Camelia, the Perl 6 bug

IRC log for #parrot, 2008-03-10

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:34 Theory joined #parrot
00:37 mdiep joined #parrot
00:40 Andy joined #parrot
00:40 kid51 joined #parrot
01:08 svnbotl r26278 | allison++ | pdd17pmc:
01:08 svnbotl : [pdd17pmc] Recommitting patch to remove pobj_t substruct abstraction, plus reorder two struct members in parrot_string_t struct to fix alignment problem on Windows MSVC when copying COW strings.
01:08 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=26278
01:17 Theory joined #parrot
01:35 Tene Hmm... no traffic on planet.parrotcode.org in a decent while.  If I were to generate a use.perl blog post, who would I harass to get it up there?  Coke?
01:40 svnbotl r26279 | chromatic++ | pdd17pmc:
01:40 svnbotl : [PMC] Removed several morph() entries that didn't do anything useful.  The
01:40 svnbotl : default behavior is almost always correct.
01:40 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=26279
01:46 skv_____ joined #parrot
01:54 skv_____ joined #parrot
02:15 peepsalot joined #parrot
02:21 Ademan joined #parrot
03:50 svnbotl r26280 | chromatic++ | pdd17pmc:
03:50 svnbotl : [WML] Updated WmlsBoolean PMC to match PDD 17.
03:50 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=26280
03:57 TimToady joined #parrot
03:57 svnbotl r26281 | chromatic++ | pdd17pmc:
03:57 svnbotl : [WML] Made WmlsFloat PMC conform to PDD 17.
03:57 svnbotl r26282 | chromatic++ | pdd17pmc:
03:57 svnbotl : [WML] Ported WmlsInvalid PMC to PDD 17.
03:57 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=26282
04:02 katie_d I cannot figure out what obj_file in imcc_run() is for, anyone know where I could look for more info?  I'm particularly interested in what the purpose of this line is:   if (!interp->lo_var_ptr) {interp->lo_var_ptr = (void *)&obj_file;}   ... I know lo_var_ptr is used for stack tracing. objfile is an int that was just declared but not defined -- so I can't see how they are connected. Anyone have any idea about this?
04:12 PerlJam Given a STRING*, how do I do the equivalent of isalpha() on the characters of the string?  I know I can use string_ord() to get at an individual character, but is there some special encoding-aware function that I need to use to determine if it's an alpha, numeric, etc. ?
04:14 PerlJam katie_d: I'd guess that it's a bug.
04:15 PerlJam katie_d: it's been a while since I've done much C coding, but taking the address of a local variable and storing it somewhere has never been a good thing.
04:16 svnbotl r26283 | chromatic++ | pdd17pmc:
04:16 svnbotl : [WML] Made test target depend on successful build.  (That explains why my
04:16 svnbotl : changes didn't seem to have an effect....)
04:16 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=26283
04:17 PerlJam katie_d: though, I suppose if it's only used within the imcc_run() function, it may be okay.
04:28 svnbotl r26284 | chromatic++ | pdd17pmc:
04:28 svnbotl : [WML] Ported WmlsString PMC to PDD 17.
04:28 svnbotl r26285 | chromatic++ | pdd17pmc:
04:28 svnbotl : [PMC] Ported WmlsInteger PMC to PDD 17.
04:28 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=26285
04:35 svnbotl r26286 | petdance++ | trunk:
04:35 svnbotl : Function pointers take NOTNULL, not ARGIN
04:35 svnbotl r26287 | petdance++ | trunk:
04:35 svnbotl : more splint target work
04:35 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=26287
04:45 katie_d PerlJam: OK, I didn't know if there was some magic going on that I didn't understand. Should I post a bug report, you think? At least I'll ask if someone can document what it does if it's not a bug ...
04:54 PerlJam katie_d: yes, at least ask.
04:54 svnbotl r26288 | petdance++ | trunk:
04:54 svnbotl : rename tosize to newsize, since "to*" is reserved
04:54 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=26288
05:07 svnbotl r26289 | petdance++ | trunk:
05:07 svnbotl : rename tosize to newsize, since "to*" is reserved
05:07 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=26289
05:07 purl i already had it that way, svnbotl.
05:22 Infinoid heh
05:51 svnbotl r26290 | petdance++ | trunk:
05:51 svnbotl : Don't complain about ints as boooleans
05:51 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=26290
05:56 katie_d Hmmm. I think I might have figured it out. lo_var_ptr is used by trace_system_stack() -- it is where memory tracing begins -- and obj_file is the first thing declared in imcc_run(), so would it not be at the bottom of the stack within imcc_run()? It seems that they are just using that as a marker so that stack tracing will start after the args for imcc_run() ... this should definitely be documented with a one line comment above the c
05:57 Tene katie_d: cut off after "above the c..."
05:57 katie_d above the conditional.
05:58 Tene katie_d: http://www.irssi.org/scripts/scripts/splitlong.pl
05:58 katie_d Tene: Excellent. Thank you :)
05:59 Tene np
05:59 Tene katie_d: nice work on figuring out that part of imcc
05:59 Tene You should open an rt ticket about it
06:00 Tene purl: parrotbug
06:00 purl parrotbug is, like, mailto:parrotbug@parrotcode.org or http://svn.perl.org/parrot/​trunk/docs/submissions.pod
06:00 katie_d Tene: Yes. I already opened one up requesting that it be documented.
06:00 Tene Ahh.
06:41 svnbotl r26291 | petdance++ | trunk:
06:41 svnbotl : Instrumenting malloc & free arguments so splint can accurately check for memory tracking
06:41 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=26291
06:42 iblechbot joined #parrot
06:43 teknomunk joined #parrot
07:06 teknomunk joined #parrot
07:25 svnbotl r26292 | fperrad++ | trunk:
07:25 svnbotl : [Lua]
07:25 svnbotl : revert r25931, I want nil as output value
07:25 svnbotl r26293 | fperrad++ | trunk:
07:25 svnbotl : [Lua]
07:26 uniejo joined #parrot
07:32 svnbotl r26295 | fperrad++ | trunk:
07:32 svnbotl : [win32]
07:32 svnbotl : add support for strawberry distribution (http://strawberryperl.com/)
07:32 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=26295
08:29 AndyA joined #parrot
08:51 kj joined #parrot
10:26 iblechbot joined #parrot
10:26 wknight8111 joined #parrot
10:26 ruoso joined #parrot
12:04 ruoso joined #parrot
12:29 UltraDM joined #parrot
12:30 peepsalot joined #parrot
12:32 ask_ joined #parrot
12:38 cout joined #parrot
13:05 gryphon joined #parrot
13:32 wknight8111 joined #parrot
13:39 Coke http://groups.google.com/group/cn.bbs.comp.la​ng.perl/browse_thread/thread/fd8a56d0b1a2c27c​/90631ea4f76d1430?hl=en&q=parrot+-site%3A​perl.org+-site%3Aparrotcode.org+-site%3Acpan.​org+-perl.perl6+-perl.cvs#90631ea4f76d1430
13:39 shorten Coke's url is at http://xrl.us/bhf5y
13:42 Coke "Dear Jeff Goff, This is a computer-generated report for parrot-0.0.8.1 on
13:42 Coke perl 5.10.0, created by CPAN-Reporter-1.11."
13:43 particle yikes
13:43 Coke (some random link triggered by a google news alert)
13:48 Coke hurm. the failures in tcl branch are slightly different than those in trunk.
13:53 jhorwitz joined #parrot
14:15 katie_d joined #parrot
14:19 rdice joined #parrot
14:36 Coke hurm. I think get_string in src/pmc/tcllist.pmc needs to be PDD17-erized.
14:37 Coke (my attempts to do so generate many more test failures that we have currently, though. =-)
15:06 * jhorwitz is happy all mod_parrot tests pass against PDD17 branch
15:06 davidfetter joined #parrot
15:35 Andy joined #parrot
15:44 * Coke tries to boil this latest failure down to some PIR for chromatic's sake.
15:46 parrot-poke joined #parrot
15:53 Coke rant: #line 444 "/home/coke/sandbox/pdd17pmc/tools/buil​d/../../lib/Parrot/Pmc2c/PCCMETHOD.pm"
15:53 Coke ... this seems to cause GDB to display the perl source when debugging.
16:00 Coke It would be most helpful if there was a way to invoke Dumper() on a PMC from gdb. any clues?
16:06 Infinoid you can say "print <c-expression>" to make gdb call a C function with specified arguments and print its return value
16:06 wknight8111 joined #parrot
16:06 parrot-poke joined #parrot
16:06 katie_d joined #parrot
16:07 Infinoid and if that C function prints some stuff as a side effect, you'll see that too
16:11 confound joined #parrot
16:11 dngor joined #parrot
16:12 particle joined #parrot
16:12 clunker joined #parrot
16:15 Coke Yes, but all the dumper stuff is a parrot function, not a C function, which  makes that much more complicated. :|
16:16 katie_d Coke: You can call anything in the symbol table w/ print.
16:16 particle i wonder if pdb would be more help than gdb
16:19 confound joined #parrot
16:20 confound joined #parrot
16:21 Coke katie_d: print (p?) just shows the top level struct info for a PMC. I want something that if I have a pmc that does list, it prints out 0: "first element"\n1: "second element", etc.
16:21 confound joined #parrot
16:21 Coke if pdb helped, I could just write PIR to dump it for me. My problem is trying to analyze a PMC that's in a method written in C, while I'm still in that method.
16:22 Coke meh. I think I've tracked down what I need, but lunch and work beckon.
16:22 Infinoid Coke: chromatic would probably know how to do this.  I'm not familiar at all with the parrot C API, sorry
16:24 particle oh my god, it's an infinoid!
16:24 particle where'd your tuits go?
16:24 confound joined #parrot
16:24 dngor joined #parrot
16:26 Infinoid my brain is jumbled mess... spaghetti tuits
16:26 clunker joined #parrot
16:26 particle you'll be sailing soon enough again. it'll all be okay.
16:26 Infinoid did that last weekend, have plenty of bruises and sunburn to prove it
16:27 particle sheesh. next time let the wind do the work!
16:27 Infinoid when the wind gets blowing, *everything* gets harder to do :)
16:45 svnbotl r26296 | fperrad++ | pdd17pmc:
16:45 svnbotl : {Lua]
16:45 svnbotl : accept the output "(unknown file):-1"
16:45 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=26296
16:55 marmic joined #parrot
17:36 Coke Infinoid: I'm trying not to bug chromatic. =-)
18:07 Dave joined #parrot
18:09 Psyche^ joined #parrot
18:12 cotto_ ack
18:12 cotto_ ack?
18:12 purl i think ack is at http://search.cpan.org/dist/ack/ or at http://petdance.com/ack/
18:12 Andy ack!
18:21 cotto_ awesome tool, btw
18:21 barney joined #parrot
18:21 cotto_ soooo much nicer than a bunch of greps
18:33 IllvilJa joined #parrot
18:36 svnbotl r26297 | coke++ | pdd17pmc:
18:36 svnbotl : [tcl]
18:36 svnbotl : Add VTABLE marker to get_string
18:36 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=26297
18:37 kj joined #parrot
18:44 Coke joined #parrot
18:47 Theory joined #parrot
18:48 Coke anyone have an example of invoking a method from C?
18:48 particle a METHOD ?
18:48 purl a method is a good start
18:48 svnbotl r26298 | coke++ | pdd17pmc:
18:48 svnbotl : [docs]
18:48 svnbotl : remove reference to obsolete "METHOD" style in branch; it's gone here.
18:48 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=26298
18:48 Coke yes.
18:49 particle i'm in the midst of debugging something. will look in a sec
18:50 Coke eh. no rush.
18:55 kj good evening
18:56 kj i seem to remember that ANSI control characters have worked some day in the past (printing them using PIR print op)
18:56 kj but unfortunately i can't get it working anymore :-(
18:56 kj (otherwise I have a nice Game of Life demo :-)
18:57 Coke should work, ja.
18:57 Coke how are they not working for you?
18:57 kj I once did it with binary: prefix
18:57 kj i just get ugly tokens on my screen, just the letters of the control character
18:58 kj a few years ago, I used binary:"\e[H"
18:59 kj and binary:"\e[2J" too
18:59 kj one is for clear screen, other for cursor I think
19:00 particle coke: see src/pmc/exporter.pmc ~322 (METHOD import)
19:01 particle funny that we still have PCCINVOKE
19:01 particle after PCCMETHOD and PCCRETURN have been converted
19:01 Coke ... If you call allison on it, she'll probably update it.
19:01 particle i can do it myself
19:01 particle just need an hour or so to go through the code
19:02 particle (not now)
19:04 barney no control chars in  examples/pir/life.pir
19:05 kj barney: yeah I know. I stole Lua's life
19:05 kj I ported it to my little tutorial-scripting language
19:05 kj I'm this >< close..
19:10 sjansen joined #parrot
19:14 Coke PCCINVOKE(INTERP, P0, "replace", STRING *orig, STRING *repl);
19:14 Coke This converts to:
19:15 marmic joined #parrot
19:17 Coke Parrot_PCCINVOKE(interp, P0, CONST_STRING(interp, "replace"), "SS->", orig, repl);
19:18 Coke which then dies on build with:
19:18 Coke ./tcllist.pmc:157: warning: passing argument 3 of 'Parrot_PCCINVOKE' makes pointer from integer without a cast
19:18 Coke ./tcllist.pmc:116: error: incompatible types in assignment
19:18 Coke does CONST_STRING even work in dynpmcs?
19:18 particle oh, you can't do that.
19:19 Coke I can't do what... use the PCCINVOKE shortcut?
19:19 particle the PCCINVOKE macro expects <type> <name>
19:19 particle so CONST_STRING(interp, "foo") is no good there
19:20 Coke I didn't write const_string. the pmc2c generator did.
19:20 particle STRING * s_replace = CONST_STRING(interp, "replace");
19:20 particle hrmm. P0 is SELF ?
19:20 Coke no.
19:21 Coke it's the result of a pmc_new, earlier.
19:21 particle ok, but it's a pmc
19:21 Coke yup
19:21 Coke lemme try this directly with the macro.
19:21 particle with a strange pir-like name. ok. which pmc? tcllist.
20:09 grim_fandango joined #parrot
20:10 x___ joined #parrot
20:12 allison joined #parrot
20:13 Coke allison: hio
20:13 allison hi coke
20:15 Coke Jerry was wondering why it's spelled PCCINVOKE. =-)
20:20 Coke there. tcl is now passing more tests in pdd17 than in trunk.
20:20 Coke ... whee.
20:20 svnbotl r26299 | coke++ | pdd17pmc:
20:20 svnbotl : [tcl]
20:20 svnbotl : String->replace() is now a PCC METHOD and not a C METHOD;
20:20 svnbotl : call it appropriately.
20:20 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=26299
20:42 Coke hurm. the {expand} syntax is broken in tcl. (trunk)
20:43 rdice joined #parrot
20:44 * particle wonders if a C sub inside pmclass {} can be marked VTABLE *and* METHOD
20:54 Coke I doubt it.
20:54 Coke VTABLEs in C don't follow the PCC.
20:54 particle figured.
20:54 Coke s/don't/don't seem to/
20:55 particle it's a small difference between PIR and C class definitions
20:58 Coke I wonder if you can have VTABLE get_string and METHOD get_string in the same PMC.
20:58 Coke if so, then the difference is just that PIR gives you sugar.
21:20 particle corrcet
21:20 particle unlike my speling
21:59 svnbotl r26300 | chromatic++ | pdd17pmc:
21:59 svnbotl : [lib] Made the PCCMETHOD rewriter ignore Parrot_PCCMETHOD calls, which are
21:59 svnbotl : functions that need no rewriting.
21:59 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=26300
22:32 Infinoid joined #parrot
22:33 allison joined #parrot
22:34 Infinoid joined #parrot
22:34 Limbic_Region joined #parrot
22:40 cognominal joined #parrot
22:52 purl joined #parrot
22:55 skids joined #parrot
22:57 purl joined #parrot
23:20 kid51 joined #parrot
23:45 svnbotl r26301 | chromatic++ | pdd17pmc:
23:45 svnbotl : [Tcl] Hoisted one string_from_literal() call out of the loop; this reduces line
23:45 svnbotl : length and makes the coding standards tests pass again.  (It also speeds up the
23:45 svnbotl : routine slightly.)
23:45 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=26301
23:47 Pabellon joined #parrot
23:58 rdice joined #parrot

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

Parrot | source cross referenced