#parrot Parrot 0.6.4 "St. Vincent Amazon" Released | parrotcode.org/ | 15 new, 648 open, 80 stalled tix | logged in irclog.perlgeek.de/parrot/today
Set by moderator on 28 July 2008.
00:10 AndyA joined
cotto_work chromatic, thanks for the tip in #ps. 00:50
00:50 bacek joined
cotto_work I'm reasonably sure my code works now. 00:50
(all tests pass, but that only means that all tests pass)
Infinoid ... which is a really good start 00:51
00:58 petdance joined 01:11 Whiteknight joined 02:02 Ademan joined
tewk gsoc_nci builds and passes t/pmc/nci.t with strawberry perl, 02:34
going to build with active state and msvc 02:35
chromatic Sweet.
JITted thunks?
tewk yep, particle reported that it fail on his windows machine with msvc. so I'm testing that now 02:37
I didn't realize that configure with strawberry perl would default to using msys, but it make sense. 02:39
02:40 SCalimlim joined
SCalimlim has anyone built parrot on freebsd lately? 02:40
tewk lots of warnings with msvc 02:41
dalek r30045 | jkeenan++ | scriptconfigure: 02:51
: [configure] Extract those options which are valid either on the command-line
: or inside a configuration file and place them in
: @Parrot::Configure::Options::Conf::Shared::valid_shared_options. Import this
: into Parrot::Configure::Options::Conf::CLI and
: Parrot::Configure::Options::Conf::File. Provide for a more robust
: configuration file. Still need to allow for multiple steps designated as
: verbose-step or fatal-step.
diff: www.parrotvm.org/svn/parrot/revision?rev=30045
r30046 | chromatic++ | trunk: 03:01
: [config] Added a dependency on key.str to src/key.c target.
diff: www.parrotvm.org/svn/parrot/revision?rev=30046
tewk hmmmm, I got two failing t\\pmc\\nci.t tests using msvc, which I fixed, just need to push to svn, but it built for me. 03:12
Maybe I got lucky, nci jitted stub errors canC[C[C[C[C cause bad things to happen 03:13
dalek r30047 | chromatic++ | trunk: 03:14
: [src] Fixed a segfault during global destruction of runloop jump points in
: optimized builds; the runloop pointers passed to
: really_destroy_runloop_jump_points may be NULL, so the compiler has to know
: about that so as not to optimize away the checks.
diff: www.parrotvm.org/svn/parrot/revision?rev=30047
03:19 Whiteknight joined
dalek r30048 | chromatic++ | trunk: 03:58
: Added the ports/cygwin directory, with files sufficient to make a Cygwin port
: of Parrot (Reini Urban, RT #57608).
diff: www.parrotvm.org/svn/parrot/revision?rev=30048
r30049 | chromatic++ | trunk: 04:01
: [config] Added support for XEmacs 21-style ctags (Reini Urban, RT #57546).
diff: www.parrotvm.org/svn/parrot/revision?rev=30049
r30050 | chromatic++ | trunk: 04:04
: [t] Made stat/lstat tests more portable on Cygwin (Donald Hunter, RT #57486).
diff: www.parrotvm.org/svn/parrot/revision?rev=30050
04:36 tetragon joined 05:03 teknomunk_ joined
cotto_home clkao? 05:13
purl rumour has it clkao is talking about the higher level thing
cotto_home seen clkao
purl clkao was last seen on #moose 1 days, 4 hours, 5 minutes and 56 seconds ago, saying: failed [Aug 5 01:07:30 2008]
05:19 Ademan joined 05:39 Psyche^ joined 06:15 uniejo joined 06:29 teknomunk joined
dalek r30051 | fperrad++ | trunk: 07:32
: [Lua] PMC
: - refactor LuaUserdata with ATTR
diff: www.parrotvm.org/svn/parrot/revision?rev=30051
07:38 barney joined 08:05 clunker3 joined 08:14 UltraDM joined 08:21 bacek_ joined 08:29 cognominal joined 08:57 masak joined 10:12 Ademan joined
dalek r30052 | julianalbo++ | trunk: 10:33
: drop directory ports/cygwin from MANIFEST
diff: www.parrotvm.org/svn/parrot/revision?rev=30052
r30053 | kjs++ | trunk: 10:34
: [pirc/new] refactor assignment rule
: + start of constant-folding implementation in grammar.
: + refactoring of several rules
: + more work on back-end (data structures)
diff: www.parrotvm.org/svn/parrot/revision?rev=30053
11:21 gaz joined
dalek r30054 | fperrad++ | trunk: 11:39
: [Lua]
: - refactor namespace : extension libraries
diff: www.parrotvm.org/svn/parrot/revision?rev=30054
11:52 clunker3 joined 11:58 kj joined 11:59 clunker3 joined
dalek r30055 | fperrad++ | trunk: 12:10
: [Lua]
: - refactor namespace : extension libraries (step 2)
diff: www.parrotvm.org/svn/parrot/revision?rev=30055
12:41 Whiteknight joined 13:01 teknomunk_ joined
dalek r30056 | fperrad++ | trunk: 13:12
: [Lua]
: - refactor namespace : standard libraries
diff: www.parrotvm.org/svn/parrot/revision?rev=30056
13:12 petdance joined
dalek r30057 | julianalbo++ | trunk: 13:19
: added runloop for debugger and some more debugger changes
diff: www.parrotvm.org/svn/parrot/revision?rev=30057
Whiteknight julianalbo++ 13:21
13:32 ambs joined
dalek r30058 | fperrad++ | trunk: 13:45
: [Lua]
: - refactor namespace : Perl Compiler
diff: www.parrotvm.org/svn/parrot/revision?rev=30058
r30059 | fperrad++ | trunk:
: [Lua]
: - refactor namespace : extension libraries (step 3)
diff: www.parrotvm.org/svn/parrot/revision?rev=30059
14:00 gryphon__ joined 14:33 petdance joined 14:35 clunker3 joined
dalek r30060 | jonathan++ | trunk: 14:52
: [rakudo] Stub in empty Perl6MultiSub PMC deriving from MultiSub, add it to the Makefile and add a sanity test to make sure we can instantiate it.
diff: www.parrotvm.org/svn/parrot/revision?rev=30060
particle points jonathan at the MANIFEST 14:54
jonathan oops 14:56
MANIFEST MUST DIE
:-)
dalek r30061 | jonathan++ | trunk: 14:57
: [rakudo] Add perl6multisub.pmc to the MANIFEST; particle++ for noticing.
diff: www.parrotvm.org/svn/parrot/revision?rev=30061
r30062 | infinoid++ | trunk: 15:11
: [src] Remove trailing whitespace in src/debug.c to pass codingstd tests.
diff: www.parrotvm.org/svn/parrot/revision?rev=30062
r30063 | infinoid++ | pdd13pbc: 15:14
: [pdd13] Merge changes from trunk (-r29903:30062) into pdd13pbc branch.
diff: www.parrotvm.org/svn/parrot/revision?rev=30063
15:29 Theory joined 16:15 Ademan joined 16:17 iblechbot joined 16:48 mj41 joined 16:49 peepsalot joined 16:59 davidfetter joined
cotto_work t/codingstd/pdd_format.t is griping 17:17
Infinoid seen rurban 17:23
purl rurban was last seen on #parrot 3 days, 6 minutes and 43 seconds ago, saying: I'll apply for 10 hours � 100 euro in october then :) [Aug 3 17:16:43 2008]
tewk particle: msvc build for me, I had a couple of t/pmc/nci.t failures, but it built and ran. 17:27
particle which version msvc?
tewk latest patch against head tewk.com/nci.diff
2008
particle cl -h returns 15.00... for me 17:28
ok, that's same 17:29
tewk 15.00... yeah
particle can you post a diff against gsoc_nci?
or just ci 17:30
i see you noticed the zillion warnings earlier
tewk Its been some time since I've compiled parrot on win32
let me see if I can post a diff agaisnt gsoc_nci 17:31
ci i mean
particle meantime i'll rebuild parrot trunk head and check for warnings
jonathan hi folks 17:40
Think I will hack on Rakudo tomorrow.
pmichaud Aloha
jonathan (As in, have Rakudo day. Today's hacking is MMD.)
Ah, hi PM.
:-)
How's vacation?
dalek r30064 | jonathan++ | trunk:
: [rakudo] Start to stub in some bits of the Perl6MultiSub PMC, plus add a few more sanity tests. Nothing to see or play with just yet, and what is here is very preliminary code and, in places, untested.
diff: www.parrotvm.org/svn/parrot/revision?rev=30064
particle was it dark in the lava tube? 17:41
pmichaud it's going good. Did Volcanoes yesterday. dark in the lava tube, yes
got to see where the lava is flowing into the sea
jonathan Wow, there's actually lava flows to see there?!
pmichaud yes
jonathan WANT!
particle yep, it's the only u.s. state that's growing 17:42
jonathan should make a vacation there sometime
17:42 jcm joined
pmichaud we didn't see the flow itself, but we could see the lava shooting into the air when it reacted with the seawater 17:42
I have a video :-)
jonathan Cool.
17:42 ambs joined
jonathan I booked my flight for Vladivostok last night. 17:42
pmichaud let's see if I have a pic
pmichaud fires up Picasa 17:43
no, pic. Let's see how long the video would be 17:44
108M . probably too long to upload via my gprs modem :-) 17:45
jonathan :-)
Well, bring it along the YAPC::EU. ;-)
pmichaud (I can cut it down a fair bit when I get to some reasonable video editing tools :-) 17:46
jonathan Will Rakudo-hack tomorrow because on Friday late afternoon I'm flying.
pmichaud sounds good.
I'll be hacking on plane tomorrow night, most likely
jonathan Will likely spend some time making sure, that my examples from my talks actually run!
Then I was considering making a branch and doing some lazy list stuff in it
particle remind me, does rakudo have given/when yet? 17:47
jonathan So that we can hit YAPC::EU with some code to review and that exposes some of the issues (like, with ResizalbePMCArrays)
particle: It 'has' them, but when doesn't throw the control exception at the end of a block.
pmichaud rakudo has given/when, but the when's .... what jonathan said
particle ok, i see some room for more tests against S04
jonathan particle: Is waiting on PCT changes, which were waiting on pdd25cx, but are unblocked now.
particle like, what happens with default {...} when * {...} 17:48
jonathan Those two are about equivalent, IIRC.
pmichaud (I think they're explicitly equivalent.)
particle they are exactly equivalent
jonathan OK. :-)
particle but, if they're repeated...
pmichaud then the second one never gets executed.
jonathan You'd never reach the second one, usually...
particle do you get a warning?
jonathan Unless you used "continue", maybe.
pmichaud We might provide warnings if there are unreachable statements, yes. 17:49
particle default { continue } when * { foo() }
jonathan Yeah. :-)
particle sick.
that needs to be a test
jonathan :-)
pmichaud that we provide a warning? I don't think the spec requires that we issue a warning.
particle no, it doesn't. 17:50
jonathan If the spec doesn't need it, shouldn't be a spectest. But we should make sure default { continue } when * { foo() } does the right thing.
particle the spec doesn't define behavior with multiple default blocks
it's ambiguous
jonathan I'm not sure it needs to.
particle it's referred to as "the default case" 17:51
not "a default case"
jonathan Ah.
pmichaud This is one of those places wehre I think the language spec doesn't need to require a warning, but implementations may be encouraged to issue one.
particle anyway, that's #perl6-speak
jonathan pmichaud: Does my Rakudo day plan sound reasonable?
pmichaud jonathan: yes
jonathan OK, great. 17:52
particle isn't that rakuday-do?
.oO(the day of the camel)
pmichaud I'm looking forward to being able to dedicate some tuits on rakudo again
jonathan needs to try and get most of his second talk written tonight too
I'm looking forward to you being able to dedicated some tuits on Rakudo again ;-)
Well, day either side of YAPC::EU and time during the conference itself, should be productive. 17:53
pmichaud yes, I'm expecting so.
jonathan I have a few days just relaxing and vacation before it too.
particle pmichaud: poke rdice about the hague grant proposal review, if he hasn't committed to a date yet
jonathan So I should arrive with some energy.
pmichaud particle: I think rdice is waiting for me to finish the mofo grant 17:54
which I told him I'd do before hawai'i, but obviously I didn't make it
so that's my friday afternoon task
particle ah, right.
ser gut. 17:55
jonathan needs food 17:57
Probably need to do some non-Rakudo tasks later today too
But may get more MMD time in. 17:59
Whiteknight your mmd algorithm sounds pretty interesting jonathan
jonathan needs to spend 40 hours on it in total before August is out. And is on vacation for the last week of August, and at YAPC next week!
Whiteknight: Well, I'm hoping it's @Larry's MMD algorithm. ;-)
But yes, it's an interesting one.
Whiteknight if you implement it before Larry gets a chance to comment on it, we will just be stuck with it 18:00
jonathan Well, it's spec'd.
So I'm trying to implement it the way the spec suggests.
I'm hoping to hack a little bit on evenings while I'm on vacation though.
So maybe during YAPC::EU I have some progress to show and get @Larry to play with. 18:01
Whiteknight how is this getting implemented, in PIR, NQP, or something better? 18:04
jonathan Whiteknight: The actual dispatcher itself is C.
pmichaud MMD probably has to be in C, for efficiency reasons
Whiteknight oh, okay. makes sense
jonathan The parsing and code-gen of everything up to the dispatcher is in NQP. 18:05
Or PIR. Or Perl6 Regex, for the parsing bits. 18:06
Signature objects will be constructed through NQP too, though the exact nature of those is for discussion at YAPC::EU, I expect.
For now I'm writing PIR tests.
And will fake up something with the right interface.
pmichaud time for me to go do more vacationing. 18:07
I'll be glad when vacationi s over so I can get some rest :-)
jonathan pmichaud: enjoy it :-)
A good vacation leaves you tired of vacationing and longing for work again. ;-)
dalek r30065 | coke++ | trunk:
: Don't complain about files in ports/ - they may need to break the rules.
diff: www.parrotvm.org/svn/parrot/revision?rev=30065
pmichaud afk 18:09
tewk gsoc_nci just needs to be deleted and recreated, maybe tonight. 18:11
18:23 ruoso joined 18:45 ambs joined 19:16 ryanc joined
dalek r30066 | fperrad++ | trunk: 19:21
: [Lua]
: - refactor namespace : extension libraries (step 4)
diff: www.parrotvm.org/svn/parrot/revision?rev=30066
19:36 donaldh joined
TimToady phone? 20:00
purl hmmm... phone is 1/4".
TimToady phono?
purl, phono is RCA. 20:01
purl OK, TimToady.
donaldh rakudo trunk fails to build on Cygwin 20:03
It doesn't look specific to Cygwin. Has anyone else noticed? 20:04
./perl6multisub.pmc:102: error: `PARROT_OP_set_args_pc' undeclared (first use in this function)
particle do you have an installed rakudo on cygwin? 20:09
er, installed parrot
donaldh nope
jonathan Ugh. 20:11
It builds on Windows.
And includes parrot.h
Maybe that's not enough, though. 20:12
particle builds with msvc, yes. just checked
donaldh It's defined in parrot/oplib/ops/h
s?ops/h?ops.h? 20:13
jonathan OK, try adding a #include to that.
donaldh Just wondering why it works elsewhere. 20:14
jonathan Not sure why, this would work one way on MSVC and not work on gcc or something, though.
Yeah, I've no idea.
cotto_work donaldh, same on linux/x86/gcc 20:18
b0rken
Tene Yeah, I get that failure on linux/x86_64/gcc 20:20
cotto_work is anyone other than kl working on pirc? 20:23
s/kl/kj/
nopaste "donaldh" at 213.123.171.12 pasted "patch for ./perl6multisub.pmc:102: error: `PARROT_OP_set_args_pc' undeclared (first use in this function)" (23 lines) at nopaste.snit.ch/13751
particle cotto_work: not that i'm aware of. svn log will help 20:24
cotto_work particle++ #of course
jonathan Anyone with commit bit and tuits now, feel free to apply that, otherwise I do it later - tied up with something at the mo. Thanks, donaldh++ 20:34
particle that patch seems so *wrong*, but who can argue if it works? 20:35
does #include "parrot/embed.h" work? 20:36
donaldh I can try it.
jonathan particle: It's OK, there's much wronger things in that file. ;-) 20:37
donaldh It doesn't fix it anyway. 20:38
parrot/embed.h that is.
particle ok, i'll try your patch with msvc 20:39
if it builds, i'll commit
jonathan thanks, particle++
Still bemused why it built in the first place on MSVC, if not on gcc...
donaldh I'm trying to track down any conditional code. 20:40
particle are you looking in the generated .c file? 20:42
.\\perl6multisub.pmc(54) : fatal error C1083: Cannot open include file 'Parrot/oplibs/ops.h': No such file or directory 20:43
ah, oops
i put a trailing 's' on oplib
donaldh Okay, the other PMCs that test PARROT_OP_* also #include "parrot/oplib/ops.h" so it's _normal_ 20:45
particle passes make test. committing. 20:48
nothing about parrot's opcode mmd is _normal_
donaldh :D
dalek r30067 | particle++ | trunk: 20:49
: [rakudo] include proper header for non-msvc builds (dunno why this works, it just does) donaldh++
diff: www.parrotvm.org/svn/parrot/revision?rev=30067
20:50 petdance joined 20:51 Whiteknight joined
dalek r30068 | fperrad++ | trunk: 21:05
: [Lua]
: - refactor namespace : extension libraries (step 5 : export_to)
diff: www.parrotvm.org/svn/parrot/revision?rev=30068
21:29 Ademan joined 21:36 vhold joined
dalek r30069 | Whiteknight++ | trunk: 21:42
: [core] Small spelling fix in src/pmc.c
diff: www.parrotvm.org/svn/parrot/revision?rev=30069
Tene Huh. Now I'm getting failure here: 21:54
../../parrot -o perl6.pbc perl6.pir
remove_attribute() not implemented in class 'ResizableStringArray'
21:56 slightlyoff joined
dalek r30070 | fperrad++ | trunk: 22:01
: [Lua]
: - refactor namespace with export_to : standard libraries
diff: www.parrotvm.org/svn/parrot/revision?rev=30070
22:04 slightlyoff left
dalek r30071 | Whiteknight++ | gsoc_pdd09: 22:08
: [gsoc_pdd09] update to trunk r30070
diff: www.parrotvm.org/svn/parrot/revision?rev=30071
22:10 timbunce joined
dalek r30072 | Whiteknight++ | trunk: 22:37
: [STM] add some function-level documentation to src/stm/backend.c
: * Open a few CAGE tickets corresponding to some TODO, XXX and FIXME items, and add RT# tags where appropriate.
diff: www.parrotvm.org/svn/parrot/revision?rev=30072
cotto_work more fun 22:43
nopaste "cotto" at 131.107.0.77 pasted "another disturbingly short segfault" (13 lines) at nopaste.snit.ch/13752 22:44
cotto_work Parrot might not quite be ready for prime time just yet. 22:45
jonathan cotto_work: That is just another case of the "fun" one I discovered
But yes, we need to do something about STRINGs. 22:46
cotto_work your earlier nopaste made me think of trying that instead of debugging my code 22:48
22:49 Limbic_Region joined
cotto_work I'm curious why the i = s line is necessary to trigger the sf, though 22:49
jonathan Because it doesn't check if s is NULL before trying to parse it and get an int. 22:52
cotto_work I see. The i = s line is segfaulting, not the say. 22:55
Whiteknight sounds like there should be an easy fix. I don't know enough about it, however
do strings, ints, and nums have vtable methods? 22:56
or are vtables reserved for pmcs?
Infinoid it segfaults at string_to_int() in src/string.c line 1942 22:57
jonathan Just PMCs. 22:58
Infinoid it segfaults because "s" is undef, and therefore, the STRING* pointer on the C level is NULL.
jonathan But I think we need to not represent null strings as a C NULL.
It's kinda just asking for segfaults.
Whiteknight I see that, and it's unhelpful that PARROT_ASSERT(s) isn't called until after s->strstart is dereferenced 22:59
Infinoid well, you can either pepper the code with checks for NULL, or make a STRINGNULL and pepper the code with checks for that... I don't really see much difference between the two.
Whiteknight I'll take a quick stab at a solution for that issue. It should be easy enough if we don't worry about beautiful code 23:00
Infinoid Whiteknight++
Whiteknight ...done! testing now
of course, it's the testing that always takes such a long time 23:01
Infinoid seems the pdd13pbc branch no longer compiles because real_exception() no longer exists... I'll fix that later tonight.
jonathan Infinoid: If the STRINGNULL is something that throws errors, like PMCNULL does, it works out better if you forget to check it. 23:02
NotFound jonathan: the price to pay can be high, a check in almost any string function. 23:04
dalek r30073 | Whiteknight++ | trunk: 23:05
: [CORE] fix to src/string.c:string_to_int(). If a NULL string is passed, the code would segfault on a null pointer dereference. Now, if a NULL string is passed, it returns 0. Not the prettiest fix ever.
diff: www.parrotvm.org/svn/parrot/revision?rev=30073
jonathan Whiteknight: Would it not have been better to throw an exception there?
Whiteknight it might have been.
jonathan NotFound: Yes, I fear that the way strings are (you don't intereact with the via a v-table) may force that. 23:06
Whiteknight it makes the most sense to me that it returns 0, but the "best response" is a personal preference
jonathan I hadn't been fro a good look.
Whiteknight I am forcing a NULL string and an empty string to be the same concept in my mind, for simplicity
cotto_work Whiteknight, what are the extra curly braces for? 23:10
NotFound Whiteknight: the function signature must be changed to no mark it as non null. 23:11
Whiteknight cotto_work: C89 specifies that in a block, all variable declarations must come before all code
since I can't add an if statement before the variable declarations in the same block, I created a different block
basically, it tells the compiler to shut up 23:12
NotFound, you're right. I'll do that in a second
NotFound Whiteknight: I put a comment in that cases, something like /* Limit scope of variable ... */ 23:13
Is helpful for people not familiar with that idiom. 23:14
Whiteknight NotFound, the function allows NULL now
NotFound Whiteknight: in string_func.h there is an __attribute_nonnull 23:15
Lack of headerizer?
Whiteknight No, you're right. I need to change ARGIN to ARGIN_NULLOK
unfortunately, that means "make" has to redo everything since I changed the headerfiles 23:16
:(
shoot. Can somebody else "make headerizer && make"? 23:17
When I did that, it screwed up some definitions in the IMCC headers, and now it wont build
NotFound Maybe we have too much dependencies, including unneeded headers in several places. 23:18
Whiteknight: let it to me, I will change the function, headerize, and revert all unrelated files. 23:21
23:25 ruoso joined
Whiteknight NotFound++ 23:29
dalek r30074 | julianalbo++ | trunk: 23:30
: allow null argument in string_to_int signature
diff: www.parrotvm.org/svn/parrot/revision?rev=30074
Whiteknight NotFound++ on the codingstd fix too 23:32
Whiteknight needs to watch his codingstd
23:41 tetragon joined
jonathan by some miracle managed to commit a pile of C today that didn't fail any of the codingstd! 23:41
Whiteknight very nice 23:45
I've never committed C code that didn't fail tests
even my fixes fail something :)
jonathan I cheated by writing new code and only checking in tests that I knew it could pass. ;-)
Whiteknight so, more of a manufactured miracle then :) 23:46
jonathan Well, it's the codingstd tests that I usually somehow manage to make an epic fail of. :-)
Whiteknight tell me about it... 23:47
purl it is just that our experimental liposuction machine... well... let's just say that even a really powerful shop vac isn't the right tool. or 10am
jonathan The real test will come when we swap the Parrot MMD implementation for my one in Rakudo. Then we'll see just what breaks. :-)
Whiteknight purl forget it
purl Whiteknight: I forgot it
jonathan it would be good if purl had a list of things just to never remember
Whiteknight (see just what breaks)++
purl permanently forget it 23:48
purl Whiteknight: huh?
Whiteknight damnit
jonathan :-)
purl would no way be *that* co-operative.
purl jonathan: sorry...
jonathan Though at least she's apologetic about it.
jonathan pets purl
purl don't mess with me
Whiteknight slaps purl with a trout 23:49
jonathan decides purl won't make a suitable girlfriend after all...
Whiteknight especially not with the trout smell
cotto_work purl, it is <reply>
purl OK, cotto_work.
cotto_work purl, it?
purl cotto_work: wish i knew
cotto_work purl, it is something
purl ...but it is <reply>...
cotto_work win 23:50
jonathan purl, it?
purl jonathan: no idea
jonathan Wow. :-)
23:50 ruoso joined
jonathan tries to get his YAPC::EU slides done... 23:51
23:56 bacek joined

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!