Camelia, the Perl 6 bug

IRC log for #parrot, 2008-04-14

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:11 pmichaud jonathan: iwbni we could just get at the lexpads using a normal hash-like interface, and then iterate over them using a standard iterator.
00:21 dalek joined #parrot
00:23 dalek joined #parrot
00:33 dalek joined #parrot
00:38 kid51 joined #parrot
01:01 dalek r26964 | chromatic++ | trunk:
01:01 dalek : [t] Untodoed an assignment inheritance test, after Jonathan fixed assign.
01:01 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=26964
01:03 dalek r26965 | chromatic++ | trunk:
01:03 dalek : [src] Improved C++ compiler compatibility (Senaka Fernando, RT #52854).
01:03 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=26965
01:53 dalek joined #parrot
01:55 guru left #parrot
01:57 pmichaud joined #parrot
01:58 davidfetter joined #parrot
01:58 dalek Krishna Sethuraman | Parrot Development on Windows:
01:58 dalek link: http://www.perlfoundation.org/parrot/i​ndex.cgi?parrot_development_on_windows
01:58 dalek bernhard.schmalhofer@gmx.de | Plumhead:
01:58 shorten dalek's url is at http://xrl.us/bjcdh
01:58 dalek link: http://www.perlfoundation.or​g/parrot/index.cgi?plumhead
01:58 dalek bernhard.schmalhofer@gmx.de | release_schedule_2008:
01:58 dalek link: http://www.perlfoundation.org/parr​ot/index.cgi?release_schedule_2008
01:58 dalek will@coleda.com | How to convert PIR to C:
01:58 dalek link: http://www.perlfoundation.org/parro​t/index.cgi?how_to_convert_pir_to_c
01:58 dalek allison@perl.org | Release Planning:
01:58 dalek link: http://www.perlfoundation.org/pa​rrot/index.cgi?release_planning
01:58 shorten dalek's url is at http://xrl.us/bjcdj
01:58 shorten dalek's url is at http://xrl.us/bjcdm
01:58 dalek allison@perl.org | milestones:
01:58 dalek link: http://www.perlfoundation.org​/parrot/index.cgi?milestones
01:58 dalek Jim Keenan | YAPC::NA::2008:
01:58 shorten dalek's url is at http://xrl.us/7nmt
01:58 diakopter ooks
01:58 diakopter oops
01:59 diakopter well, it sorta worked
02:00 kid51 What were you attempting?
02:03 particle joined #parrot
02:37 diakopter kid51: polled updates from the wiki rss list
02:37 peepsalot joined #parrot
02:37 dalek joined #parrot
02:38 diakopter http://www.perlfoundation.org/feed/work​space/parrot?category=Recent%20Changes
02:39 shorten diakopter's url is at http://xrl.us/bjcei
02:40 contingencyplan joined #parrot
02:42 kid51 Ah, I see.
03:22 dalek r26966 | jkeenan++ | trunk:
03:22 dalek : Per discussion with particle re RT 52706, we're reverting to previous version
03:22 dalek : so that we can get through this week's release without having
03:22 dalek : t/dynpmc/gdbmhash.t hang on Darwin and prevent 'make test' from reaching a
03:22 dalek : conclusion.
03:22 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=26966
03:24 sjansen joined #parrot
03:31 AndyA joined #parrot
04:08 integral joined #parrot
04:38 iblechbot joined #parrot
04:56 Psyche^ joined #parrot
05:37 GeJ joined #parrot
06:19 Senaka joined #parrot
06:19 Senaka seen Infinoid
06:19 purl Infinoid was last seen on #parrot 2 days and 16 hours ago, saying: racing sailboats &  [Apr 11 07:21:42 2008]
06:20 Senaka seen coke
06:20 purl coke was last seen on #parrot 15 hours and 22 minutes ago, saying: -> away for a bit
06:20 Senaka hi anyone here?
06:22 uniejo joined #parrot
06:25 Senaka msg chromatic please check line 4203 in imclexer.c
06:25 purl Message for chromatic stored.
06:54 Senaka left #parrot
06:58 Senaka joined #parrot
07:40 iblechbot joined #parrot
07:54 Senaka left #parrot
07:55 IllvilJa joined #parrot
08:14 cosimo joined #parrot
08:31 kj joined #parrot
09:16 kj joined #parrot
09:25 ruoso joined #parrot
09:33 Senaka joined #parrot
09:33 Senaka kj:ping
09:33 Senaka duh
09:34 Senaka kj: ping
09:34 Senaka seen kj
09:34 purl kj was last seen on purl 20 hours and 55 minutes ago, saying: <private message>
09:34 Senaka purl I see
09:34 purl Senaka: excuse me?
09:35 Senaka left #parrot
09:42 Senaka joined #parrot
09:56 Senaka left #parrot
09:57 Senaka joined #parrot
10:10 kj I'm here
10:10 kj Senaka: ping
10:10 Senaka kj: gotacha
10:10 Senaka *gotcha
10:10 Senaka is imclexer.c needed?
10:11 Senaka irrespective of imcc.l?
10:11 kj needed in the repository, you mean?
10:11 kj it's needed to compile IMCC :-)
10:12 kj "flex imcc.l" generates imclexer.c
10:12 Senaka ok
10:12 kj (well, you have to specify the output file name with the "-o" option, but that's besides the point here)
10:12 Senaka it seems that certain sections within the imclexer.c aren't replaced... at all
10:13 kj are you trying to regenerate imclexer.c?
10:13 Senaka or there is no where else that they are defined
10:13 Senaka well no build it I mean
10:13 kj sorry, I don't exactly understand what you're trying to do.
10:13 Senaka line 4203
10:15 Senaka kj: where does sections below line 4203 come from?
10:15 kj it's generated by flex
10:16 kj that's why it's documented so well :-)
10:16 Senaka from what file?
10:16 kj by the flex program
10:16 kj it's template stuff
10:16 kj ehm well it's code generated from the flex rules
10:17 Senaka ok in the switch block line 4283
10:17 Senaka say i want to add one more case
10:17 Senaka where can that be done?
10:17 kj so this is the machinery that implements the lexer
10:18 kj you can't and shouldn't
10:18 kj you should only update the imcc.l file
10:18 Senaka in that case this ain't getting built on C++
10:19 Senaka I need to add the default case
10:19 kj i think there's a flex++ program
10:19 Senaka is it so...
10:19 kj if a system has c++ compiler, it might have flex++ too
10:19 Senaka ok where is the flex program set?
10:19 kj might be called differently..
10:19 kj in the configure stage
10:19 kj i think
10:20 kj gotta go for 20 min or so
10:20 Senaka flex is the lexer is it?
10:20 kj flex is the lexer-generator
10:20 Senaka ok no issues
10:20 kj so it's the program that generates the lexer for imcc
10:20 Senaka ok
10:21 Senaka i'll get it checked
10:21 kj just like PGE generates a parser based on rules you write
10:21 kj bbl
10:21 Senaka ok
10:21 Senaka ttyl
10:23 Senaka kj: I didn't have flex all this time
10:45 Senaka kj: lex/yacc is what I have
10:46 kj Senaka: lex/yacc should be fine, flex and bison are the same I think, but just newly implemented
10:47 Senaka yep
10:47 Senaka flex/bison
10:47 purl flex/bison are close enough to clones...
10:47 Senaka btw,
10:47 Senaka I still couldn't find what writes that portion of code
10:48 kj well somewhere in lex there's code that printf()s this
10:48 kj so you can't find it; it's generated by the lex executable
10:49 Senaka ok... can u tell me the steps to generate this file?
10:50 kj http://www.gnu.org/software/fle​x/manual/html_node/flex_19.html
10:51 kj this is for the c++ compilation; you don't need flex++ I was confused
10:51 kj I confused myself :-)
10:51 kj flex++ is needed to generate a lexer *class*, which isnot what we want
10:51 kj we just need to compile with c++ compiler instead of a c compiler.
10:51 kj as for the steps:
10:51 Senaka kj: yes
10:51 kj you on windows?
10:51 Senaka Linux
10:51 kj ok
10:52 Senaka Ubuntu Gutsy
10:52 purl Ubuntu Gutsy is still on 5.8.8
10:52 kj well, what should happen is that the file is regenerated automatically
10:52 Senaka kj: I don't think so
10:52 kj so after changing the imcc.l or .y file, the imc{lexer,parser}.c files should be regneerated I think
10:52 Senaka kj: no changes in that particular section
10:52 Senaka I see that sections in this file is regenerated
10:52 Senaka but not all
10:53 kj what sections are you referring to?
10:53 Senaka line 4203
10:53 Senaka it should read #line 4204 "compilers/imcc/imclexer.c"
10:54 Senaka below that you'll find some switch/case constructs
10:54 kj yes
10:54 Senaka those seem to come from no where
10:54 kj well, the whole switch/case thing comes from nothing: it's the implementation of the lexer
10:55 kj you won't find it back in the imcc.l file
10:55 Senaka hmm... in that case we need to probably hack that or find a c++ alternate
10:55 kj did you check the link I just pasted?
10:55 kj it should compile fine
10:56 Senaka checking
10:56 kj you can easily check this by writing a sample grammar/lexer, with 1 rule or so, and the generate the lexer , and compile it with g++
10:57 Senaka hmm....
10:58 kj it seems your ticket is about code in src/pmc/bigint.c
10:59 Senaka bigint.c is another issue
10:59 kj can you try to compile imclexer.c with g++?
10:59 kj what does it say?
10:59 Senaka that has something to do with including gmp.h
10:59 Senaka yes that's another
11:00 kj I thought when using a c++ compiler, functions should be declared as extern "C" { ... }
11:01 Senaka yes and no
11:01 kj ok, need to go again for some time...
11:01 Senaka not necessarily other than someone tries to refer it
11:01 kj what happens when you compile imclexer.c with g++?
11:01 Senaka doing it at the moment
11:01 Senaka did a distclean and started make
11:01 kj just g++ -c imclexer.c
11:02 rdice joined #parrot
11:02 kj maybe it's a keyword thing; c++ has many new keywords, sometimes they're used as var.names in parrot source
11:02 kj (which should be fixed)
11:03 kj bbl
11:03 Senaka kj: gimme some time
11:03 Senaka ttyl
11:06 kid51 joined #parrot
11:09 Senaka kj: aha! that's a warning
11:14 Senaka is Buffer a dead concept?
11:19 kj dunno what buffer that is
11:19 kj s/that is/you mean/
11:19 Senaka GC has PMC and Buffer
11:19 Senaka two concepts...
11:20 Senaka but no where else (ex:- Glossary) is this being defined
11:20 kj ooh ok.
11:20 Senaka is this a dead concept?
11:20 kj ok, not even in the gc pdd?
11:20 Senaka it is
11:20 kj i don't know much about gc
11:20 Senaka but nothing like the definition of PMC in the glossary
11:21 Senaka btw
11:21 kj except that it's collecting garbage and not using a ref. counting scheme, and the current impl is slow :-)
11:21 Senaka do u find an HTML version of the glossary
11:21 kj its on the website i think
11:22 kj see resources section
11:23 Senaka yeah
11:25 Senaka kj: that switch/case thing is a warning rather
11:25 Senaka so no issues there
11:25 kj that's good.
11:26 Senaka the issue was with "input not defined" a bug in imcc.l I will send the patch shortly
11:26 Senaka In the meantime, the GMP issue in bigint still holds
11:28 kj sorry can't help you there.
11:28 kj i'm not a core hacker
11:28 Senaka kj: ok perhaps chromatic should have something to say
11:29 Senaka this has something to do beyond the scope of Parrot. A possible confusion of includes
11:29 Senaka either a wrong order, or we included some incompatible headers
11:43 man_in_ltop joined #parrot
11:43 * man_in_ltop waves
11:43 man_in_ltop are there any tutorials on how to combine PGE and PIL/PASM? :)
11:45 Senaka how to send a second patch to an issue?
11:45 kj man_in_ltop: what do you want to do?
11:45 kj Senaka: just reply on it i think?
11:45 man_in_ltop kj, play with defining a language
11:46 kj man_in_ltop: check out www.parrotblog.org
11:51 Senaka kJ: well gave me a new ticket
11:51 kj through RT you can reply on it
11:55 Senaka kj: ok I did send a mail saying that #52874 is related
11:57 Senaka kj: have you got to mail parrotbug only once?
11:57 Senaka per issue i mean.
12:00 tetragon joined #parrot
12:00 uniejo joined #parrot
12:16 Coke Senaka: to attach a patch to a ticket, you have 2 options: 1) use the web interface, 2) follow the instructions in the AUTOREPLY email you go.
12:16 kid51 The initial RT ticket is created thru mail to parrotbug@parrotcode.org.  Subsequent posts can be done thru the GUI interface (toggling CC to perl6-internals), by reply to mail received from the list, by reply to items in the news interface, or by email to parrotbug-followup@parrotcode.org referencing the RT number in the subjectline.
12:17 Coke and we can merge tickets if a new one is inadvertently opened.
12:20 man_in_ltop kj, thanks, looks pretty good
12:23 Senaka Coke: Can we merge them?
12:23 Coke I just said it was possible, yes.
12:24 Senaka 52876, 52874, and 52854
12:24 Senaka :D
12:24 Coke $work now. it's not a rush, anyway.
12:24 Senaka it took time for me to figure out that gmail screwed things up when I replied by adding parrotbug as one to address
12:24 Senaka :)
12:32 rdice joined #parrot
12:35 Senaka #include <gmp.h> has issues on Ubuntu Gutsy
12:35 Senaka I have to add a hack #define _LIBINTL_H to get through
12:35 Senaka this is related to C++ build
12:36 Senaka Coke: ??
12:37 Senaka seen Allison
12:37 purl Allison was last seen on #parrot 5 days and 17 hours ago, saying: those were bad years anyway ;)  [Apr  8 12:57:01 2008]
12:37 Coke Senaka: what?
12:38 Senaka #include <gmp.h> has issues on Ubuntu Gutsy with C++
12:39 Coke if that's a bug report... open a ticket.
12:39 Senaka or else who's paultcochrane?
12:39 Coke You're not going to have good luck pursuing things based on the author information in core stuff.
12:39 Senaka :)
12:39 Coke as it's often stale
12:40 Senaka ok... thanx for the hint
12:40 Coke and if you're looking at svn commits, a lot of people commit formatting cleanups; they're not the ones to know what's going on that section of code.
12:40 Coke your best bet... if no one speaks up on IRC, anyway, is to open a ticket.
12:40 Senaka ok
12:40 Coke then it doesn't get lost, someone not on IRC can see it, and it gets cc'd to the list.
12:41 Coke (I'm not trying to ditch you, just trying to make sure these issues do get resolved.)
12:41 Coke gotta run.
12:41 Coke work beckons.
13:02 gryphon joined #parrot
13:17 Senaka perl #50068 seems to be resolved
13:28 Senaka pmichaud: ping
13:36 paco joined #parrot
13:39 rdice joined #parrot
13:43 skids joined #parrot
13:50 rdice joined #parrot
13:51 Senaka seen pmichaud
13:51 purl pmichaud was last seen on #parrot 13 hours and 40 minutes ago, saying: jonathan: iwbni we could just get at the lexpads using a normal hash-like interface, and then iterate over them using a standard iterator.
14:00 Andy joined #parrot
14:02 Senaka well, there are a good number of commits between March 28th and April 14th... one of these introduced a new C++ build break
14:18 Senaka question regarding tower of hanoi sample
14:18 Senaka shouldn't we start from initial position?
14:18 pmichaud where is it starting now?
14:19 Senaka move one
14:19 Senaka so the initial position is not see
14:19 Senaka *seen
14:20 pmichaud makes sense to me :-)
14:20 Senaka which current? or the one with the starting point?
14:22 Senaka pmichaud: you only need one line to fix this
14:22 avar w 26
14:22 Senaka ## Now solve it.
14:22 Senaka print_towers(towers)
14:22 Senaka move_stack(size, 0, 2, 1)
14:22 pmichaud avar:  bwngo!
14:22 avar MM!
14:22 Senaka ?? go for it?
14:26 jhorwitz joined #parrot
14:26 Senaka pmichaud: ??
14:28 pmichaud Senaka: looks good to me
14:28 Senaka pmichaud: do you refer to the current one?
14:31 pmichaud your change looks good to me
15:06 ambs joined #parrot
15:06 ambs joined #parrot
15:06 ambs seen merlyn
15:06 purl merlyn was last seen on #moose 24 days and 17 hours ago, saying: ... http://methodsandmessages.vox.com/lib​rary/post/the-year-of-smalltalk.html  [Mar 20 15:09:20 2008]
15:08 moritz merlyn was last seen on perlmonks.org 1 day ago ;-)
15:41 sjansen joined #parrot
15:47 markus joined #parrot
15:52 Senaka joined #parrot
15:53 Senaka msg Coke App::SVNBinarySearch needs Version and Module::Build
15:53 purl Message for coke stored.
16:03 markus_ joined #parrot
16:16 dalek r26967 | fperrad++ | trunk:
16:16 dalek : [regex]
16:16 dalek : - update the target test
16:16 dalek : - allows the Makefile regeneration
16:16 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=26967
16:23 cognominal http://rt.perl.org/rt3/Publi​c/Bug/Display.html?id=52858  # I tripped In my patch on should read "in past mode" instead of "in pir mode"
16:23 cognominal sorry
16:26 markus joined #parrot
16:45 dalek r26968 | fperrad++ | trunk:
16:45 dalek : {regex]
16:45 dalek : - oops (duplicate regeneration)
16:45 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=26968
16:45 Senaka left #parrot
16:51 markus_ joined #parrot
17:19 robkinyon joined #parrot
17:19 robkinyon left #parrot
17:22 barney joined #parrot
17:28 Theory joined #parrot
17:49 japhb Looking for Configure advice:  I'm trying to convert my OpenGL bindings proof of concept code to work within the real Parrot build system.  My assumption is that I need to probe for OpenGL libraries late in the Configure steps, around the time that GMP, readline, gdbm, etc. are probed.
17:49 japhb The question is, do I probe for every library and header I want all at once, or do I do them piecemeal?
17:50 japhb I'll need to probe for GL itself, plus GLU, GLUT, and what *variety* of GLUT I have (original GLUT, freeglut, OpenGLUT, etc.)
17:50 Infinoid it makes sense to clump all of that stuff together
17:51 Infinoid not necessarily all in the same line of output, but certainly all in the same section
17:51 japhb The next thing is, if I find some borken thing, like having headers but not matching libraries installed, do I just say "no", or do I emit some sort of warning about that?
17:52 Infinoid maybe have one check for the header and one check for the library... compile an object and then link it, or something
17:52 japhb And how does a user find out how to enable the OpenGL stuff, if the probes fail to find all they need?
17:52 japhb hmmm
17:53 Infinoid its nice to give the user specifics about why something failed
17:53 Infinoid enabling the OpenGL stuff is platform-specific
17:53 japhb nodnod
17:54 Infinoid its sufficient to tell the user what broke
17:54 Infinoid (for instance, we don't go through the trouble of detecting Debian machines in order to tell them what to apt-get for libgmp)
17:55 japhb nodnod
17:55 dalek r26969 | bernhard++ | trunk:
17:55 dalek : Update date in PLATFORMS
17:55 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=26969
18:01 Senaka joined #parrot
18:08 lidden joined #parrot
18:20 jonathan Candlelit fondue + breeze = hard work.
18:21 japhb Real candle, or sterno?
18:21 jonathan Whatever tealights class as. ;-)
18:21 japhb realish.
18:22 japhb That's just crazy.  How many tealights are you using?
18:22 jonathan Three tends to do it.
18:22 * japhb shakes head in wonderment.  Never would have believed that would be enough heat.
18:23 jonathan Well, the first time I did it, I wasn't sure they'd heat it enough. So I touched the bottom of the pot with my hand to see. And spent the next few days with sore, burnt knuckles.
18:23 jonathan So it would appear, there is enough heat. ;-)
18:25 jonathan I probably shouldn't mention this hear, or everyone will conclude I'm too dumb to have a commit bit.
18:25 jonathan s/hear/here/ # think I just proved my point...
18:26 japhb heh
18:26 japhb Why does config/auto/readline/readline.in studiously avoid using the readline headers?
18:26 japhb Wouldn't those be needed later during the build anyway?
18:27 japhb I think I would have tried to use 7 tealights in a centered hexagon arrangement -- and thereby boiled the cheese.  :-)
18:28 japhb Though I more likely would have just gone with the sterno version
18:28 japhb if it's good enough for chafing dishes, dangit, it's good enough for fondue
18:28 jonathan I should maybe mention that I'm in the south of Spain, so the air temperature is already not so low. :-)
18:29 jonathan readline - I'm sure somebody mentioned something about that looking odd on the mailing list recently...
18:30 jonathan Certainly looks odd to me.
18:36 dalek r26970 | fperrad++ | trunk:
18:36 dalek : [regex]
18:36 dalek : - fix pmc dependencies
18:36 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=26970
18:38 pmichaud jonathan++   #  my Dog $fifi .= new()   # wow.
18:40 jonathan I'm not that many features off having all the code in my Perl 6 OO talk actually runnable on Rakudo.
18:42 ambs joined #parrot
18:44 rdice That's some sweet looking OO code.
18:44 rdice Though I have to admit to being somewhat weirded out at seeing .= used in this way. :-)
18:44 pmichaud me too.  Hey, I just implement the language, I don't design it.  :-)
18:46 ambs Hi, folks
18:48 jonathan pmichaud: I like that line. ;-)
18:48 jonathan Good way to escape design questions!
18:48 ambs anybody knows how to clear the messages list purl have for me?
18:48 pmichaud it's been my motto from the beginning -- that way I tend to stay out of any heated discussions on p6l
18:48 pmichaud /msg purl messages erase   # I think
18:48 ambs pmichaud: thanks
18:49 pmichaud see also   /msg purl messages help
18:49 ambs worked
18:57 dalek r26971 | fperrad++ | trunk:
18:57 dalek : [regex]
18:57 dalek : - fix pdd17pmc (use STATICSELF instead of SELF when MMD)
18:57 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=26971
19:02 ambs_ joined #parrot
19:20 ruoso joined #parrot
19:22 Infinoid tewk: ping
19:25 Ademan joined #parrot
19:27 cognominal can someone check and commit my patch : http://rt.perl.org/rt3/Publi​c/Bug/Display.html?id=52858
19:27 cognominal with the proviso that "pir mode" should be "past mode".
19:33 Infinoid past, or pasm?
19:35 Infinoid t/compilers/imcc/syn/symbols.t will need to be fixed up for the new error string
19:38 Infinoid t/library/md5.t is failing too, but I'm not sure that's related.
19:39 cognominal arf, I am tired. pasm indeed.
19:39 Infinoid my flex output looks a bit different from what was previously there.  looks like I have a newer version
19:40 Infinoid YY_FLEX_SUBMINOR_VERSION got bumped
19:40 cognominal mine is 2.5.33
19:41 Infinoid 2.5.35 here
19:41 Infinoid well, the output compiles and most tests are passing, so I guess the flex version doesn't matter
19:41 Senaka 2.5.33 is the one I have
19:42 Senaka so it is the latest on the Gutsy apt
19:42 Senaka apt repo I mean
19:42 Infinoid t/codingstd/tabs.t                1   256     1    1  1
19:42 Infinoid t/compilers/imcc/syn/errors.t     1   256     5    1  2
19:42 Infinoid t/compilers/imcc/syn/symbols.t    4  1024     5    4  1-4
19:42 Infinoid t/library/md5.t                   1   256     6    1  4
19:43 Infinoid I can fix up the codingstd part, it's probably something to do with my having cutpasted the diff out of RT
19:43 Infinoid the compilers/imcc/ stuff looks like its just barfing on the changed error text
19:43 Senaka C++ build succeeded with tweaks
19:43 Senaka make test started
19:54 davidfetter joined #parrot
19:54 Senaka hi all
19:55 Senaka C++ build succeeded with a great deal of work
19:55 Senaka :)
19:55 Senaka now it seems that some tests are broken
19:55 Senaka I will mail this to the list
19:56 Infinoid cognominal: your patch works around an issue that we were not previously testing for.  What do you think about adding some tests to cover it?
19:57 cognominal yes, I will work on that.
19:57 cognominal how do you run all tests in a specific folder.
19:57 cognominal ??
19:57 Infinoid prove t/compilers/imcc/syn/*.t
19:57 cognominal thx
19:58 Infinoid I've fixed up the failures there, and I'll check it in in a minute, once I've verified that the md5 failure is unrelated
20:00 cognominal Infinoid, I need to install some Perl package to get prove?
20:01 kj joined #parrot
20:01 Infinoid oh, hmm.  it normally installs as part of Test::Harness I think
20:02 cognominal visibly. I am not a smoker :)
20:02 Senaka msg chromatic C++ build succeeded on rev 26966. And I mailed a report of failed tests + diff of tweaks used to parrot-porters@perl
20:02 purl Message for chromatic stored.
20:03 ambs joined #parrot
20:07 Infinoid I think maybe my flex is broken after all.
20:07 Infinoid all tests pass, except for t/library/md5.t, but that one passes without the patch
20:08 nopaste "Infinoid" at 96.238.213.50 pasted "suspiciously flex-like output from t/library/md5.t" (22 lines) at http://nopaste.snit.ch/12690
20:09 kj Infinoid: what's your flex --version output?
20:09 Infinoid flex 2.5.35
20:09 kj oh that looks good
20:10 Infinoid judging from the version defines, the previous one was generated by 2.5.34
20:10 kj i think I have 2.5.33 runnning in my cygwin install that I use when hacking imcc (i'm on win)
20:10 kj so should be fine there
20:11 Senaka http://flex.sourceforge.net/
20:11 Senaka says that 2.5.35 is the latest stable
20:12 Senaka released 26th February 2008
20:12 Infinoid kj: do you have that environment handy at the moment?
20:12 Infinoid I'm wondering if you can try to generate a parser and see if you see the weird failures I do
20:12 nopaste "Infinoid" at 96.238.213.50 pasted "imcc_disallow_pasm_dollar_variables.diff" (94 lines) at http://nopaste.snit.ch/12691
20:13 Infinoid (that's cognominal's patch from RT#52858 plus some minor tweaks)
20:13 Infinoid bbl, work &
20:13 kj I never used REJECT; do you think it's a good idea as opposed to a parrot exception message?
20:14 Infinoid cognominal: *tosses ball into your court*
20:14 Senaka I too do have some fixes for compilers/imcc/imcc.l
20:14 kj I mean, it should emit a syntax error right? (maybe that's what reject does...)
20:14 Senaka patch sent
20:15 kj I was also thinking, to split the cases for state emit and INITIAL
20:15 cognominal REJECT permits to treat the error in one place
20:15 kj but I'm not sure if that works, i.e. whether emit is only used as a PASM state
20:16 Senaka infinoid: can u also take a look at #52874's modifications to compilers/imcc/imcc.l
20:18 Infinoid Senaka: will do
20:18 Infinoid cognominal: ...but see http://nopaste.snit.ch/12690, which seems directly related to the use of REJECT
20:20 cognominal so I should not factorize
20:21 kj adding the REJECT stuff works on my box correctly
20:21 Infinoid I have no idea, I'm a flex noob.  generally, consolidation is great, but this seems to have some side effects here
20:22 nopaste "kjs" at 86.95.212.32 pasted "output $ registers in pasm mode" (3 lines) at http://nopaste.snit.ch/12692
20:22 Infinoid work &
20:22 cognominal I get the same error as Infinoid
20:23 kj weird that it does work on my box, using flex 2.5.33
20:23 Infinoid win32?
20:23 purl win32 is the God forsaken platform.  It is not fickle about slashes; its system calls do not care. or ....DEEP HURTING. or a dying legacy system collapsing under its own weight. or see also:win32 for perl
20:23 kj yes
20:23 Infinoid maybe your platform allocates bigger buffers by default, or something.
20:24 Infinoid linux/amd64 here
20:24 kj mmm. I think flex generated scanners should be pretty stable...
20:24 Senaka purl hates windows
20:24 purl Senaka: what?
20:26 kj well actually this case of disallwoing $ registers in pasm should be handled by a rule, not by an extra check like this.
20:26 kj when imcc (and IF) is being replaced by a cleaner implementation this can be fixed easily
20:31 Senaka left #parrot
20:32 dalek r26972 | fperrad++ | trunk:
20:32 dalek : [Lua]
20:32 dalek : - minor refactor
20:32 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=26972
20:50 dalek r26973 | fperrad++ | trunk:
20:50 dalek : [eclectus]
20:50 dalek : - use petite on win32
20:50 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=26973
20:55 markus_ joined #parrot
21:17 Coke joined #parrot
21:18 Coke tcl/join #perl6
21:18 Coke messages
21:18 Coke ... damint.
21:18 Coke senaka: there's a bug queue for the binary search thing on rt.
21:19 Coke msg senaka there's a bug queue for the binary search thing on rt.
21:19 purl Message for senaka stored.
21:46 sorc|laptop joined #parrot
21:46 * sorc|laptop waves
21:47 sorc|laptop i'm trying to do the parrotblog.org tutorials
21:47 sorc|laptop i'm stuck on the exercises for episode 3
21:47 sorc|laptop "Null PMC access in find_method()" << that's all i get :(
21:48 sorc|laptop the parse tree seems to be created correctly
21:51 man_in_ltop stupid xchat
21:51 Infinoid hi!  which exercise?
21:52 man_in_ltop episode 3, exercise 5
21:54 Infinoid I take it doesn't tell you which line had the null PMC?
21:55 man_in_ltop x="42" << that's my input
21:56 man_in_ltop hmm ... running it in non-interactive mode gives more verbose output
21:56 Infinoid does http://www.parrotblog.org/2008/03/sol​utions-to-exercises-in-episode-3.html help, or have you already read that?
21:56 man_in_ltop already read that
21:56 Infinoid what's the non-interactive output?
21:58 man_in_ltop http://rafb.net/p/JMzIkj74.html
21:58 pmichaud null PMC access usually means that the invocant is undefined or hasn't been set
21:59 pmichaud unfortunately we don't (yet) have a good way to map runtime errors back to the original source
21:59 man_in_ltop ok, got something
22:00 man_in_ltop in exercise 4, solution says to use .lvalue(1)
22:00 man_in_ltop removing that line from my code makes it work
22:03 man_in_ltop so then ... what does "lvalue context" mean
22:04 pmichaud lvalue means that the PAST::Var node (or whatever) is being bound/assigned to at that point
22:05 man_in_ltop ahh ... me and my non-perl background :P
22:05 man_in_ltop problem was missing $
22:06 man_in_ltop while i'm here, i see things like ":name( ~$/ )" and ":node($/)". what's the difference between $/ and ~$/?
22:08 moritz ~$/ is the stringifcation of $/
22:09 Limbic_Region joined #parrot
22:09 man_in_ltop hmm ok
22:10 man_in_ltop i thought it was something like that
22:10 wknight8111 joined #parrot
22:15 man_in_ltop honestly, i don't think i'm fully understanding the tutorial :)
22:15 moritz it assumes some Perl 6 knowledge I guess
22:16 man_in_ltop most perl i've done is hacking a plugin for irssi that does a scrolling topic bar :P
22:16 moritz and that was Perl 5
22:16 man_in_ltop someone else did all the tricky stuff, i just hacked it so it would resize when the window resized
22:17 teknomunk joined #parrot
22:17 GNUtoo joined #parrot
22:19 teknomunk_ joined #parrot
22:19 GNUtoo hello, i've just installed parrot...now i'd like to run perl5,perl6 and python code...how do i do that?
22:20 teknomunk_ joined #parrot
22:21 wknight8111 I know perl 6 is still in active development, I dont know the current status of the Perl 5 and Python projects
22:21 jonathan GNUtoo: What did you install, exactly?
22:22 GNUtoo jonathan, gentoo's parrot 0.4.6
22:22 GNUtoo mabe i should try to install a more recent version...
22:22 jonathan Ouch.
22:22 jonathan Yes, we're on 0.6.0 now.
22:22 jonathan A lot has changed since 0.4.6.
22:23 jonathan Do you have svn?
22:23 GNUtoo yes
22:23 jonathan if so
22:23 jonathan svn co http://svn.perl.org/parrot/trunk/ parrot
22:23 jonathan cd parrot
22:23 jonathan perl Configure.pl && make perl6
22:23 jonathan Will build Parrot, the Perl 6 compiler and make a perl6 executable.
22:24 GNUtoo ok then i cd into and ./ to execute it cause i don't want to mess up my system...thanks a lot
22:25 jonathan make install is not really recommended at the moment, by the way. We're still very much developing at the moment, and there are known issues with doing make install.
22:26 GNUtoo just an idea: i saw in the fosdem talk video that it took some hours to make a python compiler...is it possible for someone that knows perl5,java and python to make a mips assembly interpretator?
22:27 moritz GNUtoo: the "some hours" are for the basic constructs. For a full featured compiler you still need more time
22:27 Infinoid GNUtoo: probably... but most of the work there would be in emulating various bits of hardware, I think
22:27 GNUtoo ok
22:34 * jonathan sleeps
22:44 skids joined #parrot
22:58 kid51 joined #parrot
22:58 zarchne joined #parrot
23:21 tetragon joined #parrot
23:29 dalek r26974 | particle++ | trunk:
23:29 dalek : [RELEASE] NEWS updates
23:29 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=26974
23:30 wknight8111 Was there a release scheduled for today?
23:30 kid51 Releases are on the 3rd Tuesday of each month.
23:31 wknight8111 so...tomorrow?
23:31 kid51 particle is on the road and is doing this and that ahead of time.
23:31 wknight8111 i'm so frazzled with school that I can't keep track of even the simplest things
23:35 Infinoid yes, tomorrow
23:35 purl o/~ the sun will come out.. tomorrow.. o/~ or the National Day of Slayer, and the National Emo Kid Beatdown day, by http://www.nationaldayofslayer.org/ and http://community.livejourna​l.com/wtf_inc/2805832.html or tomorrow and tomorrow and tomorrow creeps in this petty pace or mañana or free for all or another day
23:51 dalek joined #parrot

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

Parrot | source cross referenced