Camelia, the Perl 6 bug

IRC log for #parrot, 2008-02-16

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:33 svnbotl r25747 | jkeenan++ | tcif:
00:33 svnbotl : Change my @steps to our @steps to make it assignable in test files.
00:33 svnbotl r25748 | jkeenan++ | tcif:
00:33 svnbotl : First tests for Parrot::Configure::Parallel::Trace methods, plus a little refactoring in module.
00:33 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25748
01:11 teknomunk joined #parrot
01:13 AndyA joined #parrot
01:34 japhb joined #parrot
02:37 Khisanth joined #parrot
02:44 Alias joined #parrot
02:49 svnbotl r25749 | jkeenan++ | trunk:
02:49 svnbotl : Add CLEANUP => 1 to tempdir() calls in two locations.
02:49 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25749
02:51 cotto How does an Iterator PMC know how to iterate over an aggregate PMC?
02:56 svnbotl r25750 | jkeenan++ | tcif:
02:56 svnbotl : Add CLEANUP => 1 to tempdir() calls in two locations.
02:56 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25750
02:59 cotto I can tell it has something to to with key PMCs, but it seems magical right now.
03:07 Coke when you create the iterator, you give the PMC you care about. Everytime you shift a value off the iterator, returns the value associated with it's current keyed value, and saves the next value to be pulled internally.
03:07 Coke (so each time you peek the next key (hash) or index (array), save that, and then return the value associated with the key you kept last time. (or when you init'd)
03:07 Coke the trick is VTABLE_nextkey_keyed
03:09 Coke (this just based on a quick read of shift_* in src/pmc/iterator.pmc
03:21 svnbotl r25751 | jkeenan++ | tcif:
03:21 svnbotl : Move struct declaration out of Configure.pm and Parallel.pm and into Configure/Base.pm.
03:21 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25751
03:22 c_otto Coke++
03:22 c_otto Thanks.
03:24 c_otto Also, particle++ for the help yesterday
03:28 svnbotl r25752 | jkeenan++ | tcif:
03:28 svnbotl : Add more tests for P::C::Parallel::Trace.
03:28 svnbotl r25753 | jkeenan++ | tcif:
03:28 svnbotl : Delete commented-out code.  Specify exact number of tests in plan.
03:28 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25753
03:30 Andy joined #parrot
03:50 Andy howdy at FP
03:58 Alias FP?
03:58 purl FP is probably fp. it's not mmx or frontpage or functional programming or floating point or french press or first post or frozen perl
03:58 Andy There is apparently a Cretinous Teenager convention in this hotel.
04:12 svnbotl r25754 | jkeenan++ | tcif:
04:12 svnbotl : 1.  Add get_storable_file() method to P::C::Parallel::Trace.  Add tests
04:12 svnbotl : for that method to 061-parallel.t.
04:12 svnbotl : 2.  Modify Parallel::Trace::update_state() to require a 'step_name'
04:12 svnbotl : argument.  This is pushed on to the array which in turn is the first
04:13 kid51 purl:  FP is Frozen Perl
04:13 purl ...but fp is fp. it's not mmx or frontpage or functional programming or floating point or french press or first post or frozen perl...
04:13 kid51 FP?
04:13 purl FP is fp. it's not mmx or frontpage or functional programming or floating point or french press or first post or frozen perl
04:14 kid51 purl, repeat after me:  FP is Frozen Perl
04:14 purl OK, kid51.
04:14 kid51 FP
04:14 kid51 FP?
04:14 purl FP is fp. it's not mmx or frontpage or functional programming or floating point or french press or first post or frozen perl
04:14 kid51 You still haven't gotten it, have you, purl?
04:14 purl kid51: huh?
04:15 * kid51 must sleep
04:15 purl $kid51->sleep(8 * 3600);
04:25 SCalimlim joined #parrot
05:23 confound joined #parrot
05:27 confound joined #parrot
05:40 simcop2387 joined #parrot
05:43 Andy joined #parrot
06:21 jjore joined #parrot
06:36 petdance joined #parrot
07:00 felipec joined #parrot
07:00 felipec hi
07:01 felipec I would like to create a simple shell interpreter in C, but I'm interested on not re-inventing the wheel and using some parser generator tools
07:01 felipec do you think parrot might help?
07:04 cotto_ joined #parrot
07:18 Ademan joined #parrot
08:53 Tene felipec: it might.
09:52 HG` joined #parrot
09:52 barney joined #parrot
10:01 wknight8111 joined #parrot
10:21 ruoso joined #parrot
10:34 kj joined #parrot
11:00 iblechbot joined #parrot
11:04 wknight8111 joined #parrot
11:09 svnbotl r25755 | kjs++ | trunk:
11:09 svnbotl : [pdd06] add note about a pasm grammar in bison. This might be handy when this pdd is being reviewed and updated.
11:09 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25755
11:36 ptc joined #parrot
11:54 jq joined #parrot
12:56 JamesA joined #parrot
12:58 JamesA Can I ask a quick question that will save me hours of reading and testing. Is it possible to disable certain operators within parrot (e.g. disable io). Must this be done to the source code or can it be done similar to that in Opcodes.pm? (I am looking at the perl5 embeded parrot).
12:59 JamesA Opcodes - http://search.cpan.org/~rgarcia/​perl-5.10.0/ext/Opcode/Opcode.pm
12:59 shorten JamesA's url is at http://xrl.us/bgbga
13:06 cognominal_ JamesA, I know that Parrot operator are classified by category. I don't know if anything is done to disable t compile or runtime.
13:06 cognominal_ at least this is probably a good hook to implement that feature if (probably) missing
13:09 kid51 joined #parrot
13:18 masak joined #parrot
13:22 JamesA thanks for your help, I'll examine the source and see what I can find
13:22 JamesA left #parrot
13:36 cognominal_ you are welcome, #parrot is a friendly place
13:36 cognominal_ oops
14:10 Ron joined #parrot
14:39 svnbotl r25756 | jkeenan++ | trunk:
14:39 svnbotl : Parrot::Configure::add_steps():  Modify so that instead of assigning the list
14:39 svnbotl : of new steps en masse to the list_of_steps attribute, push them individually
14:39 svnbotl : onto the array which is the value of that attribute -- thereby averting wiping
14:39 svnbotl : out any steps which are already present in that attribute.  This should have
14:39 svnbotl : onto the array which is the value of that attribute -- thereby averting wiping
14:39 svnbotl : out any steps which are already present in that attribute.
14:39 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25757
14:46 svnbotl r25758 | jkeenan++ | tcif:
14:46 svnbotl : Minor corrections in update_state().
14:46 svnbotl r25759 | jkeenan++ | tcif:
14:46 svnbotl : Add a third test file to test Parrot::Configure::Parallel::Trace to test
14:46 svnbotl : case where two different configuration steps are tested.
14:53 Ademan joined #parrot
14:55 HG` joined #parrot
14:55 kid51 Hmm, svnbotl just mushed together two separate (though similar) commit messages.
14:57 skids /me eyeballing http://judy.sourceforge.net/doc/index.html for use in Sets/Bags/Multidim Hashes
14:58 skids which raises a ton of questions
14:58 skids like, how important is running mallocs through the GC and/or making them relocatable.
14:59 skids and licensing/maint issues -- probably best to keep a frozen version in the core, right?
15:01 skids Also, how fast/memory efficient etc need be freeze/thaw which I guess amounts to what are people using freeze/thaw for.
15:17 wknight8111 joined #parrot
15:32 amoore joined #parrot
15:45 ptc joined #parrot
15:55 Andy joined #parrot
15:56 * Andy is sitting in the FP Parrot talk
15:56 kid51 who is giving it?
15:57 Andy Allison
15:58 Andy http://www.frozen-perl.org/mpw2008/schedule
16:04 Andy Every time Allison mentions Lisp, drink!
16:08 davidfetter joined #parrot
16:13 wknight8111 joined #parrot
16:19 svnbotl r25761 | petdance++ | trunk:
16:19 svnbotl : remove unused return
16:19 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25761
16:19 c_otto Does anyone know about compatibility between the Zend license 2.00 (basically modified BSD w/ advertising clause) and the Artistic License?
16:20 Andy "All the rest of teh continuation context will get garbage collected" says Allison.
16:20 Andy IF OUR GARBAGE COLLECTOR WORKED
16:20 Andy :-)
16:22 Andy She's showing the architecture of PGE, PCT, PASM, PIR
16:22 Andy BUT WHERE IS MY HEADERIZER?
16:23 Andy Damn, ghettoized again!
16:23 * davidfetter lives in the ghetto
16:24 davidfetter would monsieur care for some greens and a 40?
16:24 svnbotl r25762 | petdance++ | trunk:
16:24 svnbotl : removed an unused return, and an extra cast
16:24 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25762
16:25 Andy can i get some grey poupon?
16:26 davidfetter certainly. it's carried at every ghetto mart i've been in in the past 5 years
16:26 Andy oh no she made finger quotes!
16:26 davidfetter apart from "doesn't work yet," what's the state of parrots' GC?
16:26 davidfetter parrot's*
16:28 Andy It's just leaky and causes some segfaults at times.
16:28 Andy chromatic's done a lot of work on it recently
16:28 davidfetter :)
16:31 Andy Whoo, a Rakudo mention!
16:31 svnbotl r25763 | paultcochrane++ | trunk:
16:31 svnbotl : [config] Applied patch from Andy Dougherty in RT#50066.  This now fixes the
16:31 svnbotl : issue whereby LIBPARROT_STATIC wasn't being expanded properly and therefore
16:31 svnbotl : caused Solaris to fail to build.
16:31 svnbotl r25764 | petdance++ | trunk:
16:31 svnbotl : consting argv, which quiets warnings, too
16:31 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25764
16:32 davidfetter Andy, url?
16:32 Andy for what?
16:32 Andy Rakudo?
16:32 purl i heard Rakudo was The Way Of The Camel. or having some failing tests as of 22JAN08 or Gotta Catch 'Em All or in languages/perl6 (see http://parrotvm.org/svn/parrot​/browse/trunk/languages/perl6) or http://use.perl.org/~pmichaud/journal/35400
16:32 Andy purl, Rakudo is also http://rakudo.org
16:32 purl okay, Andy.
16:33 davidfetter Andy, for the mention :)
16:33 Andy She's giving a talk
16:33 Andy at FP
16:34 * davidfetter slow this morning, as coffee is still metabolizing
16:34 davidfetter FP?
16:34 purl i think FP is fp. it's not mmx or frontpage or functional programming or floating point or french press or first post or frozen perl
16:37 * davidfetter went to allison's talk at SCALE. it was good, the half of it that the audience let her give :P
16:43 wknight8111 joined #parrot
16:43 svnbotl r25765 | paultcochrane++ | trunk:
16:43 svnbotl : [src,docs] Applying patch in RT#48142 which removes the deprecated
16:43 svnbotl : class_type vtable entry.
16:43 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25765
16:50 DarkWolf84 joined #parrot
16:50 svnbotl r25766 | petdance++ | trunk:
16:50 svnbotl : turning NOTNULLs into ARGMOD
16:50 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25766
16:55 Limbic_Region joined #parrot
17:01 Andy joined #parrot
17:02 rdice joined #parrot
17:15 amoore purl, FP is also Frozen Perl: http://www.frozen-perl.org/mpw2008/
17:15 purl okay, amoore.
17:16 svnbotl r25767 | paultcochrane++ | trunk:
17:16 svnbotl : [core,t] Applied patch supplied in RT#42375, with the "I am a class" test
17:16 svnbotl : modified as mentioned by Allison.  This patch removes the deprecated pmcinfo
17:16 svnbotl : opcode.
17:16 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25767
17:24 cognominal_ can someone explains me the difference between the dynamic and the static type of a pmc?
17:33 diakopter joined #parrot
17:41 bln joined #parrot
17:44 Khisanth joined #parrot
17:45 parrot-poke joined #parrot
17:51 HG` joined #parrot
17:56 * Coke still doesn't know why this warning is occurring.
17:58 Khisanth joined #parrot
18:08 svnbotl r25768 | petdance++ | trunk:
18:08 svnbotl : changed an endless while to a do/while
18:08 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25768
18:15 kid51 joined #parrot
18:23 particl1 joined #parrot
18:26 * jonathan made it back from Germany, with his bag. Despite the airport making their best efforts to convince me it was lost, then pulling it ouf of a random room... :-)
18:27 * ptc <-- afk
18:45 Andy joined #parrot
18:45 Andy back
18:45 svnbotl r25769 | jkeenan++ | tcif:
18:45 svnbotl : Add 064-parallel.t to provide additional coverage of Parallel::Trace.  Eliminate commented-out code from init_manifest-01.t.
18:45 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25769
18:52 svnbotl r25770 | jkeenan++ | tcif:
18:52 svnbotl : Simplify name of subroutine from 'store_this_step_pure' to 'store_this_step'.
18:52 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25770
18:52 Khisanth joined #parrot
19:20 bgeron is http://www.parrotcode.org/source.html generated from POD in SVN?
19:27 bgeron anyway, after #subversion: "Subversion aims is to be" has an "is" too many
19:32 mire joined #parrot
19:42 svnbotl r25771 | jkeenan++ | tcif:
19:42 svnbotl : Begin to convert configuration step tests to parallel trace mode.
19:42 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25771
19:55 svnbotl r25772 | rblasch++ | trunk:
19:55 svnbotl : [jit] Applied patch RT#44765.  This avoids reusing the interpreter argument on
19:55 svnbotl : the stack, which doesn't work if the compiler uses the stack space, e.g.
19:55 svnbotl : optimized VC++ build.
19:55 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25772
20:01 svnbotl r25773 | chromatic++ | trunk:
20:01 svnbotl : [t] This test file referred to pmcinfo.pasm, which r25767 removed.
20:01 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25773
20:02 guru joined #parrot
20:06 toft joined #parrot
20:06 jasonmay joined #parrot
20:10 iblechbot joined #parrot
20:21 particle joined #parrot
20:27 particl1 joined #parrot
20:29 chromatic joined #parrot
20:30 chromatic Our GC isn't leaky, though.
20:30 chromatic I haven't seen a GC segfault on any platform that I own in quite a while either!
20:34 cognominal_ > can someone explains me the difference between the dynamic and the static type of a pmc?
20:35 cognominal_ chromatic?
20:35 cognominal_ I discover a lot evolution since I checked out mcs,  roles, attributes...
20:36 IllvilJa joined #parrot
20:39 cognominal_ hum roles are specified but no implemented
20:39 cognominal_ ack++
20:39 Andy heh
20:39 Andy why now?
20:40 chromatic cognominal_, I'm not sure myself.  The static type may resolve at compile time to a direct C function call, while the dynamic type may need runtime lookup.  That's just a guess though.
20:42 cognominal_ Andy, cuz it help me to see that something is specced and not implemented
20:42 chromatic Wow, Tcl is done.
20:42 Andy hey c
20:43 chromatic yeah Andy?
20:44 Andy just syain' hi
20:44 Andy I'm trying to get libgmalloc to be useful for me on my Mac
20:45 chromatic You crazy Mac kids.
20:45 cognominal_ I own an Intel Macbook.  I have been disappointed by the hardware, but I love the software
20:46 cognominal_ and underneath that's an Unix.
20:47 chromatic A Unix that shows the best architectural decisions of BSD 4.3 circa NeXT circa 1987, yes.
20:49 cognominal_ I was at Sun when they merged all the ATT System V crap
20:50 Andy But I need to download XCode 3.0 and it's 1.1G
20:50 cognominal_ bad memories
20:50 cognominal_ Andy: it is worth it, not that I am a Xcode and Interface builder specialist
20:51 Andy yeah, I know, I just don't want to hammer the wifi here
20:51 cognominal_ cocoa cast will convince you if you doub it
20:51 cognominal_ too bad, perl is not hooked to Xcode
20:53 cognominal_ I think Apple is will eventually choose a dynamic language, so far these are Python and Ruby. I hope Parrot will make the difference
20:53 cognominal_ AppleScript is such a crap
20:53 cognominal_ if Perl is write only, AppleScript is read only :)
20:54 cxreg php is delete only
20:55 cognominal_ well, a lot of nice apps are built on top of php, I wish that Perl can claim as much
20:56 cognominal_ not that I want to touch the language with a ten foot pole
20:57 Andy hey chromatic, did you see my "make apilist" target?
20:57 Andy and make malloclist?
20:59 zaphod joined #parrot
21:01 petdance joined #parrot
21:09 chromatic I did, that's very nice.
21:10 chromatic Anyone want to trace a segfault in Tcl?
21:16 parrot-poke apple is on llvm
21:16 parrot-poke that makes c a dynamic language for them, especially woth objc
21:17 parrot-poke Xcode is very c centric
21:20 cognominal_ llvm?
21:20 purl i heard llvm was a big huge thing with not much docs, but it works or low-level virtual machine or at http://www.llvm.org
21:21 parrot-poke low level virtual machine allows them to optimize and run code on all kinds of hardware
21:21 parrot-poke solves their problem for them, might have interesting lessons for parrot
21:21 parrot-poke web site has a lot of docs and discussion
21:22 chromatic Allison and I both know Chris and Tanya Lattner, who work on LLVM.
21:22 chromatic I helped them move a couch x-odd years ago, before they went to graduate school.
21:51 Ron_ joined #parrot
22:09 mj41_ joined #parrot
22:13 rafl joined #parrot
22:17 Khisanth joined #parrot
22:19 chromatic Wow... parrot_split_path_ext.
22:29 Andy joined #parrot
22:33 wknight8111 joined #parrot
23:08 Andy joined #parrot
23:09 chromatic Andy, check out  parrot_split_path_ext.
23:09 chromatic It could use some lurve.
23:13 Andy joined #parrot
23:19 Coke chromatic++ fixin' the website...
23:19 svnbotl r25774 | chromatic++ | trunk:
23:19 svnbotl : [JIT] Fixed an implicit cast C++ warning; improved the other declarations in
23:19 svnbotl : this function.  PPC only.
23:19 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=25774
23:25 chromatic I'm trying to fix Tcl too.
23:28 chromatic It looks like the char * value of the name of a dynamically loaded library goes missing somehow.
23:28 chromatic Only during the bytecode thaw... I think.
23:28 chromatic Coke, which shootout fails for you on PPC?  10?
23:32 Coke (Tcl is done) what?
23:33 chromatic My mistake; I couldn't get any tests to run, but I had the PDD 17 branch patch applied locally.
23:33 Coke chromatic: I don't think shootout fails for me.
23:33 chromatic On Intel or PPC?
23:34 Coke chromatic: ... tcl runs in both trunk and pdd17pmc for me, last time I checked. It just segfaults most of the time.
23:34 chromatic Is the segfault in comparing two strings, where rhs has an invalid strstart?
23:34 Coke in ascii_compare, ja.
23:35 Coke ISTR there was another one; both had tickets, though.
23:35 Coke checking...
23:35 Coke (sent you a BT on the intlist failure.)
23:35 Coke er... macro failure. not intlist.
23:37 chromatic The problem is at stack frame #7; somehow macro_name is invalid.
23:37 chromatic Probably the part where I removed macro buffers from IMCC.
23:37 chromatic have to go though....
23:39 Coke noooo
23:39 chromatic_away It should be an easy fix; just make sure that macro_name is valid.
23:39 chromatic_away Everything after that point is just noise.  That's the problem.
23:39 Coke I'll poke.
23:40 Coke ascii_compare is #49836
23:40 chromatic_away Should be r25546 that caused the problem.
23:40 Coke other issue was #50398; started as a segfault, but now it's "I don't understand pdd17."
23:47 kid51 joined #parrot
23:48 Coke I don't think that's the problem. I think we're reading past the end of the file.
23:49 Coke (based on the test that's failing.)
23:55 Coke oh. no, that's the code path, but you're right.

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

Parrot | source cross referenced