFound Yes, the struct pointer is accepted as a forward declaration of the struct, but whithout ths struct is just an undeclared identifier,
20:45 chromatic I don't really *like* the idea of making the generator stuff declarations in the header, but that really does seem cleaner.
20:46 NotFound Allow the struct keyword in the type is a quick solution, even if a little odd.
20:48 NotFound (INTVAL|FLOATVAL|STRING\s+\*|​PMC\s+\*|\w+\s+\*|Parrot_\w*) --- (INTVAL|FLOATVAL|STRING\s+\*|PMC\s+\*|\​w+\s+\*+|struct\s+\w+\s+\*+|Parrot_\w*)
20:48 NotFound This is the change I'm testing
20:48 particle struct
20:48 particle er
20:48 particle struct--
20:48 particle we excised them from the parrot source last year
20:50 Whiteknight particle, question for you
20:50 NotFound There are two separated items: the not viewable typedef, and the pointer to pointer.
20:50 Whiteknight at the beginning of the summer, you mentioned a PGE parse tree optimizer as a needed project. Is that still a need?
20:58 dalek r30582 | Whiteknight++ | pdd09gc:
20:58 dalek : [pdd09gc] update to trunk from r30548
20:58 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30582
21:04 japhb_ moritz: sorry if you've already seen this, but: Perl-5-to-6 lessons are appearing in essentially random order.  Just an FYI.
21:04 moritz japhb_: yes, that's why I have an ordered index at the start
21:05 japhb_ moritz: Fair enough -- but that doesn't help for someone (like me) who is watching the whole blog, not just that section.
21:05 japhb_ No index.
21:05 moritz japhb_: d'oh. I should try to reorder that
21:06 japhb_ :-)
21:06 moritz japhb_: which order would you prefer? date descending?
21:07 japhb_ People are used to newest at the top in blogs, so having highest lesson number at top, and lowest lesson number (the intro) at the bottom makes sense I think.
21:08 rurban hmm, I prefer the current order topdown
21:08 rurban it's not really blog-style, more tutorial-style
21:08 japhb_ (I would love to have it the other way around, to read linearly through, but I think that's counter to expectation.)
21:09 rurban --
21:09 bacek joined #parrot
21:09 japhb_ rurban: My personal preference is with you, I just wonder about the fact that it's being added to day-by-day, blog style (plus actually *being* part of a larger blog)
21:10 rurban as long as nobody actually complains
21:10 moritz japhb_: now it's ordered chronologically, maybe that'll change
21:10 japhb_ moritz: fine by me.  Any order is better than none.  :-)
21:13 moritz japhb: thanks for the feedback
21:13 japhb moritz: Oh, and damn fine tutorials, too.  :-)
21:14 * moritz follows most blogs through the RSS feed, so the order usually doesn't matter to him
21:14 moritz thanks ;)
21:14 japhb I like the way they are laid out, and have a nice bite-sized "less filling" sort of thing going.
21:15 japhb I think it works in a similar way to the original Learning Perl.  Just enough to get you hacking and interested in learning more.
21:15 japhb Back in the day, way too many blogs had constantly-breaking RSS feeds.  So I never got in the habit.  Perhaps I should readdress that.
21:37 cognominal_ joined #parrot
21:51 paco joined #parrot
21:56 rba joined #parrot
22:13 Ademan joined #parrot
22:27 gryphon joined #parrot
22:45 Limbic_Region joined #parrot
22:47 Ademan joined #parrot
22:47 ejs joined #parrot
23:15 tetragon joined #parrot
23:36 dmknopp joined #parrot
23:56 AndyA joined #parrot

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

Parrot | source cross referenced