Camelia, the Perl 6 bug

IRC log for #parrot, 2008-02-22

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:04 AndyAway TICKET 47109 TOTALLY HAS MY NAME ON IT!!!!11!!1
00:04 particle c:\usr\local\parrot\clean\l​anguages\perl6>..\..\parrot perl6.pbc --target=past
00:04 particle say 1;
00:04 particle ---
00:04 particle {
00:04 particle "past" : !PAST::Block { ... },
00:04 particle }
00:04 particle oh, i'm *SO CLOSE*
00:05 particle ...but, naptime &
00:05 wknight8111 naptime? I agree. this exception thing has me completely confused
00:07 chromatic I could put the name Andy on several other tickets.
00:07 wknight8111 i can't find a single example of how to set the ret_addr parameter to real_exception
00:07 Andy chromatic like what?
00:08 Andy I LIKE INSTALLING SEAT BELTS LOL!
00:12 slightlyoff joined #parrot
00:13 chromatic Anything in src/string.c for example.
00:20 x joined #parrot
00:22 Andy chromatic: If you want, go ahead and assign stuff.
00:28 chromatic joined #parrot
00:42 svnbotl r25960 | petdance++ | trunk:
00:42 svnbotl : parenthesizing macros
00:42 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25960
00:54 Tene Hmm.  Looks like it's not accepting my password.
00:54 svnbotl r25961 | chromatic++ | trunk:
00:54 svnbotl : [Lua] Cleaned up LuaBoolean PMC.
00:54 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25961
00:55 Tene Trying to check in lolcode fixes.
00:56 * Tene tries 'reset password' on bitcard.
01:05 * Tene pings Coke.
01:07 svnbotl r25962 | chromatic++ | trunk:
01:07 svnbotl : [Lua] Cleaned up LuaUserdata PMC.
01:07 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25962
01:09 Coke tene;pong
01:10 Tene Coke: Seems I can't commit.  afaict, my password isn't being accepted.
01:10 * Coke jiggles in anticipation of chromatic going through tcl. =-)
01:11 Tene Tried generating a new one on svn.perl.org/account/
01:11 Coke tene; what does "svn info" tell you.
01:12 nopaste "tene" at 71.16.144.82 pasted "svn info" (11 lines) at http://nopaste.snit.ch/12380
01:12 chromatic Hm, another memory bug in Lua.
01:12 chromatic Hm, maybe not.
01:13 slightlyoff joined #parrot
01:14 Coke Tene: your rt id is 'tene@allalone.org', neh?
01:14 Tene Right.
01:15 Tene svn: Commit failed (details follow):
01:15 Tene svn: MKACTIVITY of '/parrot/!svn/act/12ed5e30-0​2bb-480f-8347-448e149bd987': 403 Forbidden (https://svn.perl.org)
01:16 Coke No clue. please send an email to webmaster@perl.org with your id, and a copy of 'svn info' and what happens when you do 'svn commit'.
01:16 Coke cc will@coleda.com
01:16 Coke sorry. :|
01:20 Tene sent
01:21 Coke Andy++ # macro cleanup
01:21 Andy I took an entire ticket
01:21 Andy I WILL OWNZOR THE PARENS
01:26 svnbotl r25963 | coke++ | trunk:
01:26 svnbotl : [codingstd]
01:26 svnbotl : Set svn properties on new files.
01:26 svnbotl r25964 | chromatic++ | trunk:
01:26 svnbotl : [Lua] Cleaned up Lua PMC.
01:26 svnbotl r25966 | coke++ | trunk:
01:26 svnbotl : [codingstd]
01:27 svnbotl : Fixup PIR coda, damaged when language was regenerated
01:27 svnbotl : RT #51086, Courtesy mkraai at beckman dot com
01:27 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25966
01:32 Coke ur... your svn login and your rt id are different?
01:32 Tene Hm?
01:33 Coke your svn login and your rt login are different.
01:33 Coke that doesn't seem right.
01:33 Tene I don't know.
01:33 Tene "Your username for Subversion/WebDav is tene"
01:34 Coke we'll see what robert says.
01:35 Tene Should I try committing with the username 'tene@allalone.org'?
01:35 Coke couldn't hoit.
01:36 Tene nope.
01:36 Tene fails.
01:36 Coke ah well
01:58 svnbotl r25967 | chromatic++ | trunk:
01:58 svnbotl : [Lua] Cleaned up LuaTable PMC.
01:58 svnbotl : Fixed one typo in LuaAny PMC.
01:58 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25967
02:00 stephen joined #parrot
02:02 cotto joined #parrot
02:03 cotto is there some nicer for printing STRINGs than that string_to_cstring/string_cstring_free stuff?
02:05 particle sprintf( "%S", string );
02:08 cotto that doesn't work for parrot strings unless I'm missing something
02:09 svnbotl r25968 | coke++ | trunk:
02:09 svnbotl : [codingstd]
02:09 svnbotl : Test was just checking for #define, when #  define is also allowed.
02:09 svnbotl : Failures now up to 3248. Andy better hurry up before I find more failures!
02:09 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25968
02:11 chromatic %Ss
02:11 chromatic but you need Parrot_sprintf_* for that.
02:12 andy_gav joined #parrot
02:12 chromatic If you're just debugging and printing temporarily, use s->strstart, but don't remember I told you that and scrub it out before checkin.
02:12 particle oh, sorry. just awoke from my nap.
02:13 cotto I don't trust myself to remember that.
02:14 andy_gav left #parrot
02:15 svnbotl r25969 | chromatic++ | trunk:
02:15 svnbotl : [Lua] Cleaned up LuaNumber PMC.
02:15 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25969
02:24 particle reboot time &
02:25 Andy coke: I allowed bracketed macro args, too.
02:26 Coke I asked here and was told that wasn't safe.
02:26 Coke but IANACP.
02:26 svnbotl r25970 | petdance++ | trunk:
02:26 svnbotl : Allow brackets instead of parens around macro args
02:26 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25970
02:27 Coke So I will let you and chromatic jello-wrestle that one out.
02:28 chromatic_away Not safe?
02:28 Andy Huh.
02:28 Andy He was on drugs.
02:28 Andy Severe, chocolate-flavored drugs
02:29 * Coke add a line of code. "The bedroom is a room."
02:29 Andy lots of these I can't test.
02:33 svnbotl r25971 | petdance++ | trunk:
02:33 svnbotl : wrapper macro args
02:33 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25971
02:37 Coke someone needs to resurrect zcode so we can play inform7 games.
02:46 svnbotl r25972 | petdance++ | trunk:
02:46 svnbotl : wrapped macro arguments
02:46 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25972
02:54 wknight-away joined #parrot
02:56 Ademan joined #parrot
02:59 svnbotl r25973 | petdance++ | trunk:
02:59 svnbotl : lots of macro wrapping
02:59 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25973
03:19 particle joined #parrot
03:20 particle gee, that was fun.
03:36 Andy joined #parrot
03:42 Andy oh look someone closed emmory leaks
03:43 kid51 joined #parrot
04:02 svnbotl r25974 | jkeenan++ | trunk:
04:02 svnbotl : Per http://rt.perl.org/rt3/Tic​ket/Display.html?id=38262, removing module which comes with Perl 5.8+ from Parrot repository/distribution.
04:02 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25974
04:03 Theory joined #parrot
04:08 AndyA joined #parrot
04:52 Theory joined #parrot
05:40 mdiep joined #parrot
06:06 svnbotl r25975 | petdance++ | trunk:
06:06 svnbotl : parenthesize macro arguments, and move the macro into the one source file where it gets used
06:06 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25975
07:13 uniejo joined #parrot
07:31 jisom joined #parrot
08:04 svnbotl r25976 | chromatic++ | trunk:
08:04 svnbotl : [Lua] Cleaned up LuaString PMC.
08:04 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25976
08:10 c_otto What's the best way to generate a patch that includes new files against svn?
08:11 HG` joined #parrot
08:11 particle svn add foo
08:11 particle svn diff
08:11 purl svn diff is clean
08:12 c_otto thanks
08:12 particle bedtime &
08:13 iblechbot joined #parrot
08:29 IllvilJa joined #parrot
08:41 svnbotl r25977 | fperrad++ | trunk:
08:41 svnbotl : [Lua]
08:41 svnbotl : - same as r25930 but in tobase
08:41 svnbotl r25978 | fperrad++ | trunk:
08:41 svnbotl : [Lua]
08:41 svnbotl : - put environment initialization in LuaUserdata constructor
08:41 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25978
08:49 mire joined #parrot
08:57 particle joined #parrot
09:13 svnbotl r25979 | kjs++ | trunk:
09:13 svnbotl : [rakudo] update STATUS: until/while/repeat has been implemented.
09:13 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25979
09:32 svnbotl r25980 | kjs++ | trunk:
09:32 svnbotl : [docs] Remove docs/imcc/macros.pod; all relevant info can be found in pdd19_pir.pod
09:32 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25980
09:46 davidfetter joined #parrot
10:03 svnbotl r25981 | kjs++ | trunk:
10:03 svnbotl : [docs] remove reference to macros.pod, which has been removed. add reference to pdd19.
10:03 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25981
10:15 c_otto joined #parrot
10:19 davidfetter joined #parrot
10:40 svnbotl r25982 | kjs++ | trunk:
10:40 svnbotl : [c99] add rules for pre-processor. I think this should become a different grammar altogether to keep things simple.
10:40 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25982
10:54 marmic joined #parrot
11:00 cognominal_ joined #parrot
11:18 GeJ_ joined #parrot
11:55 svnbotl r25983 | kjs++ | trunk:
11:55 svnbotl : [c99] start on actions.pm refactor grammar a bit for this.
11:55 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25983
11:57 dwave joined #parrot
12:21 svnbotl r25984 | kjs++ | trunk:
12:21 svnbotl : [c99] Hello world in C runs on Parrot!
12:21 svnbotl : update grammar and actions.pm
12:21 svnbotl : add myself to MAINTAINER
12:21 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25984
12:27 svnbotl r25985 | kjs++ | trunk:
12:27 svnbotl : [c99] update test harness for c99, and write new first test.
12:27 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25985
12:36 kj joined #parrot
12:40 svnbotl r25986 | kjs++ | trunk:
12:40 svnbotl : [c99]
12:40 svnbotl : - add if, while and do-while statements
12:40 svnbotl : - add tests for this.
12:40 svnbotl : - make compound-statements create blocks, instead of stmts nodes; they define a new scope.
12:40 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25986
12:47 svnbotl r25987 | kjs++ | trunk:
12:47 svnbotl : [c99]
12:47 svnbotl : /simplify TOP action a bit
12:47 svnbotl : /add a test for while
12:47 svnbotl : /add the news about c99 to NEWS.
12:57 avar joined #parrot
13:19 Coke kjs++
13:33 Coke you should add the sample c file that works under examples/ !
13:46 Coke (ah, that's make test.)
13:54 * Coke tries to decide if he should comment on the "perl6 fundraising" thread.
13:55 Andy joined #parrot
13:56 svnbotl r25988 | tene++ | trunk:
13:56 svnbotl : Lots of fixes for lolcode.
13:56 svnbotl : Properly handle dispatching arguments to functions at runtime.
13:56 svnbotl : Allow definition of functions with parameters.
13:56 svnbotl : Miscellenous cleanups.
14:06 Patterner joined #parrot
14:09 * Coke waits for tene's commit to hit the list...
14:11 Coke nice lolcode improvements. =-)
14:17 Limbic_Region joined #parrot
14:17 Limbic_Region salutations all
14:21 svnbotl r25989 | coke++ | trunk:
14:21 svnbotl : [codingstd]
14:21 svnbotl : Minor updates for new files in distro.
14:21 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25989
14:26 cognominal_ when is the cut date for 0.5.3?
14:27 Davemon joined #parrot
14:42 jhorwitz joined #parrot
14:45 orbiscerbus joined #parrot
14:46 PacoLinux joined #parrot
14:50 Coke yesterday.
14:50 purl hmmm... yesterday is seemingly so far away
14:50 Coke purl is also the sixth beatle.
14:50 purl okay, Coke.
14:51 Coke er, wednesday, now that I think about it.
14:52 Coke whoops. neither c nor I did the website.
14:52 Coke will do that shortly.
14:57 orbiscerbus hi, all!
14:57 orbiscerbus I compiled the new parrot, but with problems..
14:58 orbiscerbus The problem was undefined PARROT_REVISION var...
15:01 cognominal_ i had the sma problem
15:01 cognominal_ ...same problem using svk
15:03 orbiscerbus well, I set #define PARROT_REVISION 0 in parrot-0.5.3/include/parrot/config.h and it compiled nicely, and passed the test suite..
15:07 Coke website updated.
15:08 Coke orbiscerbus, cognominal_: can you coordinate on getting a ticket open for that issue?
15:09 orbiscerbus no problem.
15:09 Coke danke.
15:09 orbiscerbus where do I open the ticketc?
15:09 orbiscerbus tickets..
15:09 purl tickets are pretty high.
15:10 Coke forget tickets
15:10 purl Coke: I forgot tickets
15:11 orbiscerbus I forgot them, too :)
15:11 Coke http://svn.perl.org/parrot/​trunk/docs/submissions.pod
15:12 bgeron joined #parrot
15:12 spinclad well, tickets are pretty high these days, or open them in RT.
15:12 spinclad tickets
15:12 spinclad tickets?
15:12 purl tickets are pretty high these days, or open them in RT.
15:13 spinclad purl, forget tickets
15:13 purl spinclad: I forgot tickets
15:14 orbiscerbus I never submitted bug report before...
15:14 spinclad welcome aboard
15:15 orbiscerbus also, I don't know what the correct value for PARROT_REVISION should be.
15:15 orbiscerbus so, I can't submit a patch.
15:17 spinclad hm. for the release it should be 0.5.3 i believe; for HEAD i seem to recall it being removed lately...  will check
15:17 particle the REVISION is 0, unless it's an svn co, then it's whatever svn reports
15:18 particle the VERSION for the release is 0.5.3
15:18 particle and now, from svn, 0.5.3-devel
15:18 spinclad i sit corrected; particle++
15:26 orbiscerbus I browsed through the source code a little. config.h file is generated with Parrot::Configure::Compiler.pm which gets revision data from Parrot::Revision::_get_revision() function. Somewhere in between that process revision number got lost, config.h is not generated correctly, and parrot won't compile...
15:28 particle ok, we can work on that. lemme do a quick search
15:29 particle damn, building c99 gives me a segfault
15:32 particle orbiscerbus: since you have a non-svn co, could you do a test for me?
15:33 orbiscerbus I will try, particle..
15:33 particle do you know perl 5?
15:33 purl (perl 6)--
15:33 orbiscerbus what do I have to do?
15:33 orbiscerbus yes.
15:33 * particle throws purl out the window. snarky bot!
15:33 particle ok, in lib/Parrot/Revision.pm, _get_revision sub
15:34 particle make sure everywhere there's a return statement, that $revision is 0 if undefined
15:34 particle that is, that _get_revision never returns undef
15:34 particle that will likely fix the problem
15:35 particle there should be tests in t/... i'll look for em
15:35 orbiscerbus ok
15:35 particle t\configure\017-revision_from_cache.t
15:35 particle t\configure\018-revision_to_cache.t
15:35 particle t\postconfigure\03-revision.t
15:35 particle t\postconfigure\04-revision.t
15:35 particle t\steps\auto_revision-01.t
15:35 particle t\tools\revision\01-get_revision_numbers.t
15:35 particle t\tools\revision\02-print_src.t
15:41 Limbic_Region joined #parrot
15:41 orbiscerbus After I changed the _get_revision() to return 0, it compiled without errors. I'm running make test now...
15:47 particle \o/
15:48 orbiscerbus make test done
15:57 jhorwitz_ joined #parrot
16:06 rdice joined #parrot
16:10 jhorwitz opbots, names
16:13 * particle looks for AndyAway
16:13 silug joined #parrot
16:14 c_otto About how many patches should I have contributed before I ask for svn commit privileges?
16:14 c_otto (or loc, if that's a better measure)
16:14 particle c_otto: enough to have given us a history to judge
16:15 particle it took me ~6mo, maybe more. but that was long ago
16:15 Andy Yes, particle?
16:15 purl The most abundant particle in the universe is the moron. or spin 1/2, charge 2/3
16:15 particle andy: i'm having a segfault problem and imcc.l is involved
16:16 Andy a new segfault?
16:16 particle yes
16:16 Andy hmm
16:16 Andy tell me more.
16:16 particle i just submitted a ticket for it. valp is invalid
16:16 davidfetter joined #parrot
16:17 particle <emit,INITIAL>".macro" { return read_macro(valp, interp, yyscanner); }
16:17 Andy hmmm, lemme look
16:17 Andy how do you get the segfault
16:17 particle cd languages/c99 && make
16:17 particle you may not get it, but i do with msvc
16:18 particle the actual command is ..\..\parrot.exe -o src\CPP_PGE2AST.pbc --output-pbc src\CPP_PGE2AST.pir
16:18 Andy what if you go back to 25802?
16:18 particle heh, then languages/c99 didn't exist
16:18 particle i can fake that, though.
16:18 Andy ok
16:18 particle gimme a few to set up an env
16:20 Andy oh look I got a segfault
16:20 particle yay!
16:21 Andy why do you say valp is invalid
16:21 Andy it fails on 25802 too
16:21 Andy oh wait
16:22 particle my debugger tells me so. valp points somewhere valid, but everything inside is goofy
16:24 Andy in DUP_AND_RET or DUP_AND_RET_FREE
16:25 * Coke wonders if this has anything to do with macro cleanup. =-)
16:25 particle this is why i contacted andy. chromatic may be involved, too.
16:25 Andy Coke: It must, but I can't fathom why.
16:26 Coke freaky.
16:26 Coke is occasionally glad he's not a c programmer.
16:26 Coke ar.
16:26 Coke ARRRRRRRRRRRRRRRRRRRRRRR
16:26 * particle is always glad coke isn't a real pirate
16:27 Andy the diff we're talking about is
16:27 Andy -          return (token);                    \
16:27 Andy +          return token;                      \
16:27 Andy where is it segfaultng?
16:27 Coke what could token be where that would matter?
16:27 particle http://rt.perl.org/rt3/Tic​ket/Display.html?id=51104
16:27 Andy Coke: dunno
16:28 particle an expanded macro?
16:28 particle doubtful.
16:30 particle when i try to trace up to yyparse in the bt, the debugger sends me somewhere wrong (a line with a comment)
16:30 Coke ... was the generated file regenerated?
16:31 Andy i'm digging.
16:31 Andy Isure hope it's not differences between my bison and chromatic's.
16:31 particle it's possible it's missing the proper #line directive
16:32 particle i've seen wrong line happen before due to missing or invalid #line
16:32 particle but, yeah, could be bison
16:35 Coke rdice++
16:35 rdice eh?
16:35 Coke your emails on the funding thread which allow me to remain, thankfully, silent. =-)
16:36 Andy my segfault has stopped segfaulting.
16:36 rdice I'm home sick today.
16:36 rdice I make myself even sicker by participating in funding threads.
16:36 Coke hopefully nothing too serious. :|
16:36 Andy particle: Can you do a "make distclean" in the root and start over?
16:36 rdice No, I'm coping.
16:37 particle andy: with svn head?
16:37 Andy yeah
16:38 particle c:\usr\local\parrot\bug>nmake distclean && svn up && perl Configure.pl && nmake && pushd languages\c99 && nmake # this may take a while
16:38 particle bbi15m
16:38 Andy ok
16:38 barney joined #parrot
16:38 Andy oh wait, the segfault is here, it's just that the make doesn't trigger it.
16:39 Andy but hey, a remake won't hurt
16:40 Andy see updated ticket
16:40 Andy http://rt.perl.org/rt3/Tic​ket/Display.html?id=51104
16:42 Andy ok, your problem is most likely NOT related to macro fixes.
16:46 Andy Aha, it's something I broke, and it's a good thing.
16:46 svnbotl r25990 | petergibbs++ | trunk:
16:46 svnbotl : In thawfinish, delay setting constant flag on lib_name until Parrot_load_lib
16:46 svnbotl : has been called, to avoid corruption in iglobals during GC.
16:46 svnbotl : This fixes segfault attempting to run tcl tests on some platforms.
16:46 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25990
16:47 Andy If someone is available to help me hash through this, ping me.
16:47 parrot-poke joined #parrot
16:50 particle back
16:50 Andy see my updated notes.
16:50 DarkWolf84 joined #parrot
16:51 particle i didn't get a segfault this time
16:51 Andy I think that's luck.
16:51 Andy run that one line directly.
16:52 Andy ..\..\parrot.exe -o src\CPP_PGE2AST.pbc --output-pbc src\CPP_PGE2AST.pir
16:52 Andy brb
16:52 particle yep, segfaulted this time
16:56 Coke petergibbs++
16:58 Andy well, I'm throwin' in an assertion
17:03 pipping joined #parrot
17:04 pipping hi, just wanted to point out that it says '15 January 2008' at http://www.parrotcode.org/n​ews/2008/Parrot-0.5.3.html, which is when 0.5.2 was released (copy&paste error)
17:04 pipping that's all, thanks for listening :D
17:04 pipping bye
17:04 Coke curses.
17:05 Coke fixed.
17:10 sjansen joined #parrot
17:10 Andy Nobody likes cursing.
17:10 Andy Who's coming to Chicago?
17:11 Coke for yapc::na? me.
17:12 Andy I added an assertion, Particle, but it doesn't fix your problem.
17:12 Andy just should assert rather than segfault. :)
17:17 Tene I'm probably coming.
17:17 svnbotl r25991 | petdance++ | trunk:
17:17 svnbotl : Add an assertion in scan_string, and hoist some redundant code.  Also consting here and there.
17:17 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25991
17:19 particle well, my compiler doesn't do squat with asserts, so it segfaults still
17:21 Andy it's a PARROT_ASSERT, so it should.
17:21 Andy anyway, if you look at the code, you can see the bummer.
17:21 particle yep
17:29 jhorwitz_ joined #parrot
17:32 Theory joined #parrot
17:39 Psyche^ joined #parrot
17:42 Coke I thought the problem was not that valp was null, but that valp->s was screwed up: the assert isn't going to help with that, issit?
17:44 particle valp isn't null. it points to a valid location in memory. the contents of that location are uninitialized.
17:44 particle valp should be nulled out at that point, i suppose, or it's contents should be reinitialized
17:45 Coke Right. So the assert is useless.
17:45 Coke (in your case.)
17:46 Coke (it's not your compiler that's ignoring the assert; it's that the assert isn't tripped.)
17:48 slightlyoff joined #parrot
17:52 Coke is valp a STRING?
17:52 Coke (perhaps a COW bug?)
17:53 Coke oh, I see andy added a comment to the ticket.
17:53 Andy multiple
17:54 Andy How do you mean "valp->s is screwed up"?
17:54 Andy That it poitns to a bad location?  Or that the location is bad?
17:54 Andy What I've changed recently is setting valp->s to NULL after freeing.
17:55 spinclad well, either way the location is bad
17:55 particle yes, everything under valp is null
17:55 particle valp itself isn't null
17:56 Coke even if valp->s is null, you're ... ah, you're asserting VALP->s, so that should work, yes.
17:56 Coke I heavily amend my previous statement. =-)
18:03 Andy oK, so for whatever reason, valp isn't getting populated.
18:03 Andy and most importantly, NOT MY FAULT. :-)
18:03 Coke there's so many other things to blame on you, though: we're good.
18:04 Andy Jews & Arabs?  My fault.
18:04 particle wait. didn't schwern have something to do with that?
18:05 Andy No, he's German, they have their own Jewish issues.
18:09 Andy I wish I understood the flex/bison stuff.
18:11 Coke Once you do, there's a ton of tickets to take. =-)
18:11 Coke You might be able to take the "eliminate global variables in imcc" one; seems like more of a C issue than a flex issue
18:12 pmichaud (generic pong)
18:12 Coke pmichaud: have you arrived?
18:13 pmichaud I have indeed.  Got here this morning (approx 10 hrs ago), and caught up on sleep.
18:13 pmichaud thanks for taking care of the website updates
18:13 particle yay! now you can go looking for bad food. bet you can't find any!
18:14 pmichaud already ate.  :-)  no, no bad food yet.
18:16 * pmichaud trembles at the thought of reading the funding thread.
18:17 Coke It's stopped on a pretty sane note.
18:18 pmichaud I had already decided not to comment, other than to perhaps ask that it not be crossposted to every perl6-* list
18:22 pmichaud hey, here's an oddity on capn
18:23 pmichaud er, cpan
18:23 purl i think er, cpan is now 1.3 GB
18:23 pmichaud 1)  go to search.cpan.org
18:23 pmichaud 2)  enter "Parrot"
18:23 pmichaud 3) follow the link to "Parrot"
18:23 pmichaud (i.e., the first link in the search results)
18:27 particle hrmm, 'tis odd indeed.
18:27 particle i don't know how to fix that.
18:32 pmichaud me neither.
18:33 particle ask andreas or brian_d_foy?
18:39 Coke what's the oddity: the really old module?
18:40 Coke hurm. I was getting "parrot 0.0.5" earlier today doing that. I thought we removed the "for Dummies" link since it's a TM violation.
18:40 pmichaud the oddity:  the documentation for Parrot is the PCT "getting started" document that kjs developed
18:42 Coke I would definitely ask somebody to fix that. :|
18:46 Coke Where is that normally pulled from. README?
18:46 particle probably. check an older distro
18:46 svnbotl r25992 | coke++ | trunk:
18:46 svnbotl : remove potential TM violation (and match file name)
18:46 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25992
18:46 Coke Don't see how.
18:46 Coke when you're on latest, it's different.
18:47 Coke I bet that 'gettingstarted' is preferred over whatever we used to happen to get caught.
18:48 Coke I'm going to go ahead and remove versions 0.4.10 and 0.4.17 from CPAN.
18:49 particle i'll remove 0.4.8,14,16
19:05 cotto Would someone mind looking over my PMC code and giving me pointers about how to improve it?
19:07 Coke nopaste?
19:07 purl hmmm... nopaste is at http://nopaste.snit.ch/ (ask TonyC for new channels) or http://rafb.net/paste or http://paste.husk.org/ or http://poundperl.pastebin.com/ or http://paste.scsys.co.uk/ or don't bother me while I'm eating
19:08 Coke simon?
19:08 purl simon is in the mystic west or the anti-XML or the defintion of perl  or not okra or scary or Men are from Macs, women are from VMS or an SMTP server or an alpha geek or at simon-cozens.org or http://use.perl.org/comment​s.pl?sid=7297&amp;cid=11131 or A Real Unix Guru(tm) or at http://lumacode.com/simon/ (the game)
19:09 particle cotto: nopaste it, we'll look
19:10 nopaste "c_otto" at 131.107.0.104 pasted "PHPArray PMC code" (1920 lines) at http://nopaste.snit.ch/12383
19:11 cotto Thanks.
19:14 Coke cotto: is a PHPArray basically a hash?
19:14 particle pmichaud: did you see my work on yaml_dumper.pir?
19:14 cotto yes, except it can also use integer keys
19:15 Coke and [1] stores something in a different slot than ["1"] ?
19:15 cotto and a couple other constraints that PHP's arrays have
19:15 cotto yup
19:16 particle parrot's Hash PMC allows both str and int keys
19:16 PerlJam cotto: you're working on plumhead?
19:16 cotto yes
19:16 PerlJam cool
19:16 particle get_string_keyed_int, get_string_keyed_str
19:16 Coke particle: yes, but 1 and "1" end up in the same bucket.
19:16 Coke no?
19:16 purl no is is <reply>
19:16 particle not sure.
19:17 cotto That depends on the string's hash.
19:17 nopaste "coke" at 72.228.52.192 pasted "hash example" (10 lines) at http://nopaste.snit.ch/12384
19:19 Coke Sounds like for this, you can't expect those to be identical.
19:20 Coke actually, based on my reading of http://www.php.net/manual/e​n/language.types.array.php, it looks like you should expect that to work.
19:20 particle cotto: i should have mentioned earlier that hll contributors usually don't need to wait as long for a commit bit as core contributors
19:20 cotto that's good to know
19:21 stephen left #parrot
19:21 cotto Coke, you're right.  That should work.
19:22 Coke if that works, I wonder if you need to create an entire new pmc at all.
19:22 Coke which is avoiding the review you were asking for, I know. =-)
19:22 pmichaud I've always treated digit string keys and numeric keys equivalently in PHP
19:22 pmichaud there is a difference that crops up in the ordering that results, however.
19:23 cotto I'll need to make some changes, but I don't think they'll be too big.
19:23 cotto Thanks for catching that.
19:23 cotto Coke++
19:26 PerlJam pm: how's eu?
19:34 peeps[work] joined #parrot
19:51 pmichaud PerlJam: eu is pretty nice, thus far
19:51 pmichaud it was fun to walk around some of the streets and look at the various shops
19:52 jmcadams joined #parrot
19:53 jmcadams Conrad Schneiker or Joshua Gatcomb hang out here?
19:53 Coke LR?
19:53 purl somebody said LR was long-range, so that makes up for it :)
19:53 Coke limbic region?
19:54 svnbotl r25993 | petdance++ | trunk:
19:54 svnbotl : Heavy consting of hash-key building.  This is the first true consting I have been able to do lately.
19:54 svnbotl r25994 | pmichaud++ | trunk:
19:54 svnbotl : [rakudo]:
19:54 svnbotl : * Get postcircumfix < > to use <quote_expression> instead of <anglewords>
19:54 svnbotl :   (especially since quote_expression already knows how to handle
19:54 svnbotl :   splits and other coding issues.
19:54 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25994
19:54 jmcadams limbic = Joshua?
19:54 * jmcadams was wondering because i was following the Perl6 rss and saw a big email flurry about funding work on parrot
19:54 particle yes, limbic_region is joshua gatcomb
19:55 jmcadams thx
19:55 particle jmcadams: the yapcchicago website is somewhat sparse on date details
19:55 particle has it not been finalized yet?
19:55 jmcadams 16th-18th
19:55 jmcadams June
19:55 jmcadams should be right at the top
19:55 jmcadams lemme check
19:56 particle ah, it's there
19:56 rdice_ joined #parrot
19:56 jmcadams i'll add it in a few more places around the page
19:56 jmcadams so that it's easier to notice
19:56 particle thanks, that'll help
19:56 pmichaud it did take me a long time to see it when I was looking a few days ago
19:57 jmcadams ah, on the topic of parrot funding... I was going to point out http://www.thepoint.com
19:57 AndyLunch hey this looks bad
19:57 jmcadams seemed like a decent way of having many people raise $$ for a specific purpose
19:57 AndyLunch <emit,INITIAL>{DOT}{LETTER}{LETTERDIGIT} can run off the end of the function.
19:57 pmichaud there's also fundable.com
19:57 pmichaud (or perhaps it's .org)
19:57 jmcadams .com did the trick
19:58 jmcadams cool, just wanted to make sure that they knew about some place like that
19:58 Oniichan joined #parrot
19:58 jmcadams seems like a decent idea
19:58 pmichaud I've used fundable to get browsercam subscriptions.  It seems to work pretty well.
19:59 jmcadams fundable does a 7% deduction it seems
19:59 pmichaud yes, it is a little pricey
19:59 jmcadams thepoint claims to only take 3.5%
19:59 jmcadams for cc processing
19:59 pmichaud (they waive the deduction for browsercam subscriptions, though)
20:00 jmcadams other than that, they seem compariable
20:00 pmichaud I'll keep thepoint in mind
20:00 * jmcadams is going to have to figure out what a browsercam is
20:00 pmichaud browsercam.com -- it lets you see how web pages appear in many different browsers at once
20:00 jmcadams neat
20:00 pmichaud normally it's $60/month, but they also offer a $500/yr option with 20 separate login accounts
20:01 pmichaud thus, if you can get 20 people to contribute $25 apiece, it ends up costing $25/yr instead of $60/mo
20:01 pmichaud I've used it for quite a bit of web design projects -- it's great to be able to test a layout and see exactly how it will look in many different browsers
20:02 jmcadams yeah, seems like a good tool
20:02 jmcadams thx
20:02 Andy Who is hep to any of this parser stuff?
20:02 pmichaud andy: parser... as in imcc?
20:02 particle andy: not i. how can that run off the func name?
20:03 Andy go look at imcc.l
20:03 Andy line 522
20:03 Andy see how the func can run off the end without a return
20:04 particle you mean no 'return' after mem_sys_free ?
20:04 Andy lemme run it thru splint.
20:04 Andy yes
20:05 particle ok, i see. thought you meant something else. that's no good.
20:05 Andy indeed it is not.
20:10 spinclad TM?
20:10 purl i guess TM is TextMate (http://macromates.com/) or TicketMaster
20:11 spinclad TM violation?
20:11 pmichaud TM == trade mark
20:11 pmichaud "XYZ for dummies"
20:11 spinclad gotcha, thx
20:12 spinclad purl, TM is also trademark
20:12 purl okay, spinclad.
20:15 cotto TM?
20:15 purl TM is TextMate (http://macromates.com/) or TicketMaster or trademark
20:15 cotto Is purl documented anywhere?
20:15 Coke google for infobot.
20:16 pmichaud purl, purl?
20:16 purl i am going alone. or almost an anagram of Donaudampfschiffahrtskapitaensmuetzenkordel or a perv or an auto-triage bot or a she or so stupid or a smartass or the sixth beatle.
20:16 pmichaud purl is also http://www.infobot.org
20:16 purl okay, pmichaud.
20:19 pmichaud Coke:  should I remove the old versions of parrot cpan that I have?  How far back are we removing?
20:19 particle everything before 0.5 for sure
20:20 Coke pmichaud: you esp. be careful (don't remove the most recent. =-)
20:20 Coke what particle said is fine.
20:20 pmichaud so, it looks like I can delete the 0.4.9 and 0.4.15 releases... do I delete the .meta and .readme files also?
20:22 particle yep
20:22 Coke Danke.
20:23 pmichaud done.
20:27 x joined #parrot
20:30 pmichaud afk for a bit
20:36 cotto Is there a quick way to make an Integer PMC from an INTVAL?
20:37 cotto (other than pmc_new)
20:45 Coke in C or PIR?
20:45 Coke (guessing C)
20:47 Coke i think you need pmc_new and then to set its invtval.
20:49 Andy TM is also transcendental meditation
20:49 purl okay, Andy.
20:49 cotto that's what I though
20:49 cotto I meant C
20:53 svnbotl r25995 | paultcochrane++ | trunk:
20:53 svnbotl : [pmc] VTABLE_() function lines need to be all on one line for icc to be able
20:53 svnbotl : to build them...
20:53 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25995
20:58 Coke that icc requirement is like, crazy, yo.
20:58 particle some compilers suck differently than others
21:38 jmcadams particle: http://conferences.mongueurs.net/yn2008/
21:38 jmcadams i think this is beter
21:38 Theory joined #parrot
21:38 jmcadams i have the dates on the front page in three places
21:39 PerlJam jmcadams: when is yapc again?  ;)
21:39 jmcadams :)
21:56 pmichaud I think it's in July.  :-)
21:56 pmichaud jmcadams: *much* better, thanks
21:56 * pmichaud considers submitting a talk for YAPC::NA.  :-)
21:57 jmcadams pmichaud: considers?
21:57 jmcadams you should :)
21:57 jmcadams you've got tons of parrot stuff going on
21:57 * Coke wonders when the submission deadline is.
21:58 * Coke doesn't really want to give a talk, honestly.
21:58 Coke would be happy to assist someone, however.
21:58 pmichaud oh, I'm definitely planning a talk for YAPC::NA
21:58 jmcadams Coke: March 15th
21:58 pmichaud actually, probably multiple talks
21:58 Coke jmcadams: perhaps I'll have a change of heart by then. ^_^
21:58 Coke you're not running it again, are you?
21:59 jmcadams in a round-about way
21:59 Coke ... not that I didn't have a great time last time. =-)
21:59 jmcadams Chicago.pm got the bid
21:59 jmcadams there are a few people on the organizing committee
21:59 jmcadams but a lot of the stuff fell back to me for one reason or another... some of it just being that I already had been-there-done-that
22:04 Coke poor sod.
22:05 zev_ could I get a commit bit to parrot?
22:12 Coke zev?
22:12 purl i guess zev is Zero Economic Value or a term used by rescuers to distinguish refugees in NOLA... conversely "taxpayer"
22:12 Coke I can't map your id to an email address; Typically commit bits are handed out after several patches are submitted and approved.
22:14 zev_ I have submitted several patches.  the email address used has usually been either zev@strangersgate.com or zev-perl6@strangersgate.com
22:14 pmichaud zev:  have you sent in a cla yet?
22:15 zev_ cla?
22:15 purl cla is, like, Contributor License Agreement or http://www.perlfoundation.org/​contributor_license_agreement
22:16 zev_ appearently not.  I can can do so...
22:17 pmichaud are there some outstanding patches yet to be applied?
22:17 pmichaud I've been a bit behind on applying rakudo patches
22:17 Coke it's a necessary but not sufficient condition to the actual commit bit.
22:17 pmichaud correct.
22:17 purl no, it's not!
22:18 zev_ yes: #50558 and #50550
22:19 zev_ I also have a patch ready (but not yet submitted) for the glue to let you write the rakudo runtime in perl6.  I would like someone to look that over, however
22:19 * pmichaud checks
22:20 pmichaud is the 'run' builtin documented somewhere?
22:20 Coke can always submit patches with a "comments requested" note.
22:20 zev_ pmichaud: yes.  one sec
22:20 pmichaud and coke is correct, feel free to submit the patch as a "request for comments" patch
22:21 zev_ SubjectStatusQueueOwnerPriority
22:21 zev_ ack.  sorry
22:21 zev_ http://feather.perl6.nl/syn/S29.html
22:23 pmichaud re: #50550 (patch for exponential numbers) --- STD.pm doesn't allow exponential notation on integers, which is why rakudo doesn't have it
22:24 pmichaud we might need to ask on perl6-compiler
22:25 zev_ yes, I realize that STD.pm doesn't allow it (hence my comment).  However the spec indicates that it should be possible.  I can send mail
22:25 nopaste "cotto" at 131.107.0.104 pasted "updated phparray pmc code" (1976 lines) at http://nopaste.snit.ch/12385
22:25 TimToady hmm, that's just an oversight
22:25 zev_ ... or not
22:26 pmichaud TimToady: any particular points about Perl 6 that you want me to hit (or avoid) in my FOSDEM presentation tomorrow?
22:29 cotto That fixes the string -> int index issue that Coke mentioned.
22:31 zev_ ok, latest patch is #51112 (http://rt.perl.org/rt3/Tic​ket/Display.html?id=51112).  I'd appreciate if someone would take a look at it at some point
22:31 TimToady I'd suggest not talking about fundraising.  :)
22:31 cognominal_ pmichaud and our usual guest are welcome to talk in the French Perl Worshop  in Albi Friday  May 30 and Saturday may 31
22:31 pmichaud I definitely wasn't going to mention that.
22:31 pmichaud (fundraising, that is)
22:31 pmichaud mmmm, French Perl Workshop
22:32 cognominal_ we will probably rent large cars to go there from Paris
22:32 cognominal_ Jonathan can tell you we are nice :)
22:32 pmichaud oh, Paris (Texas) isn't far from where I live :-)
22:32 cognominal_ not the same Paris
22:33 pmichaud But Albi is definitely not easily reachable by car.
22:33 cognominal_ and may be someone living in Central Paris like rgs can lodge you a few day before for you to do tourism
22:33 pmichaud well, I don't know that I get to do much tourism-like stuff until this summer
22:34 pmichaud until Paula (my wife) gets done with her chemo treatments, my trips will tend to be get-there-do-business-return-home
22:34 Andy so, relaly, who can I bug about imcc.l?
22:35 cognominal_ ho, I hope here disease is not serious or easily curable
22:35 cognominal_ s/here/her/
22:35 pmichaud oh, it's serious enough, unfortunately (ovarian cancer)
22:35 zev_ also, if a CLA is required for commit access, perhaps that requirement should be added to docs/submissions.pod
22:36 pmichaud zev:  correct
22:36 TimToady pmichaud: okay, I think I've fixed dec_number
22:36 cognominal_ family comes before anything
22:36 pmichaud cognominal_: agreed.  The main reason I'm doing FOSDEM (besides having committed to it) is that Paula insisted
22:37 pmichaud and, as it turns out, I'll be back in time for the rough phase of treatment
22:37 cognominal_ I wish both of you luck
22:37 pmichaud thanks
22:37 cognominal_ doyou have any kids or intend to do?
22:38 cognominal_ or does ovarian cancer will comprimise that?
22:38 pmichaud we have two kids from Paula's first marriage.  We don't have any of our own (but weren't planning any either)
22:38 cognominal_ good
22:38 pmichaud and yes, ovarian cancer normally compromises that.
22:39 pmichaud unless the patient is quite young and the cancer is not too extensive
22:39 pmichaud TimToady: dec_number looks good to me
22:39 TimToady was tempting to make both parts ? but that wouldn't work
22:40 TimToady would parse integers as dec_number for one
22:40 pmichaud yes, I was wondering how you would resolve that.  :-)
22:40 TimToady and mess up the 1.method interpretation on longest-token
22:40 TimToady if we used a <?before>
22:40 TimToady so it looked like the easiest way was the way I did it
22:40 TimToady with a small duplication of code
22:41 cognominal_ I wish the other talk interested me but only your talk interested me in FOSDEM so I decide not to go.
22:41 pmichaud I was thinking something like   [ '.' \d+[_\d+]* | <?before E> ]
22:41 TimToady yeah, that'd probably work too
22:42 pmichaud it would avoid the duplication :-)
22:42 TimToady but this way gets the whole number into the DFA matcher
22:42 pmichaud true.
22:42 TimToady and some token rules can probably be taught not to reparse as an optimization
22:43 TimToady though I haven't worried about that much yet
22:43 cognominal_ a previous girl friend had a overian kyst the size of a first and the first gynecologist tell she was pregnant despite us having protected sex
22:44 cognominal_ so I hope you wife is in more competent hands
22:44 cognominal_ s/first/fist/
22:44 pmichaud thus far we're very happy with the g-oncologist that we have
22:44 cognominal_ good
22:45 pmichaud he did a fantastic job with the surgery, and he seems to know exactly how to manage the whole process
22:46 pmichaud and whenever we've asked about other treatment possibilities, he's been very aware of them and been able to explain the pros and cons and when they might be appropriate
22:46 allison joined #parrot
22:47 jonathan cognominal_: French Perl workshop is in May?! That's only just over six months since the last FPW!
22:47 pmichaud allison (and anyone else on channel) -- anyone have any specific topic they think I should/should not cover in my Perl 6 talk tomorrow at FOSDEM?
22:47 cognominal_ I wish some day PGE or its progeneture wil help to sequence dna and help to save life
22:47 cognominal_ jonathan, we are so impatient to have you again :)
22:48 pmichaud we've already decided _not_ to cover fundraising possibilities
22:48 pmichaud :-)
22:48 pmichaud although if someone offers a large contribution tomorrow, I might change that position :-)
22:48 jonathan cognominal_: May might not be so good for me. :-(
22:49 ruoso joined #parrot
22:49 cognominal_ and Albi being the south west of France, the fodd should be good
22:49 jonathan Though I'll certainly say, I enjoy attending FPW.
22:49 allison pmichaud: lately I've changed my "easy to use" message to "powerful tools". Ultimately it's the same thing (powerful tools make writing compilers easy), but seems to be more exciting to the audience
22:49 pmichaud allison: good point
22:49 purl Thanks, pmichaud, I'm glad you agree with me.
22:50 cognominal_ we had nicholas two in one previous occasion. Nice chap too
22:50 jonathan pmichaud: Similar to what Allison said, one thing that occurred to me is to mention that people are using NQP and Perl 6 rules - currently implemented parts of Perl 6 - Ilready.
22:50 allison pmichaud: and while I wouldn't ask for funding, it's entirely appropriate to thank NLNet and Mozilla who have made the work go so much faster
22:50 cognominal_ s/two/too/
22:50 jonathan s/Ilready/already/
22:51 pmichaud allison: also good point about nlnet/mozilla
22:51 jonathan Similar in the sense of, it relates to the compiler tools. :-)
22:51 pmichaud jonathan: good idea, thanks.
22:51 jonathan But mentioned that we already have bits of Perl 6 implemented and people are doing useful stuff with them...
22:51 pmichaud much of my talk is about Perl 6 the language as opposed to Rakudo the implementation
22:51 pmichaud although about 1/3 to 1/2 of the talk will undoubtedly include Rakudo, Parrot, and the compiler tools
22:52 jonathan The stuff I'm presenting here in Kiev tomorrow is about the language, but much of the stuff I'm showing actually runs on Rakudo already. :-)
22:52 jonathan Of course, I kinda cheated by knowing what I was going to be talking about ahead of time and then implementing those bits.
22:52 slightlyoff joined #parrot
22:52 Limbic_Region joined #parrot
22:52 pmichaud that's not cheating, it's progress.
22:53 jonathan Only if my implementation is good. ;-)
22:53 jonathan I'm sure what I've done will need lots of tweaks.
22:53 jonathan But, it's a start.
22:53 pmichaud I do have some changes to make to class/role/etc., but those will have to wait a .... right
22:54 pmichaud I'll probably revisit them first week of march
22:54 pmichaud there's a lot of refactoring that needs to take place in order for .HLL to work
22:54 cognominal_ I have seen you have roles right into pmcs
22:54 jonathan I've got at least one hack in the codebase already that I think is related to not having .HLL in place.
22:55 jonathan I put a comment in saying so, so it's grep-able.
22:55 cognominal_ ack-able!
22:55 pmichaud right.  One of the challenges is that some of the tools (like HLLCompiler and Protoobject) arent .HLL-aware yet
22:55 Andy cog++
22:56 jonathan Sure; there's plenty of work to do there.
22:58 jonathan Anyway, I'm tired and have talks tomorrow, so time for sleep.
22:58 cognominal_ good luck
22:58 pmichaud good luck on your talk tomorrow -- wish I was there to see it!  :-)
22:58 jonathan Same with yours!
23:20 cotto CLA?
23:20 purl somebody said CLA was Contributor License Agreement or http://www.perlfoundation.org/​contributor_license_agreement
23:28 kid51 joined #parrot
23:30 svnbotl r25996 | petdance++ | trunk:
23:30 svnbotl : good to splint the imcc stuff, too
23:30 svnbotl r25997 | petdance++ | trunk:
23:30 svnbotl : use appropriate macros in the generated code.  Const, too
23:30 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25997
23:49 mncharity joined #parrot
23:52 rdice joined #parrot

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

Parrot | source cross referenced