Camelia, the Perl 6 bug

IRC log for #parrot, 2008-10-04

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:00 Robrt I have no clue if this is going to help anything or not.
00:00 particle me neither, but at least nobody can complain about the old version anymore :)
00:00 Robrt It wasn't that old.  It was stable!
00:01 particle the occasional '200 errors' were annoying, i wonder if this will fix them
00:01 Robrt I couldn't replicate them, and nobody was able to help me do so.
00:01 particle ci worked, but merge failed, so needed to do update manually after commit
00:01 particle it was sporadic and usually lasted a few hours
00:01 Robrt weird.
00:01 particle hard to pin down
00:02 Robrt I had this plan to have someone help by writing a script to do commits (to a different repo) every few minutes.
00:02 Robrt so we could see if there was a pattern.
00:04 particle1 joined #parrot
00:08 cotto Is any PMC that does hash or array iterable?
00:09 AndyA joined #parrot
00:16 particle1 sure
00:16 particle1 hash, for example
00:21 Theory joined #parrot
00:22 pmichaud some "hashes" such as LexPads aren't iterable
00:24 particle ok, so i just figured out (i think) that i was missing the :namespace() param for my PAST::Var to bind to
00:24 pmichaud yes, you'll want that.
00:25 particle not quite sure how to get it, though
00:26 pmichaud ...the namespace?
00:26 particle let's say i have: module Foo; sub bar is export {1};
00:26 particle how do i get bar's namespace in the routine_def action?
00:26 pmichaud what works now, or what should work?  ;-)
00:27 particle ah. no wonder it's hard.
00:27 pmichaud I _think_ that $?PACKAGE ought to have it.
00:27 particle what works now, dear sir.
00:27 pmichaud since every module is (I think) a package
00:27 pmichaud I don't think that it's been implemented that way yet, though.  :-(
00:28 particle nope, seems $?PACKAGE is blank in this case
00:28 particle it seems i may be shaving a yak soon
00:29 pmichaud might check with p6l to see if "module Foo;"  also sets $?PACKAGE
00:29 particle well, we haven't implemented package Foo...
00:30 pmichaud I'm not sure that's relevant
00:30 pmichaud I can envision that 'class Foo;'   will have the effect of setting $?PACKAGE, $?MODULE, and $?CLASS
00:30 particle As in Perl 5, a module is just a kind of package. Unlike in Perl 5, modules and classes are declared with separate keywords, but they're still just packages with extra behaviors.
00:31 particle so, i believe, yes, $?PACKAGE is set
00:31 particle *should be set
00:31 particle why the heck don't we do packages yet. i ought to fix that.
00:31 particle crap.
00:31 pmichaud sure, that would be the next step.
00:32 pmichaud (there's a reason that "is export" wasn't moving ahead quite so quickly :-)
00:33 particle i don't see any references to $?MODULE
00:33 pmichaud right.  I think jonathan jumped directly doing classes
00:34 particle ah, just in S02, once
00:34 particle S02.pod
00:34 particle 1946:    $?MODULE    Which module am I in?
00:34 pmichaud also $@MODULE
00:34 pmichaud er, @?MODULE
00:34 particle of course, they're a pair.
00:34 particle ok
00:35 particle do classes and roles set $?PACKAGE too?
00:36 pmichaud I've been thinking yes....
00:36 pmichaud but haven't confirmed it on p6l or anything
00:36 particle sure, but it's a direction in which to head
00:37 particle so class isa module isa package?
00:37 particle grammar isa module isa package?
00:37 pmichaud "A class is a module declared with the 'class' keyword."  (S12)
00:37 pmichaud I think a grammar isa class isa module isa package.
00:37 particle ah, yes
00:38 particle role isa module, too
00:38 particle but not a class
00:38 particle there's no switch in nqp, is there
00:39 pmichaud no.
00:40 dalek r31602 | julianalbo++ | trunk:
00:40 dalek : pirric: SAVE and LIST refactoring, and make more consistent use of quotes and apostrophes
00:40 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31602
00:49 dalek r31603 | julianalbo++ | trunk:
00:49 dalek : pirric: example of a pirric program that genrates and run a piric program
00:49 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31603
00:53 pmichaud afk for a while
00:53 cotto Sorry.  It's not good enough unless it has at least 3 levels of indirection, and preferably a configurable number.
01:04 petdance joined #parrot
01:11 PerlJam Holy cow! NYTProf is awesome!  We need to get Tim working on parrot.
01:15 dalek r31604 | cotto++ | trunk:
01:15 dalek : [pipp] implement clone, add, i_add and assign_pmc and enable associated tests
01:15 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31604
01:20 dalek r31605 | cotto++ | trunk:
01:20 dalek : [pipp] better exceptions for i_add and assign_pmc
01:20 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31605
01:24 cotto particle, ping
01:26 particle what's up, cotto?
01:26 cotto Is assign_pmc() in languages/pipp/src/pmc/phparray.pmc how HLL interop is supposed to work?
01:27 cotto latest revision
01:29 particle yes, mainly
01:29 particle i mean, it looks like the way it should
01:29 particle but the code could use some cleanup
01:29 Khisanth PerlJam: to do what?!
01:30 particle PMC_\w+_val should be turned to attribute access
01:30 particle but that's file-wide, not just in the vtable
01:30 particle and assign_pmc should be marked VTABLE
01:31 particle i like to point to src/pmc/exporter.pmc as an example of current-day pmc syntax
01:31 cotto That'll help.
01:32 particle oh, also, do you really want to test the exact type of the phparray pmc, or do you want to check its inheritance chain?
01:32 cotto It's good to know where to look for current code, rather than legacy stuff that only works because nobody's bothered to deprecate it.
01:32 cotto good point
01:32 particle i wrote exporter, so it is the pmc i know best
01:34 Hinrik hm, something in S03 caught me by surprise
01:35 Hinrik there's a 'constant' keyword? why not a 'constant' trait?
01:36 Theory joined #parrot
01:38 cotto particle++
01:38 particle glad to help
01:39 particle there's no 'my' trait
01:39 Hinrik yes but a constant is just a lexical that can't change
01:39 Hinrik why not be able to say "my $int is constant = 1" ?
01:40 particle constants can't be assigned to
01:40 Hinrik constant Dog $foo is woof = 123;  <-- then why does this work?
01:41 particle ah, right, just with parens you cant' do =
01:45 Hinrik apocalypse 2 mentions "my int $pi is constant = 3;"
01:45 Hinrik so is that still possible?
01:46 Hinrik ah, there's an update there
01:46 Hinrik Run-time properties are set with but rather than is. They're really just mixins. And there is no "constant" trait, but rather a "readonly" trait. Use the special declarator "constant int $pi = 3;"
01:46 Hinrik ...to declare a compile-time constant.
02:20 Robrt left #parrot
02:23 dalek r31606 | particle++ | trunk:
02:23 dalek : [rakudo] implement 'package' package declarator
02:23 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31606
02:29 Andy joined #parrot
02:54 dalek r31607 | chromatic++ | trunk:
02:54 dalek : [t] Appended PID to name of tempfile used in test to avoid race condition
02:54 dalek : during parallel testing.  (Guess who it just bit?)
02:54 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31607
03:23 dalek r31608 | tene++ | hllmagic:
03:23 dalek : [p6object]
03:23 dalek : * Code cleanliness
03:23 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31608
03:37 s1n does parrot have any requirement of python? for any version of python, to either build, execute, or otherwise have on a system?
03:39 cotto just for Parrot, no
03:39 cotto why
03:40 s1n i just noticed my gentoo install has a use flag for it
03:40 s1n not sure why it's checking for python, but they are checking specifically for python-2.4
03:41 s1n i was going to ask in here before i file a bug
03:41 s1n (on gentoo, not parrot)
03:41 cotto pynie, perhaps?
03:41 s1n i don't think they're building any of the languages yet
03:42 Hinrik a suggested ebuild on bugs.gentoo.org has a use flag for rakudo
03:42 s1n Hinrik: yes, i saw that too
03:42 s1n still a bug though, and nothing in the ebuild, 0.6.1 has anything that needs python that i can tell
03:42 Hinrik I wondered about that python dep, I just ended up copying the ebuild for the latest parrot version and removing the python dependency
03:43 s1n i'll file a bug and maybe by time parrot hits 1.0 the slow moving gentoo devs will get around to it
03:49 s1n Hinrik: what is up with those wacky ebuild requirements?
03:49 Hinrik you mean the python dependency? I don't know
03:49 s1n gpm, gdbm, ncurses, readline, icu? are any of those required?
03:50 s1n it has those as DEPENDS _and_ RDEPENDS!
03:50 Hinrik beats me :P
03:50 Hinrik especially the gpm one...
03:50 s1n heh, this doesn't seem right
03:51 s1n unicode isn't a requirement, is it?
03:51 cotto I know gdbm and icu are optional components
03:51 Hinrik it's like someone copied an ebuild from another package and forgot to change the deps or something
03:51 s1n sounds like it, ncurses is just plain wrong
03:51 s1n perl is the only DEPEND, right?
03:54 Hinrik according to the README, it's perl, icu and a C compiler
03:54 Hinrik https://svn.perl.org/parrot/trunk/README
03:54 Hinrik oh and make
03:54 s1n that's an "if" on icu
03:55 Hinrik hm, is the README wrong, then?
03:55 s1n no, the readme says "if you will be linking with the ICU library"
03:55 Hinrik You need a C compiler, a linker and a make program of course. If you will be linking with the ICU library you have to download and install it before configuring Parrot."
03:55 s1n note the "if" part
03:55 Hinrik oh, hah
03:55 Hinrik didn't see that
03:58 s1n does the parrot release url change depending on who's the current pumpking?
04:06 dalek r31609 | tene++ | hllmagic:
04:06 dalek : [p6object]
04:06 dalek : * Create namespaces for classes we're asked to create that don't have namespaces yet.
04:06 dalek : * This needs to be refactored into 'register'
04:06 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31609
04:09 s1n http://bugs.gentoo.org/show_bug.cgi?id=239492
04:09 s1n proper ebuild for parrot 0.7.1
04:10 s1n it might be addressed before the year is over, but don't hold your breath
04:27 Andy joined #parrot
05:43 cotto In PHP, arrays of same size with different keys are not considered comparable, and such a comparison returns null.
05:43 cotto Since the cmp VTABLE only returns an INTVAL, would it be a good idea to throw an exception?
05:44 cotto (from the PhpArray PMC's cmp function)
06:37 bacek joined #parrot
06:52 barney joined #parrot
07:03 bacek joined #parrot
07:13 dalek r31610 | tene++ | hllmagic:
07:13 dalek : [p6object]
07:13 dalek : * accept and use an hll named argument in new_class
07:13 dalek : * pass our hll on to 'register'
07:13 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31610
07:24 dalek r31611 | tene++ | hllmagic:
07:24 dalek : [p6object]
07:24 dalek : * Use the hll argument in 'register'
07:24 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31611
07:24 Tene I'm *still* surprised to see my hilights go off in here when dalek announces commits.  I keep thinking someone is commenting on whatever I committed.
07:26 mberends joined #parrot
07:29 bacek joined #parrot
07:32 * Tene sleeps, hopes to wake to instructions from pmichaud
07:51 clunker3 joined #parrot
08:00 iblechbot joined #parrot
08:12 dalek r31612 | fperrad++ | trunk:
08:12 dalek : [Markdown]
08:12 dalek : - implement Symbol
08:12 dalek : - and add test
08:12 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31612
09:01 iblechbot joined #parrot
09:02 johbar joined #parrot
09:06 AndyA joined #parrot
09:06 particle1 joined #parrot
09:16 dalek r31613 | cotto++ | trunk:
09:16 dalek : [pipp] implement cmp and is_equal, enabled several newly passing tests
09:16 dalek : general code cleanup, thanks to particle++ and chromatic++ for suggestions
09:16 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31613
09:21 rurban joined #parrot
09:39 dalek r31614 | allison++ | pdd27mmd:
09:39 dalek : [pdd27mmd] Revise test to match current mmd behavior.
09:39 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31614
09:40 masak joined #parrot
09:59 Ademan joined #parrot
10:55 dalek r31615 | rurban++ | trunk:
10:55 dalek : [ports] update to cygwin 0.7.1 release
10:55 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31615
11:20 rurban joined #parrot
11:23 ruoso joined #parrot
11:54 dalek r31616 | julianalbo++ | trunk:
11:54 dalek : pirric: use stdout as default in list
11:54 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31616
12:17 dalek r31617 | julianalbo++ | trunk:
12:17 dalek : use int instead of size_t in Parrot_print_backtrace, RT#59590
12:17 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31617
12:18 rurban joined #parrot
12:48 Whiteknight joined #parrot
12:49 tetragon joined #parrot
13:13 Andy joined #parrot
13:38 tetragon joined #parrot
13:45 masak NotFound: is there something I can do to help you along with the failing t/examples/library.t test?
13:46 NotFound masak: What OS tou have in the test failed system?
13:46 masak Mac OS X 10.5.4
13:46 tetragon joined #parrot
13:47 Bzek joined #parrot
13:47 NotFound Mmmm... I have no idea about how the linkers works on that.
13:48 masak neither do I. but I can fetch/nopaste any information you think you need.
13:49 dalek r31618 | Whiteknight++ | trunk:
13:49 dalek : [GC] Add function-level documentation to src/gc/resources.c
13:49 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31618
13:49 NotFound What prove -v t/library/pcre.t says ?
13:52 tetragon joined #parrot
13:52 NotFound Try also: pcre-config --prefix
14:18 tetragon joined #parrot
14:26 NotFound masak: I've found a possible pcre specific problem, unrelated to generic linking ones: there are place in the tools that use HAS_PCRE and other that use has_pcre
14:26 masak oki
14:27 Zaba_ pcre makes me sad
14:28 NotFound Unfortunately, ./parrot_config --dump shows that both HAS_xxx and has_xxx are frequently used.
14:31 dalek r31619 | Whiteknight++ | trunk:
14:31 dalek : [Charsets] Added some function-level documentation (for the functions I understand) to src/charset/ascii.c, and fixed some formatting of long lines.
14:31 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31619
14:31 pmichaud purl: msg Tene  I don't understand the need for r31609 -- Parrot already creates namespaces automatically.... why should we be creating them manually?
14:31 purl Message for tene stored.
14:32 pmichaud purl: msg Tene the next "big" task on this is to get PGE and PCT to compile with the new changes to P6object -- currently they're failing to build.
14:32 purl Message for tene stored.
14:57 dalek r31620 | julianalbo++ | trunk:
14:57 dalek : Fix has_pcre/HAS_PCRE inconsistent usages
14:57 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31620
15:01 NotFound masak: Can you try this fix? I don't think the problem was related, but just in case...
15:02 * masak tries
15:09 masak no, same error.
15:09 NotFound masak: try: ./parrot_config HAS_PCRE
15:09 masak at what stage? before make?
15:09 NotFound After
15:10 masak hokay.
15:10 masak it replied with a '1'
15:11 NotFound masak: try: prove -v t/library/pcre.t
15:11 masak all tests successful.
15:11 purl pass(); ok(1); is(1,1);
15:11 masak purl, forget all tests successful
15:11 purl masak: I forgot all tests successful
15:12 masak but that's nothing new. it's t/examples/library.t that gives the error.
15:12 NotFound Too much magic in those test, I think the solution is drop all attempts to check availability from the test and put them in the pit library, if needed.
15:13 NotFound s/pit/pir
15:14 masak sounds like a good idea
15:18 NotFound masak: please execute: pcre-config --prefix
15:18 masak NotFound: /opt/local
15:18 purl /opt/local is evil
15:18 masak purl: forget /opt/local
15:18 purl masak: I forgot /opt/local
15:18 masak purl: you're the evil one.
15:18 purl masak: excuse me?
15:21 NotFound I think I have a diagnose: t/library/pcre.t looks for the prefix, and adds prefix/lib to the parrot library search path (what BTW is an undocumented feature), but examples/library/pce.pir does not do that thing.
15:23 NotFound So if your pcre lib is not in a standard lib location, such as /opt/local.... is evil, purl was right ;)
15:25 NotFound masak: Now the problem is to think about what the correct solution must be.
15:26 masak NotFound: aye. I understand what you're saying, but am not in a position to suggest anything constructive.
15:30 NotFound I think that letting any program or pir library to manipulate the library search path, and even forcing to do it, is not a solution.
15:30 masak right.
15:33 NotFound I'll try a temporary fix: if the library is not found, look at the process image.
15:34 tetragon joined #parrot
15:44 dalek r31621 | Whiteknight++ | trunk:
15:44 dalek : [Charsets] Added some function-level documentation (for the functions I understand) to src/charset/binary.c, and fixed some formatting of long lines.
15:44 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31621
15:49 NotFound Why library/pcre.pir needs to load_bytecode 'library/pcre.pir' ?
15:54 jhorwitz joined #parrot
15:56 dalek r31622 | pmichaud++ | trunk:
15:56 dalek : [rakudo]:  spectest-progress.csv update: 200 files, 4437 passing tests
15:56 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31622
16:14 sjansen joined #parrot
16:23 NotFound nopaste?
16:23 clunker3 http://pasta.test-smoke.org/ or http://paste.husk.org/ or http://nopaste.snit.ch:8001/ or http://rafb.net/paste or http://poundperl.pastebin.com/ or http://paste.scsys.co.uk/
16:23 purl rumour has it nopaste is at http://nopaste.snit.ch/ (ask TonyC for new channels) or http://rafb.net/paste or http://poundperl.pastebin.com/ or http://paste.scsys.co.uk/ or App::Nopaste or tools/dev/nopaste.pl or at http://www.extpaste.com/
16:24 nopaste "NotFound" at 213.96.228.50 pasted "Temporary fix for pcre tests" (20 lines) at http://nopaste.snit.ch/14224
16:24 NotFound masak: can you try this patch?
16:29 masak NotFound: hokay
16:32 masak NotFound: problem persists.
16:32 masak same error.
16:34 NotFound Same output than in ticket?
16:34 masak aye
16:36 masak I'm heading home in a few minutes, but will be back later, and will backlog.
16:37 NotFound Can you try this?; ./parrot examples/library/pcre.pir Hello h
16:41 masak sure
16:41 masak says "Bus error"
16:41 masak what did you do to my bus? :)
16:42 masak & # home
16:47 NotFound Mmm... then the problem is harder, someone with better knowledge of Mac OS than me must take a look.
16:57 jan joined #parrot
17:10 Tene pmichaud: the issue was that parrot wasn't creating namespaces automatically.  get_root_namespace was returning null.
17:12 nopaste "tene" at 166.70.38.237 pasted "demo of get_root_namespace not creating a namespace" (16 lines) at http://nopaste.snit.ch/14225
17:22 dalek r31623 | julianalbo++ | trunk:
17:22 dalek : pirric: add CLEAR instruction and his support, give better diagnostic for undefined variables
17:22 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31623
17:42 dalek r31624 | julianalbo++ | trunk:
17:42 dalek : pirric: add ERROR instruction
17:42 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31624
18:07 pmichaud Tene: okay, why do we need a loop for creating the namespace?
18:08 Tene pmichaud: how would you do it?
18:08 pmichaud just create it.
18:09 Tene Using the 'create_namespace' opcode, eh?
18:09 pmichaud make_root_namespace
18:10 pmichaud although now that I think about it, that might be unimpl even though it's defined in pdd21
18:10 Tene Yeah, I don't see that anywhere.
18:10 pmichaud we could always "cheat" and use the set_root_global opcode for now, though.  :-)
18:12 pmichaud it just feels funny that we have to create all of the intermediate namespaces.
18:12 Tene should get_root_namespace be creating the namespace if it doesn't exist?
18:12 pmichaud I think the other get_*_namespace opcodes do it.
18:12 Tene oh?  Lemme test.
18:12 pmichaud no, they don't.
18:13 pmichaud okay, so I had that wrong in my head.
18:13 Tene I can simplify that loop to elimite the repeated 'get_root_namespace calls.
18:13 pmichaud set_*_global will create the namespace if needed, though.
18:14 pmichaud so, given a namespace of   'hll';'Foo';'Bar';'Baz'   we could create it quickly with    set_root_global ['hll';'Foo';'Bar'], 'Baz', ns
18:14 pmichaud where ns is a newly created namespace
18:15 pmichaud even better might be a 'make_namespace' method
18:15 pmichaud instead of a 'make_namespace' opcode.
18:15 pmichaud $P0 = get_root_namespace ['hll']
18:15 pmichaud $P0.'make_namespace'($P1)    # $P1 is array of names
18:16 pmichaud _another_ thought I had is that it would be really nice if a sub could change its hll alignment in midstream.
18:16 pmichaud i.e., upon entering one of these subs, it could change its current context's affiliation to the hll of the caller
18:17 pmichaud I could see how that would be problematic, though.
18:17 pmichaud anyway, I'd like to see us avoid the loop if at all possible -- that seems like something Parrot should be doing.
18:18 pmichaud for one, we have to be careful that we don't accidentally clobber an existing symbol (i.e., a protoobject) when creating the namespace.
18:20 * Tene nods.
18:22 pmichaud there's a ticket that references make_namespace:  RT #46859
18:26 pmichaud anyway, the big task at the moment is getting the existing stuff that uses the 'Foo::Bar' namespace names to switch to 'Foo';'Bar'
18:26 Tene Starting with PGE and PCT, right?
18:26 pmichaud PGE and PCT are the biggies, yes.  I think the others are mostly already set correctly.
18:26 pmichaud i.e., I don't think NQP has a problem with it.
18:27 Tene does NQP use pge or pct?
18:27 pmichaud yes, but it does so in the standard protoobject way.
18:27 pmichaud i.e., it doesn't ever make use of the 'PAST::Var' form
18:28 pmichaud it always does things based on the ['PAST'], 'Var'   protoobject, which isn't going to change here.  :-)
18:28 Tene nqp doesn't run in the branch right now
18:28 pmichaud well, nqp can't run until pge and pct run, no.
18:28 Tene Right, okay.
18:28 pmichaud I'm just saying it probably doesn't have to change once those are fixed.
18:28 * Tene nods.
18:29 pmichaud I need to get some lunch, bbiaw
18:29 pmichaud (we can leave the ns creation loop in p6object for now, but that's something I'd really like to get cleaned up at some point.)
18:30 Tene perhaps make a ticket for it.
18:30 pmichaud that would work.
18:30 pmichaud at some point we'll also have to figure out how to handle hll-specific namespace object types
18:31 pmichaud anyway, gone to lunch, bbiah
18:31 Tene seeya
19:03 NotFound joined #parrot
19:04 NotFound perl6.pir has a .namespace [ 'Perl6::Compiler' ] This is not what you are trying to avoid?
19:05 Tene Yes, that will need to be 'Perl6';'Compiler'
19:06 NotFound And there are also several others :: in that file
19:29 NotFound In perl6.pir there is a push_eh _handler in line 71, and this handler seems no never be popped. Must no be a pop_eh before the label _handler: on line 87?
19:32 Theory joined #parrot
19:36 pmichaud since there's not really anything that might cause an exception, there doesn't really need to be a pop_eh
19:36 pmichaud yes, it would probably be cleaner if one is there.
19:37 NotFound pmichaud: I caused one, doing a simple test of changing :: to ;
19:38 pmichaud in trunk, or hllmagic branch?
19:38 NotFound trunk
19:38 pmichaud right, changing the :: to ; won't work in trunk yet, because P6object doesn't split names yet.
19:38 pmichaud (in trunk)
19:39 NotFound I imagined that, but the point is that the lack of pop_eh can give misguided diagnostics.
19:39 pmichaud sure.
19:39 pmichaud feel free to add one.
19:39 NotFound Ok.
19:40 pmichaud but I expect most of the :: to ; conversions to take place in the hllmagic branch, at least until we get it merged back to trunk.
19:42 dalek r31625 | julianalbo++ | trunk:
19:42 dalek : add a missing pop_eh in perl6 initialization
19:42 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31625
20:07 Theory joined #parrot
20:19 Zaba joined #parrot
21:03 ruoso joined #parrot
21:17 Tene particle1: why do you need ' and - for 'is export'?
21:22 dalek r31626 | julianalbo++ | trunk:
21:22 dalek : pirric: add ON ERROR GOTO and ON ERROR EXIT instructions and some interface methods to Runner, and update examples
21:22 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31626
21:25 dalek r31627 | chromatic++ | trunk:
21:25 dalek : [IMCC] Changed SymHash entry information to use unsigned ints.  C doesn't give
21:25 dalek : much type protection, but it can give compilation warnings when we do things
21:25 dalek : which might walk outside of an array.  We should take advantage of this when
21:25 dalek : possible.
21:25 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31627
21:33 ruoso joined #parrot
21:40 dalek r31628 | chromatic++ | trunk:
21:40 dalek : [IMCC] Made Basic_block index member an unsigned int, as a negative index would
21:40 dalek : be a memory violation.
21:40 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31628
21:54 Tene pmichaud: what should I do with "new 'PGE::OPTable'"?  call new on a RSA, get the protoobject out of the namespace and call .new() on that, edit 'new' to do splitting on '::'?
21:55 dalek r31629 | chromatic++ | trunk:
21:55 dalek : [distro] Updated NEWS per some of my more interesting commits.
21:55 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31629
21:55 Andy joined #parrot
22:01 dalek r31630 | julianalbo++ | trunk:
22:01 dalek : [distro] NEWS update
22:01 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31630
22:01 dalek r31631 | chromatic++ | trunk:
22:01 dalek : [IMCC] Converted indexes in Instruction struct to unsigned integers, again to
22:01 dalek : let the compiler warn about invalid indexes.
22:01 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31631
22:05 gmansi joined #parrot
22:08 Andy joined #parrot
22:14 omega joined #parrot
22:14 dalek r31632 | chromatic++ | trunk:
22:14 dalek : [IMCC] Fixed signedness of all values passed to system calls such as malloc()
22:14 dalek : and realloc().
22:14 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31632
22:14 TiMBuS joined #parrot
22:15 petdance joined #parrot
22:21 Tene Hmm.  I'm having trouble with 'subclass'.  Let's see what other things I can try...
22:28 tetragon joined #parrot
22:50 Ademan_ joined #parrot
22:51 dalek r31633 | chromatic++ | trunk:
22:51 dalek : [IMCC] Made the number of symbols in an IMC_Unit an unsigned quantity to avoid
22:51 dalek : signedness errors.
22:51 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31633
23:01 pmichaud Tene:  it should be     $P0 = get_hll_global ['PGE'], 'OPTable';   $P1 = $P0.'new'()
23:01 pmichaud that's the standard mechanism for all P6object-based things.
23:02 pmichaud you may run into some issues in various parts of PGE with that -- if so, let me know and I'll look at them pretty quickly.
23:10 dalek r31634 | allison++ | pdd27mmd:
23:10 dalek : [pdd27mmd] Change 'floor_divide' so it doesn't immediately morph Integer result
23:10 dalek : to a Float. Change 'modulus' to a multi, so it doesn't mask default multi
23:10 dalek : dispatch.
23:10 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31634
23:10 dalek r31635 | chromatic++ | trunk:
23:10 dalek : [src] Fixed signedness of several string functions (avoiding overflow and
23:10 dalek : negatives in string sizes seems sane).
23:10 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31635
23:10 dalek r31636 | allison++ | pdd27mmd:
23:10 dalek : [pdd27mmd] Change direct multi calls to use correct vtable names of the multis
23:10 dalek : for modulus.
23:10 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31636
23:14 dalek r31637 | Whiteknight++ | trunk:
23:14 dalek : [Book] Add some more information about PMCs.
23:14 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31637
23:14 s1n pmichaud: ping
23:16 pmichaud s1n: pong
23:16 s1n sweet, you're around
23:16 s1n pm
23:18 s1n i've been daydreaming about working on parrot/rakudo at my day job :)
23:26 Andy joined #parrot
23:33 masak s1n: who hasn't? :)
23:33 s1n masak: unfortunately i can't do anything other than browse the svn sources (no cheating at work and working on it on the clock)
23:35 masak s1n: I'm a bit more unrestricted than that, but there have definitely been times when I wished myself away from $work to hack on Rakudo instead
23:35 Whiteknight i can do some writing/documenting stuff, but I dont have a compiler on my work computer so I can't do much coding
23:36 Whiteknight but I definitely would much rather be working on Parrot then working on work
23:36 s1n my linux machine at work isn't externally connected and the windows (email) machine is fairly useless
23:46 Ademan joined #parrot
23:56 davidfetter joined #parrot
23:59 ruoso joined #parrot

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

Parrot | source cross referenced