Camelia, the Perl 6 bug

IRC log for #parrot, 2008-06-07

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:09 Limbic_Region joined #parrot
00:09 AndyA joined #parrot
00:21 diakopter TimToady: :P
00:25 cjfields joined #parrot
00:32 bacek \
00:36 kid51 joined #parrot
00:47 Maddingu1 joined #parrot
00:47 d4l3k_ joined #parrot
00:59 skv__ joined #parrot
01:01 skv__ joined #parrot
01:03 paco joined #parrot
01:03 clunker3 joined #parrot
01:15 grim_fandango joined #parrot
01:24 DietCoke token & <panic> ?
01:25 DietCoke ... not that that's the right syntax, of course. let me find something.
01:26 DietCoke 'HAI' <panic: 'Should never get here'>
01:26 DietCoke (that will die if it matches the literal HAI, methinks.
01:27 jhorwitz davidfetter: you can steal my time at YAPC.  the price is beer.  :)
01:27 davidfetter oh noes!
01:27 davidfetter what kind of beer?
01:27 jhorwitz the good kind
01:27 * davidfetter did *not* intend to steal, and was told there'd been a cancellation
01:28 davidfetter what kind of beer do you like?
01:28 jhorwitz all kinds.  i'm a fan of belgians in particular.
01:28 davidfetter k
01:36 cotto-work DietCoke, it's for php's little-known __halt_compiler() directive, which causes PHP's parser to completely ignore the rest of a script after that point
01:37 cotto-work http://www.sitepoint.com/blogs/200​6/05/12/__halt_compiler-how-nuts/
01:37 shorten cotto-work's url is at http://xrl.us/bmnws
01:38 cotto-work (it dresses like a function, but obviously it has to operate at a much lower level)
01:38 bacek joined #parrot
01:39 Infinoid cotto-work: could it be implemented as some fancy sort of comment?
01:40 Infinoid like # or // but more so...
01:40 cotto-work I feel like I should have thought of that
01:41 cotto-work Infinoid++ for stating the retrospectively obvious
01:42 Infinoid well, if it works, great :)
01:42 Infinoid I was just curious... I know nothing about PHP.
01:43 cotto-work It'll be a while before I can find out, but I can't poke a hole in it
01:43 cotto-work You're lucky that way.
01:43 cotto-work going home &
02:02 dalek r28147 | jkeenan++ | trunk:
02:02 dalek : Move print_tree() out of Parrot::Pmc2c::Pmc2cMain and into
02:02 dalek : Parrot::Pmc2c::PMC::PrintTree.  Eliminate '--tree' option from
02:02 dalek : tools/build/pmc2c.pl; replace it with a new program, tools/dev/pmc_tree.pl.
02:02 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28147
02:04 dalek r28148 | jkeenan++ | trunk:
02:04 dalek : Simplify program name:  get rid of the underscore.
02:04 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28148
02:07 dalek r28149 | coke++ | trunk:
02:07 dalek :
02:07 dalek : Remove the Version PMC. (RT#55366)
02:07 dalek : It was only referenced in its own instantiation
02:07 dalek : test and a few other infrastructure locations.
02:07 dalek : This changes the list of available core PMCs, so a reconfig is probably
02:07 dalek : required.
02:07 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28149
02:19 cotto_home Infinoid, it's sensitive to scope, since halting the compiler in a block or function definition (etc) would cause a syntax error
02:19 cotto_home It might be possible to make a scope-sensitive comment, though.
02:23 TiMBuS joined #parrot
02:32 dalek r28150 | jkeenan++ | trunk:
02:32 dalek : Transform t/tools/pmc2cutils/06-print_tree.t into t/tools/dev/pmctree.t.
02:32 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28150
02:33 dalek r28151 | jkeenan++ | trunk:
02:33 dalek : Update MANIFEST to reflect last file name change.
02:33 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28151
02:37 dalek r28152 | jkeenan++ | trunk:
02:37 dalek : print_tree() no longer needs to be tested here.
02:37 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28152
02:38 dalek r28153 | jkeenan++ | trunk:
02:38 dalek : Correct typo causing syntax error.
02:38 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28153
02:51 confound joined #parrot
02:52 DietCoke joined #parrot
02:53 dalek r28154 | jkeenan++ | trunk:
02:53 dalek : Test for correct time at which print_tree() can be run should have been when
02:53 dalek : 'parrot' has been created, not when 'Makefile' was created.
02:53 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28154
02:55 confound joined #parrot
03:03 DietCoke damnit, I can't 'svn commit' on feather because /usr/bin/editor is getting invoked and screwing things up. I just want 'vi', damnit.
03:05 dalek r28155 | coke++ | trunk:
03:05 dalek :
03:05 dalek : [codingstd]
03:05 dalek : add coda for perlcritic.t
03:05 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28155
03:06 DietCoke ... even using vi fails with svn: system('vi svn-commit.tmp') returned 256
03:07 DietCoke ... (P*#&$ TERM...
03:07 kid51 DietCoke:  Thanks for that fix.  I have to fix the pod formatting too.
03:10 cotto_home svn -m "your commit message here" ?
03:10 cotto_home svn commit -m "..."
03:10 dalek r28156 | jkeenan++ | trunk:
03:10 dalek : Correct POD formatting.
03:10 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28156
03:22 dalek r28157 | jkeenan++ | trunk:
03:22 dalek : Correct documentation re printing PMC inheritance tree.
03:22 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28157
03:22 dalek r28158 | coke++ | trunk:
03:22 dalek : [docs]
03:22 * kid51 must sleep
03:22 dalek : Remove reference to long dead weekly summaries
03:22 purl $kid51->sleep(8 * 3600);
03:22 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28158
03:24 Zaba joined #parrot
03:32 dalek r28159 | coke++ | trunk:
03:32 dalek : [docs]
03:32 dalek : RT #46711 - add links to google groups for RSS goodness.
03:32 dalek : As long as we're in here, delete the defunct google URLs, clean up a few
03:32 dalek : more and delete the obsolete VERSION information.
03:32 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28159
03:40 Andy joined #parrot
04:17 dalek r28160 | pmichaud++ | trunk:
04:17 dalek : [rakudo]:
04:17 dalek : * In scalar assignment, only call .item() on those objects that support it.
04:17 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28160
04:19 bacek pmichaud, good yesterday :)
04:19 pmichaud good tomorrow!
04:19 purl somebody said good tomorrow was mocha_haze's birthday and mocha_haze's new job sucks
04:19 pmichaud although it will be tomorrow in about 41 minutes.
04:19 * bacek thinks that pmichaud is robot. He never sleeps... :)
04:21 pmichaud I do keep weird hours.
04:21 pmichaud always have.
04:21 bacek pmichaud, I have patch for 'eval()' and 'eval_dies_ok'.
04:21 pmichaud send it to rakudobug, please
04:22 * bacek fear imagine pmichaud's 'weird hours'
04:22 bacek pmichaud, This mailinglist does not accept postings crossposted to newsgroups - Contact perl6-internals-owner@perl.org for help
04:22 bacek What I did wrong?
04:22 pmichaud should just send it to rakudobug@perl.org .  That doesn't work?
04:23 bacek ok, I'll try again.
04:23 pmichaud or, if there's an existing ticket, reply to one of the messages in the ticket.
04:23 bacek ok.
04:24 bacek btw, S29-list/grep.t works for me. Add it to spectest_regression?
04:24 pmichaud it's not already there?
04:24 pmichaud hmmm, I guess not.  Let me run it quickly.
04:25 pmichaud I get "2 subtests UNEXPECTEDLY SUCCEEDED"
04:26 bacek TODO?
04:26 purl i think TODO is at http://www.AmbrosiaSw.com/A​mbrosia_Products/ToDo.html
04:26 pmichaud yes, that means two TODO tests are passing.
04:27 bacek it about 'mutating map/grep'
04:27 pmichaud ok 28 - mutating $_ in grep works (1)# TODO unspecced
04:27 pmichaud ok 29 - mutating $_ in grep works (2)# TODO unspecced
04:28 bacek pmichaud, is there any decision for 'bugward compatibility' about mutating map()?
04:29 pmichaud not that I've seen.  We may need to repost the question.
04:29 pmichaud anyway, I'll add it to regression.
04:29 pmichaud it'll be more likely to get a decision.  :-)
04:29 bacek :)
04:30 dalek r28161 | pmichaud++ | trunk:
04:30 dalek : [rakudo]:
04:30 dalek : * Add t/S29-list/grep.t to spectest_regression (bacek++)
04:30 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28161
04:33 Eevee perhaps this is insane, but microsoft's recent silverlight flailing makes me wonder
04:33 Eevee how feasible would it be to someday have parrot in a browser, either as core or a plugin
04:33 pmichaud lots of things about microsoft make me wonder recently.  :-)
04:33 pmichaud having parrot in a browser is something we'd like to see, yes.
04:33 pmichaud so feasible, yes.
04:36 bacek pmichaud, question about lazy list: should we separate classes for 'lazy' and 'eager' lists?
04:38 Eevee actually, hm.  as core, you'd have a hell of a time getting IE to implement it.  as a plugin, you'd need a specific canvas and controls and ways to draw to it and etc
04:46 pmichaud bacek: I wasn't planning on having a separate class for eager lists.
04:47 pmichaud unless there's a class for it already in the synopses :-)
04:48 bacek pmichaud, no. Such class doesn't exists :)
04:49 bacek But 'lazy-list' is quite different from 'eager-list'.
04:49 pmichaud why is that?
04:50 bacek "It's hard to be lazy" :)
04:50 bacek Actually 'lazy' should morph self into eager on 'set_pmc_*'.
04:50 pmichaud you mean on set_keyed_pmc ?
04:50 pmichaud sorry, set_pmc_keyed_* ?
04:51 bacek indeed
04:51 pmichaud for that we only need to be eager enough on the elements up to the key
04:51 pmichaud we don't have to make the entire list eager.
04:52 pmichaud i.e.,   my  @b = 1..*;  @b[5] = 'five';  should still work
04:52 bacek what about grep { $_%2}, @list? I think 'grep' should produce lazy list too
04:52 pmichaud yes, grep should be lazy also
04:53 pmichaud grep should return an object that when iterated (i.e., shifted) produces the next element in the sequence
04:53 bacek ok. what about get_pmc_keyed?
04:53 pmichaud same thing -- get_pmc_keyed only needs to be eager enough to produce the element we want
04:53 bacek eg (grep{}, @list)[5]?
04:54 Eevee pmichaud: is that treated as "1..* except for index 5", or "1, 2, 3, 4, 5, five, 7..*"?
04:54 bacek so you thinks get_pmc_keyed should materialize list?
04:54 pmichaud Eevee: the list ends up being 1, 2, 3, 4, 5, five, 7..*
04:54 davidfetter joined #parrot
04:54 pmichaud (iiuc)
04:54 bacek say (grep{}, 1..Inf)[1000000000]
04:55 bacek out-of-memory exception thrown...
04:55 Eevee okay; never saw that mentioned explicitly in a synopsis so I wasn't sure
04:55 pmichaud bacek:  that might not even throw an exception
04:55 bacek segmentation fault, core dumped :)
04:56 bacek (on side note) anyone tried to adapt t/02-test-pm/ for rakudo? It seems very pugs specific.
04:58 pmichaud I think it's somewhat pugs specific.  I'd like to see it rewritten for rakudo (with the assumption that Rakudo's Test.pm is more universal)
04:59 pmichaud (the assumption may be wrong, but I prefer Rakudo's very simple Test.pm to Pugs' more complex one.)
04:59 bacek pmichaud, +1
05:06 bacek is it possible to override 'proclaim' in Test.pm from external program?
05:06 bacek pmichaud, good today :)
05:06 pmichaud I dunno.
05:07 pmichaud (proclaim)
05:09 bacek hmm... Probably adding 'sub set_proclaim(Sub procaim)' is not best idea, but it may works
05:09 pmichaud why override proclaim?
05:09 bacek to check that 'ok(2+2==5) actually works?
05:10 pmichaud might also try:   my &proclaim = sub ....
05:10 pmichaud or   my &proclaim := sub ....
05:10 bacek ... and add 'set_proclaim' in Test.pm :)
05:10 pmichaud although that would just do it lexically, and not for Test.pm's ok() function.
05:11 pmichaud maybe  &Test::proclaim would work.
05:12 pmichaud although I don't think we support namespace'd sigiled vars yet.
05:12 pmichaud although we might.  :-)
05:15 bacek future is bright. Only one problem - we lives now :)
05:21 bacek sub s1 { say 's1'; }; sub s2 { say 's2'; };
05:21 bacek our &s := &s1; &s(); &s = &s2; &s();
05:21 bacek segmentation fault.
05:21 purl (Core dumped)
05:21 cotto_home future?
05:21 purl You will waste much time on IRC
05:21 cotto_home future?
05:21 purl You will waste much time on IRC
05:21 cotto_home future?
05:21 purl Ask again later.
05:22 bacek first is ':=', second '='. It causing infinite loop in 'returncc'.
05:28 nopaste "bacek" at 202.7.166.165 pasted "It works: testing Test.pm with proclaim overriding." (20 lines) at http://nopaste.snit.ch/13200
05:30 dalek r28162 | pmichaud++ | trunk:
05:31 dalek : [rakudo]:
05:31 dalek : * Rework handling of named unaries a bit.
05:31 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28162
05:52 bacek Ho.
05:53 bacek I've implemented test for Test,pm
06:19 DietCoke Feedback on 52776 appreciated.
06:21 Psyche^ joined #parrot
06:22 silug joined #parrot
06:23 * bacek will never again build parrot with --optimize --inline for developing rakudo...
06:54 pmichaud DietCoke: (#52776)  +1
07:02 cotto_home joined #parrot
07:09 dalek r28163 | pmichaud++ | trunk:
07:09 dalek : [rakudo]:
07:09 dalek : * Make C<pop> and C<shift> into named unaries.
07:09 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28163
07:13 moritz pmichaud: re try <statement>*, does this ever DWIM?
07:13 moritz oh no, it was {*}, not *
07:13 pmichaud moritz: it works for things like     try { ... };
07:13 * moritz tries to wake up entirely
07:14 moritz is type checking guarantueed to be at run time?
07:15 moritz I mean a clever compiler could try to do some type inference and prove at compile time that there well be a clash
07:15 moritz so shouldn't it be eval_lives_ok in the test case anyway?
07:16 pmichaud I don't know.  That might be a good followup to my original post.
07:16 moritz ok, I'll write something
07:21 pmichaud anyway, sleep time
07:21 pmichaud bbl
07:22 bacek moritz, I don't think, that even clever compiler should check content of string...
07:22 bacek good night, pmichaud
07:22 moritz bacek: did you read the mail on p6c?
07:22 bacek moritz, which one?
07:23 bacek btw, #55438 :)
07:23 moritz t/spec/S02-builtin_data_types/type.t and 'try' statement
07:23 bacek in process
07:25 bacek yeah... There is a _LOT_ of such usage of try/grep/whatever...
07:25 iblechbot joined #parrot
07:25 bacek I just thinks, that priority of 'infix:,' is too high
07:27 * bacek still thinks that this handled in overcomplicated EXPR rule...
07:27 moritz are there still statement likes like 'a(), b() while c()' in perl 6?
07:27 bacek moritz, no idea.
07:28 bacek Probably EXPR/expect_term combo handles this, adverbial closure and much more...
07:31 bacek moritz, can you make me a favour? Test test.t/Test.pm from #55438?
07:32 bacek (most recent test.t from ticket :)
07:33 moritz don't you mean #55304?
07:34 bacek moritz, no :)
07:35 bacek but there is a patch for 'eval()' in 55304. It required to apply before test.
07:35 bacek eval2.diff
07:37 bacek If test passed, try to comment out line 119 in Test.pm and rerun test... I've got very nasty error...
07:37 moritz ok, wait a sec
07:38 moritz don't you want to publish a git mirror that I can simply track?
07:39 moritz that would simplify all that patching stuff ;)
07:39 bacek moritz, hmm... It will take some time. But I can do it if you really want :)
07:40 bacek afk, coffee break :)
07:40 moritz that was more a "this patching is so cumbersome" than a real suggestion ;)
07:49 dalek r28164 | allison++ | trunk:
07:49 dalek : [cage] Update the MacPorts description to match other descriptions.
07:49 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28164
07:49 * bacek broke coffee and switched to wine...
07:50 moritz bacek: dunno why, your test.diff didn't apply cleanly for me
07:51 moritz (or did I try to apply it twice? don't think so...)
07:51 bacek There is clean Test.pm attached :)
07:51 bacek s/clean/patched/
07:53 moritz your test.t fails for me with ok 16 - ... it eval_lives_ok
07:53 moritz set_pmc_keyed_str() not implemented in class 'Eval'
07:53 moritz current instr.: 'fake_proclaim' pc 1391 (EVAL_13:414)
07:56 bacek shit...
08:06 Ivatar joined #parrot
08:09 nopaste "bacek" at 202.7.166.165 pasted "eval_lives_ok replacement for moritz" (7 lines) at http://nopaste.snit.ch/13202
08:11 bacek moritz, can you replace eval_lives_ok in Test.pm with nopasted one?
08:11 moritz nah
08:11 moritz that one is wrong
08:11 moritz either eval can propagate excpetions
08:11 moritz than it could happen that a call to eval_lives_ok doesn't call proclaim() at all
08:11 moritz or it can't
08:11 moritz then the try block is superflous
08:12 bacek no, its looks like stack corruption.
08:12 bacek so I just try to find exact place where it happens.
08:14 moritz now it runs the tests up to 16
08:14 moritz ok 16 - ... it eval_lives_ok
08:14 moritz # Looks like you planned 17 tests, but ran 16
08:15 bacek shit... It dies inside of parrot...
08:16 bacek can you comment out line 119 in Test.pm? It will die early.
08:17 moritz not ok 10 - ... it dies_ok
08:17 moritz with 119 commented out
08:22 bacek moritz, i expected it...
08:59 bacek moritz, i cought this bug!
08:59 moritz sorry, have to run... just write on, I'll backlog
09:02 bacek moritz, perl6: try { die }; say (defined $!) xor Bool::True
09:03 bacek rakudo outputs nothing...
09:45 moritz re
09:46 moritz is xor tested somewhere? and does 'make fulltest' run those tests?
09:46 moritz if not, don't rely on it ;)
09:50 bacek moritz, it works for just Bool.
09:50 bacek check last my tests in #perl6...
09:51 dalek allison@perl.org | Articles of Incorporation:
09:51 dalek link: http://www.perlfoundation.org/parrot​/index.cgi?articles_of_incorporation
09:51 shorten dalek's url is at http://xrl.us/bkyfj
09:52 bacek can you try replace proclaim(defined...) with proclaim(0+defined...)
09:52 bacek ?
09:52 bacek afk,
09:55 dalek allison@perl.org | Bylaws:
09:55 dalek link: http://www.perlfoundation.o​rg/parrot/index.cgi?bylaws
10:00 barney joined #parrot
10:05 dalek allison@perl.org | Bylaws:
10:05 dalek link: http://www.perlfoundation.o​rg/parrot/index.cgi?bylaws
10:06 dalek allison@perl.org | Bylaws:
10:06 dalek link: http://www.perlfoundation.o​rg/parrot/index.cgi?bylaws
10:32 dalek allison@perl.org | Bylaws:
10:32 dalek link: http://www.perlfoundation.o​rg/parrot/index.cgi?bylaws
10:59 jan joined #parrot
11:05 skv__ joined #parrot
11:36 dalek allison@perl.org | Bylaws:
11:36 dalek link: http://www.perlfoundation.o​rg/parrot/index.cgi?bylaws
12:05 kid51 joined #parrot
12:12 Zaba_ joined #parrot
12:19 davidfetter joined #parrot
12:41 dalek allison@perl.org | Bylaws:
12:41 dalek link: http://www.perlfoundation.o​rg/parrot/index.cgi?bylaws
13:13 Andy joined #parrot
13:27 iblechbot joined #parrot
13:46 DietCoke joined #parrot
13:48 diakopter eee
13:54 masak joined #parrot
13:57 pmichaud particle++ allison++ # parrot foundation documents
13:58 bacek pmichaud, goot today :)
13:59 pmichaud good morning, afternoon, or other localtime
14:00 * bacek thinks is 'good now' is good enough :)
14:03 bacek Hey! It's tomorrow already! :)
14:03 bacek 00:03
14:03 smash joined #parrot
14:04 smash hello everyone
14:04 smash DietCoke: ping
14:28 tetragon joined #parrot
15:07 dalek r28165 | fperrad++ | trunk:
15:07 dalek : fix RT #55332
15:07 dalek : On Windows, unlink is not POSIX compliant :
15:07 dalek :   unlink on an opened file does nothing (now & after closing).
15:07 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28165
15:23 pmichaud what would I need to do to get languages/perl6/t/harness to honor the -v option?
15:33 Theory joined #parrot
16:31 Andy joined #parrot
16:40 ruoso joined #parrot
17:18 khisanth_ joined #parrot
17:20 Auzon joined #parrot
17:24 nopaste joined #parrot
17:32 dalek r28166 | fperrad++ | trunk:
17:32 dalek : [Lua]
17:32 dalek : - factorization in lua_loadfile
17:32 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28166
17:41 cognominal joined #parrot
17:52 ruoso joined #parrot
17:56 teknomunk joined #parrot
18:00 dalek r28167 | fperrad++ | trunk:
18:00 dalek : [Lua]
18:00 dalek : - lua_loadfile : give a chunkname when stdin
18:00 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28167
18:06 Andy joined #parrot
18:16 Zaba joined #parrot
18:41 gryphon joined #parrot
19:28 moritz pmichaud: re lazy and eager lists, 'List' is lazy, 'Seq' is eager
19:30 pmichaud I did think of that, but I tend to think of Seq as simply "completely evaluated".  At any rate, I'm not sure I want List to automatically morph to Seq
19:30 moritz what's the difference between eager and "completely evaluated"?
19:31 moritz is the latter "recursively eager"?
19:31 pmichaud I don't know that there is a difference, but I'm not yet ready to claim they're exactly the same.  They may be, but I haven't thought that far ahead yet :-)
19:31 moritz ok
20:11 Limbic_Region joined #parrot
20:19 dalek r28168 | fperrad++ | trunk:
20:19 dalek : [Lua]
20:19 dalek : - lua_load handles binary chunk via bytecode translation
20:19 dalek : (don't work, namespace problem)
20:19 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28168
20:20 gmansi joined #parrot
20:35 Ademan joined #parrot
20:40 davidfetter joined #parrot
20:42 kid51 joined #parrot
20:55 guru joined #parrot
20:56 * kid51 has spent the day reinstally Mac OS X 10.4 on his old iBook G4.  Had to reinstall Xcode Tools, rebuild Perl 5 and Python 2.5, re-install Subversion.  All due to file corruption in Software Update.
21:00 * cizra has spent all the day installing Debian on his old G3 PowerMac.
21:00 cizra :-p
21:01 * pmichaud has spent all day working on various parts of rakudo.
21:01 pmichaud (mostly running spectests)
21:01 * moritz has spent nearly all day (5 hrs) playing table tennis
21:01 * tetragon has left her iBook G4 at OS X 10.5.2 until enough disk space has been freed up for the combo update for 10.5.3 to download
21:06 * tetragon is also noting down all sorts of details needed to get a Toughbook's touchscreen to work under Linux
21:09 iblechbot joined #parrot
21:31 * diakopter spent all day working on yap6.  gigantic commit message forthcoming.
21:44 Zaba_ joined #parrot
21:48 bacek morning
21:49 moritz morning is in about 9 hours ;)
21:49 bacek moritz, no... It's already 7:49...
21:50 moritz ok then
21:50 moritz morning
21:50 moritz ;)
21:50 Infinoid happy now :)
21:50 Infinoid and happy 1 moment from now too (accounting for net lag)
21:52 kid51 tetragon:  I spent last Saturday at the Apple Genius Bar getting this Software Update problem diagnosed.  The 'Genius' said that there wouldn't be enough memory on my iBook G4 to install 10.5.  So I'll have to buy an Intel Mac to get 10.5.
22:03 tetragon kid51: And if you were to buy an Intel with 10.5, it would help complete the set of architectures and OS releases
22:04 pmichaud or someone could donate one to kid51.  :-)
22:04 * tetragon digs out a Powerbook with pre-OS X Mac OS
22:05 * jonathan spent the day at a BBQ and then watching the first two matches of Euro 2008.
22:05 tetragon And I don't think any of the OS X releases would run on that Powerbook (just on the ppc side of the m68k -> ppc switchover)
22:06 davidfetter joined #parrot
22:07 bacek summon pmichaud to discuss #55446...
22:07 nopaste "pmichaud" at 76.183.97.54 pasted "Rakudo spectest regression history" (29 lines) at http://nopaste.snit.ch/13204
22:08 gmansi joined #parrot
22:08 bacek pmichaud, good progress!
22:08 pmichaud bacek: yes, thanks to many contributors.
22:09 bacek pmichaud, you very welcome :) (It's all vany fault. He'd convenient me :)
22:10 bacek pmichaud, I need help with 55446...
22:10 bacek something very strange happens with handling exceptions in parrot.
22:11 dalek r28169 | pmichaud++ | trunk:
22:11 dalek : [pge]:
22:11 dalek : * Improvement to OPTable so that word tokens also look
22:11 dalek :   for word boundaries (e.g., so that 'if' doesn't match 'ifonly').
22:11 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28169
22:16 dalek r28170 | pmichaud++ | trunk:
22:16 dalek : [rakudo]:
22:16 dalek : * Add a preliminary implementation of Complex.
22:16 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28170
22:20 dalek r28171 | pmichaud++ | trunk:
22:20 dalek : [rakudo]:
22:20 dalek : * Add 'test_summary.pl' script used to provide an alternate view
22:20 dalek :   of running spectest_regression (and generate the regression history).
22:20 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28171
22:25 braceta joined #parrot
22:27 kid51 tetragon:  I suspect we have *many* people who have Intel Macs with 10.5.  I just haven't identified them for the purpose of testing your patches (and those of others).
22:30 kid51 One positive side effect of losing /usr/local:  I lost that botched attempt to build my own gcc that required me to specify 4 options to get Configure.pl to work.  For the first time ever, all I did to build was call 'perl Configure.pl'.  'make' and 'make coretest' all passed.
22:30 bacek how I can curry function in PIR?
22:31 * kid51 watches Belmont
22:34 kid51 No Triple Crown this year; winner at 38-1 odds
22:35 tetragon kid51: Does Apple still ship Perl on 10.4 without hidden, nasty -arch flags?
22:36 tetragon My 10.4 box hasn't been updated in a while
22:37 kid51 How would I determine that?  Since this was a reinstall from DVD, it meant that I got the same Perl 5.8.6 that came with this originally 4 years ago.
22:37 tetragon Take a look at the output of perl -V
22:38 tetragon ccflags is one of the values that on 10.5 has the -arch flags in it
22:38 tetragon (10.5 comes with 5.8.8)
22:39 nopaste "kid51" at 68.237.19.81 pasted "The Perl 5.8.6 that came with Mac OS X 10.4.6" (55 lines) at http://nopaste.snit.ch/13205
22:39 tetragon What I suspect is that Apple built Perl differently between 10.4 and 10.5
22:39 bacek summon pmichaud
22:39 bacek spell failed...
22:39 purl failed... looks funny
22:39 kid51 Yes, I believe you demonstrated that a couple of months back.
22:39 * bacek cast fireball on purl
22:40 tetragon 10.5 is a full multiarch build, 10.4 had both an i386 and a ppc build done separately and glued together with lipo
22:40 tetragon And parrot's build system would need some alterations to handle the all at once multiarch
22:41 kid51 No -arch flags on this Apple build.
22:41 tetragon Still no -arch flags
22:41 kid51 I think 10.4.6 was still in the ppc-only days.
22:43 tetragon My Intel's perl doesn't have -arch flags, and that box's initial OS X was either 10.4.6 or 10.4.7
22:46 pmichaud pong
22:48 bacek pmichaud, hi again. Is it possible to curry sub in PIR?
22:48 Theory_ joined #parrot
22:49 pmichaud bacek: it will be, yes.  We're not doing it yet.
22:52 bacek pmichaud, ok...
22:53 bacek so it's not I'm so stupid. It's just not yet implemented :)
22:53 pmichaud right.
22:53 bacek pmichaud, second question: #55446
22:53 bacek something wrong with exception/variables handling.
22:53 pmichaud right, we haven't corrected $_, $/, and $! yet.
22:55 pmichaud but what did you expect instead of Undef ?
22:56 bacek Bool...
22:56 pmichaud xor doesn't return a Bool.
22:56 bacek Ok, Int. Something but not undef.
22:57 bacek rakudo: try{die}; say defined $!; say (defined $!) xor 1;
22:57 bacek it return [1\n\n]
22:57 pmichaud oh.
22:57 bacek pmichaud, exactly...
22:58 pmichaud drat.  The spec changed on me:  # infix:<xor>, exclusive or
22:58 pmichaud $x xor $y
22:58 pmichaud Returns the true argument if there is one (and only one). Returns Bool::False if both arguments are false or both arguments are true. In list context forces a false return to mean (). See ^^ above for high-precedence version.
22:59 bacek pmichaud, problem not with 'xor'. Problem with variables...
22:59 bacek <bacek> rakudo: say (defined $!) xor 1;
22:59 bacek <p6eval> rakudo r28171 OUTPUT[1␤]
22:59 pmichaud 22:53 <pmichaud> right, we haven't corrected $_, $/, and $! yet.
22:59 pmichaud although what you wrote there looks right to me.
23:00 pmichaud unless $! is supposed to be undef... okay.
23:00 cotto_home what's a good example of what a language spec should look like?
23:00 pmichaud actually, what you wrote there looks write to me.
23:00 pmichaud s/write/right/
23:01 bacek <bacek> rakudo: say undef xor 1;
23:01 bacek <p6eval> rakudo r28171 OUTPUT[1␤]
23:01 pmichaud correct.
23:01 jan joined #parrot
23:01 * bacek still dont understand how 'xor' can return undef...
23:02 pmichaud rakudo:  say (1 xor 1).WHAT;
23:02 pmichaud ww
23:02 bacek wrong channel :)
23:04 * davidfetter mischans to keep bacek company
23:10 nopaste "pmichaud" at 76.183.97.54 pasted "specte_regression results in head" (62 lines) at http://nopaste.snit.ch/13206
23:15 bacek pmichaud, why you 'rewrite' xor in PAST/Compiler.pir instead of invoking xor directly???
23:15 pmichaud xor is short circuiting
23:15 pmichaud a xor b xor c  is not the same as  ((a xor b) xor c)
23:16 bacek ouch...
23:16 bacek and how it should work???
23:16 purl somebody said how it should work was that M::B should generate enough stuff for ...
23:17 pmichaud a xor b xor c   returns the true element if there's exactly 1, otherwise it returns Bool::False
23:19 bacek pmichaud, pugs disagree with you :)
23:19 pmichaud pugs is wrong then.
23:20 pmichaud or the spec changed.
23:20 pmichaud hrm.
23:20 bacek in which spec xor defined? I cant find it in S29...
23:20 pmichaud it's an operator, so S03
23:21 bacek pmichaud, thanks. Looks like spec really changed.
23:26 pmichaud bacek:  here's what I was working from:  http://groups.google.com/group/perl.​perl6.language/msg/36b680b042a3aece
23:26 shorten pmichaud's url is at http://xrl.us/bmow7
23:29 pmichaud there's also:  http://groups.google.com/group/perl.​perl6.language/msg/fae98f310e4dfcf4
23:29 shorten pmichaud's url is at http://xrl.us/bmoxb
23:30 bacek to trust Larry or not to trust? This is the question!
23:30 bacek :)
23:31 pmichaud well, S03 was modified to have its current definition of C<xor> in 2007, so I presume the later interpretation rules.
23:31 pmichaud Which would mean I can tear out the 'xor' code altogether from PCT.
23:31 pmichaud which would be nice.
23:32 pmichaud but I'll ask on p6c.
23:33 dalek r28172 | jkeenan++ | trunk:
23:33 dalek : Eliminate trailing white space at line 415.
23:33 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28172
23:46 pmichaud message sent
23:46 purl Sorry, I've never seen sen before.
23:46 pmichaud time for shopping.
23:47 pmichaud ...."sen"?

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

Parrot | source cross referenced