#parrot Parrot 0.5.2 Released | parrotcode.org/ | see www.parrotcode.org/misc/parrotsketch-logs/ for logs
Set by moderator on 11 February 2008.
svnbotl r25858 | chromatic++ | trunk: 00:11
: [src] Added documentation to undocumented functions (RT #48260,
: ajr@ippimail.com).
diff: parrotvm.org/svn/parrot/revision/?rev=25858
00:42 AndyA joined 00:43 AndyA joined
Coke chromatic++ 00:50
svnbotl r25859 | chromatic++ | trunk: 00:54
: [src] Improved the formatting of this file and fixed a couple of typos.
diff: parrotvm.org/svn/parrot/revision/?rev=25859
01:27 Ademan joined 01:46 rdice_ joined 01:51 Ademan joined 01:55 kid51 joined
diakopter ∞ 02:09
Coke +1 02:24
purl 1
TimToady ∞+1 02:43
Coke tries not to get sucked into Inform 7.
TimToady ∞+1 # --buzz lightyear
except buzz is more like ∞ + $n if $n > 0 02:44
svnbotl r25860 | jkeenan++ | reconfigure:
: Removing branch; all its good stuff is now in tcif branch.
diff: parrotvm.org/svn/parrot/revision/?rev=25860
TimToady wait, "and beyond", so make that >= 02:45
02:48 skids joined
svnbotl r25861 | jkeenan++ | trunk: 02:51
: Removing superseded tag.
diff: parrotvm.org/svn/parrot/revision/?rev=25861
02:58 petdance joined
Tene takes a machete to languages/cardinal/src/parser/* 03:09
apparently cardinal doesn't parse () in function calls 03:12
PerlJam Makefile:631: *** Recursive variable `MAKE' references itself (eventually). Stop. 03:22
Am I the only one who gets this?
(running on cygwin btw)
Coke ... cygwin works? =-0 03:37
infinity... and even further! 03:38
PerlJam: I don't seem to be getting an error on 'make' 03:39
(osx)
Tene comments on getting a commit bit? 03:51
03:52 avar joined 04:07 AndyA joined 04:16 shamu joined 04:37 particle joined 05:09 grim_fandango joined
svnbotl r25862 | allison++ | pdd17pmc: 05:50
: [pdd17pmc] Update the Class PMC to use a generated attribute struct.
diff: parrotvm.org/svn/parrot/revision/?rev=25862
06:39 skv joined 06:54 Coke joined
svnbotl r25863 | chromatic++ | trunk: 07:11
: [src] Removed unreachable code after exceptions (RT #50886, Andrew Whitworth).
diff: parrotvm.org/svn/parrot/revision/?rev=25863
07:16 uniejo joined
svnbotl r25864 | chromatic++ | trunk: 07:24
: [src] Removed unnecessary casts (RT #50238, julian.notfound@gmail.com).
diff: parrotvm.org/svn/parrot/revision/?rev=25864
07:32 jjore joined
svnbotl r25865 | petdance++ | trunk: 07:43
: hoist repeated code
r25866 | fperrad++ | trunk:
: [Lua]
: - aligned with Lua 5.1.3 (part 2)
07:49 Ademan joined
svnbotl r25867 | petdance++ | trunk: 07:50
: the ext parm is NULLOK
r25868 | petdance++ | trunk:
: nosync_pop_entry cannot actually return NULL
diff: parrotvm.org/svn/parrot/revision/?rev=25868
08:10 iblechbot joined 08:15 HG` joined
svnbotl r25869 | paultcochrane++ | trunk: 09:23
: [core] Removed deprecated instantiate opcode as per patch in RT#50938 (with
: the one minor modification that I needed to update the test plan).
diff: parrotvm.org/svn/parrot/revision/?rev=25869
r25870 | kjs++ | trunk: 09:42
: [src] add function doc to oo functions.
diff: parrotvm.org/svn/parrot/revision/?rev=25870
09:54 IllvilJa joined 10:05 mj41 joined 10:39 cognominal_ joined 10:48 e joined 11:16 IllvilJa joined 11:39 ruoso joined 11:42 marmic joined 12:07 kid51 joined 12:24 diakopter joined 12:42 diakopter joined 13:15 mj41 joined 13:18 kj joined
Tene is finding ast problems in cardinal similar to those found in lolcode. 13:50
Tene needs to create doc patches.
kj Tene: what kind of problems? 13:54
Tene kj: parameters to blocks weren't being passed in right. 13:56
I've got function calls working now.
kj ok. was it in the parameter passing stuff?
I didn't include tests for that i think 13:57
Tene And a couple of other things, iirc.
the () in foo(1) wasn't supported anywhere in the grammar.
As it is right now, though, foo(bar(1)) doesn't parse.
kj i'm not sure if that's the only thing missing ;-)
Tene There are some things I don't quite understand the reasoning behind in the grammar. 13:58
Are you particularly attached to the grammar as it stands?
kj no not at all 13:59
please go ahead and claim it ;-)
Tene Ahh, okay.
kj I just wanted it running on PCT
at least some basics
Tene I'll consider more-significant restructuring, then.
kj like nodes for literals and functions. once those are working, the rest is easy :-)
Tene I need to get a commit bit eventually. 14:00
kj yes. please note that some grammar references found on the internet don't seem to be correct
at some point I looked at the yacc file found in ruby distr. 14:01
Tene My reference right now is /usr/share/doc/ruby-1.8.6.111/sample/**.rb
kj ok
Tene Like you, iirc, I don't actually know ruby at all. :) 14:02
kj heh. i found that it doesn't really matter if you don't. Of course it helps a bit 14:03
Tene: i would suggest to take at least a look at several sources for the grammar, and then select the one that looks best 14:09
(and preferably the most up-to-date one)
14:18 gryphon_ joined 15:00 mdiep joined 15:10 skids joined 15:15 rdice joined, jhorwitz joined 15:46 peeps[work] joined 16:14 e joined 16:29 jmcadams joined 16:34 Dave joined, parrot-poke joined 16:40 sjansen joined 17:03 particl1 joined 17:42 DarkWolf84 joined
Coke 17:45
spinclad too 18:29
Tene not
spinclad not! :( 18:30
todo el mundo sería sí 18:31
18:31 allison joined
Tene Well, only on Thursdays. 18:32
18:32 chromatic joined
spinclad (#ps?) 18:33
(yes!)
chromatic yes
Coke crap! 18:35
chromatic carp!
particle i'm in, too 18:36
18:48 proq joined
skids chromatic: re your testing article, I have a crusty debian distro on hp-parisc, if that's bizarre enough. But I can't leave it on 24/7 (300Watts, ouch) 18:49
chromatic That's bizarre enough. 18:50
Would you consider running a smoke once a week on there?
svnbotl r25871 | kjs++ | trunk:
: [NEWS] add cardinal and function documentation to news.
diff: parrotvm.org/svn/parrot/revision/?rev=25871
skids When I can remember to power it on, sure. And assuming I can get it to run in the first place. I'll give a try sometime this week or next. 18:52
proq how do I register my nick on this network? O_o
Coke I don't think that's an option.
chromatic Thanks, skids. I'm sure the first test results will be horrifying! 18:53
Coke "It came from beyond the repository"
particle proq: no nick registration on magnet 18:54
skids Yeah, probably. If I have time I'll look into whatever passes for atomics on that CPU.
proq oh, true enough 18:55
skids I also have an alpha and a non-Darwin G4, but those are probably much easier to come by.
chromatic I don't think we have anyone running on the Alpha either. 18:56
What's on the G4?
spinclad yeah, alpha seems rare these days
skids Fresh debian.
particle do we have anyone running linux-ppc?
i don't think so
chromatic I just fixed Linux PPC on Saturday. All green
(That's why JIT works now on Darwin PPC)
proq I was considering running YDL on my G4 a few days ago :) 18:58
skids almost got debian wedged onto that G4 without plugging in a monitor. Almost.
particle anybody running haiku?
Coke (jit on darwin ppc) oooh, really?
now if we could get jit on osx x86... =-)
chromatic Coke, try the shootout on Darwin PPC. 18:59
particle sweet. haiku has nightly vmware images to download :) 19:00
Coke that will take a day. =-)
but ok.
chromatic It shouldn't; the JIT works now.
Coke most of that day doesn't involve building or running. 19:01
particle it's only 30MB zipped.
Coke chromatic: two tickets open. looks like it's probably PARROT_API wonkery. 19:08
svnbotl r25872 | kjs++ | trunk: 19:09
: [NEWS] More news. this is as much as I can think of.
diff: parrotvm.org/svn/parrot/revision/?rev=25872
chromatic Existing tickets, or new ones?
19:10 Ron joined
Coke new, just opened. 19:10
... though they haven't hit the list yet.
chromatic Beware. I hit back.
Coke 51000 and 51002
purl 102002
Coke ... stupid purl. 19:11
10 minus 3?
ha!
particle ha! haiku vm running :) 19:13
Coke chromatic: ... did you intentially fix the ascii_compare segfault?
chromatic No. 19:14
Coke I can't reproduce it. :(
kj particle: can you summarize what's new in rakudo? THen I'll update NEWS on that.
chromatic In tcl or anywhere?
Coke (currently chugging through a full 'make test')
tcl was the only place it was happening before.
svnbotl r25873 | kjs++ | trunk:
: [NEWS] I forgot the new PCT documentation. Yay!
diff: parrotvm.org/svn/parrot/revision/?rev=25873
Coke ... well, damn.
So, it's not fixed, necessarily, but it's hidden again. whee. :( 19:15
chromatic Maybe r25851.
... which makes a lot of sense, actually.
Coke could be. once this test run finishes, I'll revert that patch and see if it explodes again. 19:16
chromatic I have a good feeling about it.
Coke (locally)
... that's what you said about the nail gun.
... now if you could just make tcl.pbc run code about 60 times faster... 19:17
chromatic When you see me in a pink tutu waving a fairy wand, then you'll know I'm your fairy godmother and can do things like that. 19:18
proq the FAQ on using parrot for, say, ruby code says I need to modify languages.pm. I take it this means I need to download the parrot source, modify and build? 19:20
chromatic proq, there's already a Ruby compiler in progress (languages/cardinal/), so you don't have to modify languages.pm.
kj proq: using parrot for ruby code: you mean, using the ruby compiler for parrot
Coke ok. tcl current is failing tests, but no segfaults. 19:21
proq I mean using parrot to compile ruby code to ELF 19:22
kj proq: parrot is the runtime engine, it can't compile ruby code
it can *run* ruby code
proq oh, I see
kj using the ruby compiler targeting parrot, which is cardinal
(though, I should mention, it's not complete yet) 19:23
chromatic No reason you can't build a Ruby-to-ELF compiler with Parrot though. You just won't emit PBC/PIR.
spinclad which compiles to pir, not ELF
(which = /languages/cardinal)
integral hmm, you could compile parrot's ops to LLVM IR, then output code that's calls to the op functions, then use LLVM's inliner, then compile from IR to native 19:27
Coke chromatic: I don't think that was one that fixed it.
svnbotl r25874 | kjs++ | trunk:
: [NEWS] add deprecation entry. add "attribute" scope variant for PAST::Var nodes.
diff: parrotvm.org/svn/parrot/revision/?rev=25874
Coke if I revert 25851, tcl doesn't segfault like it used to.
ah well.
onward and upward. :|
proq oh, I get it. a cardinal is a ruby parrot? :D 19:29
spinclad exactly! 19:30
purl it has been said that exactly! is it not awesome?
particle we're *so* clever here :)
Coke Yah, I can hack on tcl again! Crap, it's so slow!
particle holy crap, who screwed up trace output? 19:31
Coke kj: have you added every bit of news through r25873?
(if not, the blurb in the NEWS file is somewhat misleading. =-)
particle: ?
chromatic Coke, you'll have to clear out bytecode if you revert 25851.
At least Tcl bytecode.
kj Coke: nope, this is what I know about
burb?
nopaste "particle" at 24.19.3.148 pasted "tail of trace output on pdd17pmc segfault" (487 lines) at nopaste.snit.ch/12358
kj ehm. blurb?
what's that?
purl hmmm... that is what he is using :)
Coke kj;you said "news up through r<foo>". It's not. 19:32
kj mmm. that's not what I meant either. but I guess you can interpret it that way
which i didn't
Coke that was originally meant a placeholder if some intrepid soul was doing ALL the news. If you're just cherry picking, don't include that.
what does it mean to you?
chromatic msdn.microsoft.com/events/hero/sfbio/
Coke chromatic: yah, it was a clean tclsh build.
kj as a placeholder too, as there's no version number yet 19:33
Coke but what the placeholder mean?
kj should I remove it? 19:34
spinclad use it as a fence between the complete part and the cherries
Coke with all the updates, I don't think we know where the fence is. 19:35
kj ok. I'll leave it alone next time. I don't know exactly what is new.. there's so much new
chromatic Coke, before that patch, IMCC always tried to load a library named '' when it encountered .HLL 'lang', '' 19:36
Coke tcl does that a lot. 19:37
chromatic Yeah, that's the only way to specify an HLL.
Coke so yes, it was plausible; I just didn't see an effect.
chromatic Thus the interpreter froze an empty string into the bytecode for the associated HLL library name.
Thawing that bytecode would give you an empty string for the HLL lib name.
Coke I could binary search, but doing that to find when something began working is far less helpful than finding when something stopped working. 19:38
chromatic I agree.
Still, it would be good to know if we've really fixed it.
Coke True.
chromatic If you get really, really bored and have a lot of spare time and nothing productive to do, like clip your toenails or whatever, I'm curious. 19:39
But you'd have to be very bored.
Coke crap. I just clipped my toenails yesterday.
chromatic Re-reading Howard Roarke's 40-page speech bored.
davidfetter if you're *that* bored, i have ideas ;) 19:40
19:40 peepsalot joined
chromatic I'd rather start my own steel mill, with robots, and blackjack, and... 19:41
Coke So; while I want to track all of my tcl issues, I am no longer convinced that rt is the place to do that.
I can see keeping RT for languages that *ship* with parrot, but I'm not sure tcl's going to make that cut. 19:42
any thoughts? (ditto for any other HLLs except rakudo which has its own special RT queue.)
chromatic Depends on whether an issue is a Parrot issue or a Tcl issue. 19:44
Coke anything truly parroty would have an rt ticket. 19:45
particle chromatic: nopaste.snit.ch/12358 19:46
chromatic particle, I'll poke at that. Thanks!
How weird that the output goes wonky at 958 though. 19:47
particle it goes wonky on and off throughout the run 19:49
19:54 Crassworm joined
proq Cardinal - Ruby for the Parrot VM irb(main):001:0>> 20:01
^^ sweet!
kj proq: :-) It's not complete yet. but patches are welcome ;-) 20:02
jhorwitz considers mod_ruby... 20:03
chromatic Cardinal runs irb? 20:05
kj chromatic: nope , that's just the same command prompt as you get when running interactive ruby
cognominal_ where is the code that dumps POST and PAST trees, in POST dumps, source and pos with value null is just noise 20:10
20:21 PacoLinux joined 20:26 Steve_p joined
particle look! it's a steve_p! hey there. 20:29
20:42 allison joined
Coke OMG, it's allison! =-) 20:42
particle and just miles away from me :) 20:43
Steve_p hey 20:45
Steve_p digs himself out from his work
cognominal_ anyone here going to FOSDEM apart pmichaud? 20:48
s/apart/but/ ?
svnbotl r25875 | chromatic++ | trunk: 20:49
: [src] Silenced warnings when compiling with Intel C++ (RT #51004, Steve Peters).
diff: parrotvm.org/svn/parrot/revision/?rev=25875
Coke is not. 20:54
allison hi coke :) 21:03
Coke is no longer getting paid to hack java. 21:09
Tene Would you like to be?
Coke not especially. 21:10
Though my current deployment is mostly CFMX at the moment; I almost miss java. =-)
Steve_p I'm have grown very sick of Java 21:16
Why are the tests phoning home? 21:24
chromatic Which tests are these? 21:27
Steve_p t/distro/file_metadata
nopaste? 21:28
purl nopaste is at nopaste.snit.ch/ (ask TonyC for new channels) or rafb.net/paste or paste.husk.org/ or poundperl.pastebin.com/ or paste.scsys.co.uk/ or don't bother me while I'm eating
nopaste "Steve_p" at 165.189.41.11 pasted "Test phoning home..." (14 lines) at nopaste.snit.ch/12359 21:29
svnbotl r25876 | rblasch++ | trunk: 21:32
: [src] Force values into range where appropriate, so we can find the places
: where it's not, and values are truncated.
diff: parrotvm.org/svn/parrot/revision/?rev=25876
chromatic Oh yeah, those tests. 21:33
Did you check out Parrot via svn? 21:34
Steve_p Yes
Coke that shouldn't have to phone home.
Steve_p Shouldn't is the operative word :) 21:36
Coke does perl -Ilib -MParrot::Revision -e1; phone home? 21:40
file_metadata.t seems to just be invoking 'svn pg <attr> <filelist.' 21:42
Ron Coke: It seems to be the "svn ls" 21:53
Coke ... where is that done? 21:54
Ron In the BEGIN block, line 244. 21:55
Coke oh, in the beginning.
Steve_p Coke, no 21:56
Coke Steve_p: yah, ron found it. 22:00
I wouldn't expect 'svn ls .' to do that, though.
Ron Me neither.
BUT...
Coke Can one of you open a ticket? 22:01
Ron svn help ls
Sure.
Coke oh! '.' doesn't mean what I think it means. :|
particle i need a good header location for internationalization settings
#ifdef HAS_GETTEXT ... 22:02
Coke locale.h ?
Ron Yep, svn ls always heads for the server.
svnbotl r25877 | rblasch++ | trunk:
: [src] Declare variable at beginning of block.
diff: parrotvm.org/svn/parrot/revision/?rev=25877
particle i suppose i could create a new header, but hoped to stick these defs in an existing header
Coke conversation of headers? 22:03
misc.h, then?
particle ah. "conservation"
Coke whoops. 22:05
Coke disengages autopilot on his fingers. 22:06
particle i'm thinking intl.h
or i18n.h
Coke I like my bikeshed betteer. 22:07
particle alright, misc it is ;)
svnbotl r25878 | rblasch++ | trunk:
: [src] PARROT_ASSERT is used to assert pointers, so it should be at least as
: long as them. long may not be large enough (e.g. on LLP64), but INTVAL
: probably should.
diff: parrotvm.org/svn/parrot/revision/?rev=25878
Coke do we have a type that == "size of a pointer" ? 22:08
particle Ron: are you working towards parrot compilation with msvc on x86_64?
we have ptrsize 22:09
Ron Yep.
particle \\o/ ron++
Coke (if we need something as big as a pointer, is that better than INTVAL ?
particle on msvc/x86_64, intsize >= ptrsize 22:10
Ron But I only got as far as the last time we tried....yet
Is INTVAL not supposed to be long enough to hold a pointer?
particle that question is too vague for me to answer 22:12
parrot frequently and mistakenly assumes intsize == ptrsize
Ron Uups, I thought that too. 22:13
particle we need func to data and data to func macros that do bounds checking 22:15
22:15 dalek joined
chromatic If you have the Perl 5 source, look at F2DPTR and D2FPTR 22:16
particle i haven't it, but that can be fixed 22:17
PerlJam mugwump has a git repository of it :) 22:19
Ron FPTR2DPTR, right?
chromatic That's right. 22:20
Ron The first thing I stumbled over on Win x64 was config/auto/pack.pm. 22:26
The _set_ptrconst test
I'm still a bit puzzled about the relative test ($longsize == $ptrsize) and the absolute value "ul". 22:28
particle my parrot configs and builds on msvc/x64, but when i run it interp is at a different location in every stack frame
Ron That's just a regular build, right? 22:29
particle yes
Ron Not a "--intval=long long" 22:30
ahh, ok.
particle vanilla configure/nmake
Ron I haven't tried that yet.
particle there's a long way to go
Ron With intval "long long" I got a basic interpreter working. 22:31
Something like "hello world" and "ack 1"
particle cool. can you compile pge, or do you just try examples?
ok
Ron Well, that's the problem. The first fault occurs when building PGE. :-( 22:32
But I'm on it.
particle ja, happens here, too
Ron I tried building with -RTCcsu, to enable compiler bounds checking. 22:33
Nothing spectecular, just for example the PARROT_ASSERT thingy.
But you're right, it won't be an easy fix. 22:35
particle yep. has to be done, though.
Ron So, an INTVAL can basically be any size, unrelated to a pointer, right? 22:36
I got a bit distracted by the ptrsize != intvalsize "warning" in config/auto/sizes.pm 22:37
I need to be off for today, thanks particle++ and chromatic++ 22:41
Hope I'll soon have good news from x64. 22:42
22:43 peepsalot joined 22:49 Limbic_Region joined 22:54 wknight8111 joined 22:57 silug joined 23:21 allison joined
particle Coke: if you get a chance, see if there are old files to be removed in languages/APL/ after pmichaud's conversion to PCT in r25139 23:41
particle is working on NEWS updates
chromatic 25 tickets creates this week. 25 tickets closed this week. 23:43
particle (treading water)++ 23:44
23:44 teknomunk joined
chromatic It'd be nice to cut down on the 110 open too. 23:48
109 open. 23:50
108 23:52
feel free to join in everyone
or else finish editing this article for me... 23:53
particle has ~2450 lines of svn log still to review
wknight8111 what article? 23:54
particle coke: please summarize your macports work in NEWS when you get a chance 23:55
chromatic An article completely unrelated to Parrot, Perl, or virtual machines. 23:58
wknight8111 good, because i'm not an expert on parrot, perl, or virtual machines!

Powered by Cro and the Raku® Programming Language.

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

Thank you!