www.parrotcode.org | Last Release: 0.7.0 "Severe Macaw"
Set by moderator on 28 August 2008.
00:06 cotto_work joined 00:12 davidfetter joined 00:22 bacek joined 00:37 bacek joined 00:40 kid51 joined 00:41 kid51 joined 00:42 kid51 joined 00:43 kid51 joined 00:45 AndyA joined 00:57 Zaba joined 01:01 megamic joined 02:08 teknomunk_ joined 02:13 Ademan joined 04:30 Zaba_ joined 04:33 tetragon joined 05:18 Psyche^ joined 06:02 zostay joined 06:12 Ademan joined 06:32 uniejo joined 06:44 mapott joined 07:02 dmknopp joined, Ademan joined 07:03 Zaba joined 07:34 iblechbot joined, viklund joined
dalek r30718 | chromatic++ | trunk: 07:41
: [dotnet] Migrated ops and PMC to new exception style (Reini Urban, RT #58176).
diff: www.parrotvm.org/svn/parrot/revision?rev=30718
r30719 | chromatic++ | trunk: 07:49
: [config] Generate an installable parrot_config fakecutable (Reini Urban, RT
: #58354, with thanks to Gerd Pokorra for reporting).
diff: www.parrotvm.org/svn/parrot/revision?rev=30719
r30720 | chromatic++ | trunk: 07:53
: [ports] Updated specfile to reflect installable parrot_config fakecutable (Gerd
: Pokorra, RT #58354).
diff: www.parrotvm.org/svn/parrot/revision?rev=30720
08:24 clunker3 joined 08:28 masak joined 08:42 AndyA joined
dalek r30721 | kjs++ | trunk: 08:52
: [pirc/new] MAJOR UPDATES to allow all PIR keywords as identifiers.
: * grammar is probably not as complete as before, but this will be fixed soon.
: * semantic actions must be fixed; just parsing now.
: * constant folding and instruction selection is mostly removed at this point, but will be restored shortly
: * some compilation warnings, which will also be fixed shortly.
diff: www.parrotvm.org/svn/parrot/revision?rev=30721
08:59 kj joined, cosimo joined 09:03 kjs_ joined 09:06 kjs_ left, kjs_ joined 09:13 pako joined
dalek r30722 | allison++ | pdd27mmd: 09:15
: [pdd27mmd] Fix wrong vtable function description.
diff: www.parrotvm.org/svn/parrot/revision?rev=30722
09:48 bacek left 09:56 cosimo joined 10:31 Zaba_ joined 10:44 mapott left
cognominal tries to grasp the relationship between parrot interpreters, task and threads. 11:19
hope pdd25 will help him 11:21
11:52 masak joined
dalek r30723 | smash++ | trunk: 12:22
: [ports/debian]: updates from last build
diff: www.parrotvm.org/svn/parrot/revision?rev=30723
12:23 tetragon joined
dalek r30724 | allison++ | pdd27mmd: 12:26
: [pdd27mmd] Shift 'mod' and 'cmod' over to real opcodes, eliminate 'cmodulus'
: vtable functions.
diff: www.parrotvm.org/svn/parrot/revision?rev=30724
12:29 iblechbot joined 13:30 gryphon joined 13:43 eee|busy joined 13:57 Zaba joined 14:07 AndyA joined
masak is there a Rakudo day planned for this week? 14:31
15:13 Andy joined
dalek r30725 | allison++ | trunk: 15:16
: [cage] Note deprecation of 'cmodulus' vtable functions.
diff: www.parrotvm.org/svn/parrot/revision?rev=30725
15:20 gmansi joined, moritz_ joined
dalek r30726 | allison++ | pdd27mmd: 15:22
: [pdd27mmd] Remove implementations of deprecated 'cmodulus' vtable function.
diff: www.parrotvm.org/svn/parrot/revision?rev=30726
16:32 ilbot2 joined
moderator www.parrotcode.org | Last Release: 0.7.0 "Severe Macaw"
16:33 moritz joined
cotto_work Parrot VM: PANIC: ippt is an unknown signature type. 16:34
NotFound cotto_work: need my last commit and configure 16:35
Tene I'm apparently incompetent with make.
particle make competent
Sorry, i don't know how to make 'competent'.
nopaste "tene" at 166.70.38.237 pasted "tene's makefile fail" (23 lines) at nopaste.snit.ch/13934 16:36
Tene I patch cardinal's makefile like that, and run 'make test.pir', and it says 'Nothing to be done for `test.pir'.'
particle does test.pir exist? 16:37
NotFound Tene: --output=test.pir ?
Tene test.pir does not exist.
cotto_work NotFound++
NotFound Ah, is cardinal, output, sorry.
particle try make -n test.pir
Tene particle: same response
cotto_work it's like MSPaint v.01 for Parrot
particle tene: did you make Makefile? 16:38
NotFound cotto_work: v.01 VHS
Tene Oh, Make is sensitive to tab vs. space.
particle hee, yep
NotFound (not even Beta)
Tene That's awesome.
cotto_work Tene: very
particle i think python stole that from make ;)
cotto_work it doesn't exit cleanly and it loses my artwork when I switch between virtual desktops, but that's pretty awesome 16:39
NotFound cotto_work: press any key to exit
particle can anybody say LOGO?
16:39 Theory joined
cotto_work ok. I was using the WM's close button 16:40
NotFound cotto_work: too much for a first test
cotto_work I shouldn't really be pointing out bugs, as it's just a demo 16:41
dalek r30728 | tene++ | trunk:
: [cardinal]
: * Set :load on the main sub of each file for precompiled modules to work.
: * Precompile test.rb 16:42
diff: www.parrotvm.org/svn/parrot/revision?rev=30728
Tene NotFound: I get a segfault 16:43
NotFound Tene: can you trace it? 16:44
nopaste "tene" at 166.70.38.237 pasted "nci xlib test backtrace" (22 lines) at nopaste.snit.ch/13935 16:45
Tene NotFound: dies as soon as there's any mouse activity in the window
NotFound Tene: What architecture? 16:46
purl architecture is great and all, but what I don't want to do is make it work right in every damn browser in all the right ways
Tene NotFound: x86_64
NotFound Tene: sorry, the event ManagedStructure is hand coded for x86_32 16:47
Tene ah
nm, then
NotFound Can I add it to examples/nci/ ? 16:50
16:51 kj joined
cotto_work NotFound, can you make it fail gracefully on non-x86 archs? 16:55
NotFound cotto_work: don't know how to detect unsigned long and pointer sizes from pir 16:57
17:01 Zaba joined
particle NotFound: use Parrot/Config.pasm 17:03
dalek r30729 | moritz++ | trunk: 17:11
: [rakudo] add traits test to spectest_regression
diff: www.parrotvm.org/svn/parrot/revision?rev=30729
17:11 cognominal joined
NotFound intsize, longsize and ptrsize ? 17:15
particle or platform, or whatever you think is best 17:23
nopaste "NotFound" at 213.96.228.50 pasted "nci xlib test - 2" (686 lines) at nopaste.snit.ch/13936 17:26
NotFound Tene: can you test it now?
Tene Sure. 17:30
NotFound: it doesn't crash, but it doesn't do anything else either. 17:31
When the mouse cursor is down and I drag, it prints numbers to stdout
NotFound Tene: Does not close when pressing a key?
Tene Oh, yes, it does close 17:32
NotFound Tene: please do this: grep "typedef .* Time" /usr/include/X11/* 17:33
dalek r30730 | chromatic++ | pdd27mmd: 17:35
: [PMC] Changed the second occurrence of get_number to set_number_native, fixing
: a compilation warning. 17:36
diff: www.parrotvm.org/svn/parrot/revision?rev=30730
Tene /usr/include/X11/X.h:typedef unsigned long Time; 17:37
/usr/include/X11/X.h:typedef CARD32 Time;
dalek r30731 | moritz++ | trunk: 17:39
: [rakudo] one more test for spectest_regression
diff: www.parrotvm.org/svn/parrot/revision?rev=30731
nopaste "NotFound" at 213.96.228.50 pasted "nci xlib test - 3" (691 lines) at nopaste.snit.ch/13937 17:43
NotFound Tene: again, please. 17:44
Tene Still acts the same. Numbers on stdout when click+drag, and close on keypress. 17:45
NotFound Tene: move the pointer close to upper left and tell me the numbers
The las two numbers 17:46
What is the size of int int x86-64? 17:47
Tene 6: 386547056697 0 0 208638299
6: 390842023993 0 0 208638695
particle NotFound: would it help if tene pasted his Parrot::Config? 17:48
NotFound Sure 17:49
Tene lib/Parrot/Config.pm ?
particle yep 17:50
nopaste "tene" at 166.70.38.237 pasted "Parrot::Config" (46 lines) at nopaste.snit.ch/13938
particle oops, i mean Parrot::Config::Generated of course :(
i long ago forgot about that switch
nopaste "tene" at 166.70.38.237 pasted "Parrot::Config::Generated" (1525 lines) at nopaste.snit.ch/13939
18:10 peepsalot joined
nopaste "NotFound" at 213.96.228.50 pasted "nci xlib test - 4" (699 lines) at nopaste.snit.ch/13940 18:21
NotFound Another try 18:22
Tene Same behavior as before. 18:23
What's it supposed to do?
NotFound Tene: draw points when button down and moving
Tene ah
NotFound Mmmm... forgot the 'Bool' type 18:25
please do this: grep "typedef .* Bool" /usr/include/X11/*
Tene /usr/include/X11/Xdefs.h:typedef int Bool; 18:27
cotto_work particle: ping 18:29
particle pong
18:33 rurban_away joined
nopaste "NotFound" at 213.96.228.50 pasted "nci xlib test - 5" (699 lines) at nopaste.snit.ch/13942 18:52
NotFound Last attempt
rurban I've added more cygwin070patches to parrot-porters. Anybody knows about my cla status? 18:55
particle rurban: chromatic generally checks the po box, best to contact him 18:56
rurban he was here yesterday. hmm. 18:57
NotFound rurban: Did you have the cygwin xserver?
rurban I do have. Do want the xlib nci to be tested?
NotFound rurban: please do. 18:58
rurban I suppose it should be in runtime/parrot/library/xlib.pir, right?
NotFound rurban: is just a test program, not a library yet. 18:59
dalek r30732 | moritz++ | cygwin070patches:
: [config] more Makefile hacking
: * fix a perlcritic test for Compiler.pm
: * add DESTDIR to languages make install and test-installable
: * fix forth and tcl build/installed seperation and fixed pbc destination paths
: * make builds the default targets, all builds all.
: * fix wrong #ELSEIF to #ELSIF in language makefiles
: Patch courtesy by Reini Urban, rurban++
diff: www.parrotvm.org/svn/parrot/revision?rev=30732
rurban moritz++ thanks! 19:00
19:00 rdice joined
rurban NotFound: it said: Init xlib 19:03
No xlib
I'll have to check if I have it.
NotFound rurban: xlib = loadlib "libX11" --> This works in cygwin or needs some change? 19:04
rurban loadlib 'libX11' should work 19:05
nope: it's cygX11-6 on cygwin 19:08
Init xlib
Init xlib Display
Parrot VM: PANIC: pppiiiiiii is an unknown signature type.
CAN_BUILD_CALL_FRAMES is disabled, add the signature to src/call_list.txt!
purl i already had it that way, rurban.
rurban simple error, I'll add it
19:09 Zaba_ joined
NotFound rurban: requires r30727 at least 19:09
19:15 mj41 joined
rurban NotMuch: But this sig is still not in call_list/misc.in 19:18
NotFound rurban: I added it to trunk in r30727 19:19
rurban cygX11-6 is for the current x11-org package, but this name changes with every major update
particle you need to reconfigure, too
rurban just doing, need time... 19:20
NotFound rurban: this is a quick test, later a stable way to find the lib may be added.
rurban Ths only one which comes close is: i ppiiiiiiii
NotFound rurban: at the end of the file must a be a section: # Testing xlib 19:22
rurban OK, now I get a simple segfault 19:25
NotFound Nice 19:27
rurban I've looked in the wrong branch. The sig is there. xlogo and xclock works fine.
sorry: it works fine. Hello DisplayName: "" 19:30
nopaste "rurban" at 84.169.232.81 pasted "xlib.pir on cygwin" (7 lines) at nopaste.snit.ch/13944 19:31
rurban And with DISPLAY set it even paints a window with title "Hello, parrot" 19:32
NotFound rurban: good!
rurban Just the close button causes an XIO fatal 19:33
fatal IO error 11
NotFound rurban: move the mouse with a button pressed, and see if it draws dots.
rurban It does
NotFound Close button does not work, press any key to close.
rurban Ok, this is fine
NotFound Good, thanks. 19:34
rurban Now I have to find out why opengl does not work with X11 on cygwin.
freeglut needs DISPLAY set at Configure. OK 19:35
MSVC also? I have not set thi up yet.
dalek r30733 | julianalbo++ | trunk: 19:36
: add examples/nci/xlib.pir
diff: www.parrotvm.org/svn/parrot/revision?rev=30733
moritz NotFound: that thing opens a white window with no contents for me - is that how it should be? 19:37
NotFound moritz: with "Hello, parrot" in the title bar. 19:38
rurban yes, just draw with the left mouse
moritz ah, yes
cool ;) 19:39
NotFound Not very eye candy, but is a good test of nci capabilities. 19:40
moritz indeed ;)
now somebody please write a connector to mysql or pg ;)
rurban should we really add such libs via plain pir, or better with nqp or such? 19:41
pq is already there
19:41 gryphon joined
rurban qt, gtk2, xlib, win32api and such 19:42
moritz is it? where?
rurban library/posg
NotFound rurban: don't know, this is my first attempt with nci, don't even know if it's usable from HLLs.
rurban library/postgres.pbc 19:43
library/Pg.pir 19:44
and so on. 19:45
mysql is missing though. This should be done via nci 19:46
ncigen probably
moritz doesn't mind which open source database is available 19:47
NotFound A parrot DBI will be a good thing.
moritz if I could use one of them through Rakudo that would rock
rurban better write that in a HLL 19:49
perl6 e.g. those pbc's are installed for sure and available for every other HLL 19:50
NotFound By the way, I think parrot links with too much libraries, will be better to dinamically load anything not required by the core.
moritz if HLL interoprability works
rurban ++
pg and opengl esp.
NotFound Even readline
rurban hmm, readline is needed by the utils
NotFound rurban: I disagree, is nive to have it available, but not a requirement. 19:51
s/nive/nice
rurban it's just a pbc, not in the shared lib.
japhb rezzes for a moment 19:52
NotFound ldd parrot --> ...... libreadline.so.5 ....
rurban okay. but this is a basic requirement for almost every such binary 19:53
japhb NotFound: last I heard from the DBI maintainer, the plan is to try to get Parrot able to use JDBC drivers, and build DBI 2 on top of that.
NotFound Not a bad idea. 19:54
This will allow, for example, talk to Oracle without a Oracle client. 19:55
particle yep, timbo wants to use the jdbc api
moritz basically because it is documented ;)
and tested
and established
particle and everywhere
japhb His argument was that the JDBC drivers have a very well defined surface (unlike Perl 5 DBDs), and they are where the vendors put their most stability effort.
rurban but we can read .net bytecode, but not the java bytecode yet. 19:56
so shouldn't we try such .NET lib instead?
japhb shrugs
moritz do we need to run the bytecode to be able to use their API?
rurban technically its about the same. 19:57
particle no
japhb I'd bet dollars to donuts that JDBC APIs are WAY more stable than anything for .NET. Microsoft seems to have a fetish for changing DB APIs
rurban agreed
If we could read / convert the java vm bytecode (as dotnet) we could simply use the jdbc. 19:58
NotFound Some comment about the xlib.pir code? Is my first pir program of some size, and don't know, for example, if I'm using the globals in recommended ways. 20:00
Tene NotFound: should it work with 64 yet? 20:01
NotFound Tene: I think not, nothing change from your last test, and I've runned out of ideas. 20:02
Tene Oh, right.
I'm a bit distracted today. I forgot about our several tests. 20:03
NotFound Somenone wants to donate a x86-64 machine to test in? ;)
Tene Want to ssh to my laptop?
NotFound Tene: I can try 20:04
Tene If you want me to, I can set it up. 20:05
NotFound Too late for me today, we can talk about that tomorrow, if no one is interested. 20:06
Tene Okay.
NotFound I've already reach today much more result than expected :) 20:07
Tene Yes, it's very nice.
rurban anyone knows why Jonathan didn't try to convert the JVM also? 20:15
particle he did dotnet for a school project
Tene Man, the ruby grammar is pretty crazy.
particle so, ENOTUITS|ENOCREDITS 20:16
rurban because they are very similar. And there are a lot of java libs out there which we could just use.
pmichaud particle: phone? 20:17
purl hmmm... phone is 1/4".
particle dotnet bytecode translator is far from complete
rurban and we can even convert them from a stack machine to a register machine
particle oops
Tene I'm trying to get it to give reasonably useful parse errors right now. 20:19
Right now if there's a parse failure, the error is reported as occurring quite a bit earlier 20:23
moritz not enough sequence points? 20:24
Tene sequence points? 20:25
purl sequence points are at the ? and the :
moritz Tene: stuff that prevents backtracking
Tene I'm really not sure.
I don't have a good handle on when PGE backtracks. 20:26
pmichaud if using rule and token, then it doesn't backtrack 20:27
Tene That's what I thought. 20:28
pmichaud at least not in the normal sense of backtracking
but moritz said is essentially correct -- you probably need more sequence points in the grammar 20:29
Tene The problem when trying to track down the issues is that it unwinds all the way back before failing, so I have trouble seeing what it was trying to parse. 20:34
dalek r30734 | tene++ | trunk:
: [cardinal]
: * Parse fail more appropriately in some cases
: * Update hash/hash.t to use test.rb
diff: www.parrotvm.org/svn/parrot/revision?rev=30734
20:37 Ademan joined
pmichaud Tene: yes, that's why you need the sequence points, to stop the unwinding somewhere more helpful 20:59
Tene What's the syntax?
purl hmmm... the syntax is a nightmare
pmichaud I'll give an example... just a sec
Tene Aw. purl is discouraging.
pmichaud languages/perl6/src/parser/grammar.pg:147 21:00
Tene Ah, so just what I did in my last commit.
pmichaud yes, although that looks a little funny to me. 21:01
Tene Does PGE do Perl 5 regexes?
pmichaud it does, but P6Grammar does not.
Tene Ah.
21:01 Anandeep joined
pmichaud but yes, what you did in your last patch is the general idea 21:02
once you've successfully parsed something, add a sequence point that says "if we backtrack here, the error occurred after this point because everything up-to-now was okay." 21:03
21:03 peepsalot joined
dalek r30735 | julianalbo++ | trunk: 21:03
: use XDrawLine in example xlib.pir
diff: www.parrotvm.org/svn/parrot/revision?rev=30735
pmichaud I'm looking for a good way to embed PIR into Perl 6 subs... any ideas? 21:04
Tene eval :lang<pir> ?
rurban .include ??
purl hmmm... .include is like if you'd just copy and pasted the code from the other file where the .include is.
moritz pmichaud: q:pir{...} ?
pmichaud eval is a runtime operation, so that's not really it.
moritz: you mean as a term in void context? 21:05
moritz pmichaud: if it returns nothing, yes
pmichaud conceivably the PIR could return something.
Tene pmichaud: do you know quickly if src/parser/quote_expression.pir supports parsing p5regex?
pmichaud Tene: it does not. 21:06
Tene Okay.
pmichaud but I guess if the PIR could return something, that'd be just the way it is.
Tene For now cardinal only gets P6 rules, then.
pmichaud so, q:pir might actually work out. hrm.
cotto_work is the perl6 phone meeting still in progress? 21:07
pmichaud no
typically they last at most 60 minutes
cotto_work good to know
21:09 toddr joined 21:10 toddr left
dalek r30736 | tene++ | trunk: 21:13
: [cardinal]
: * Slightly more appropriate requirements for rebuilding test.pir
: * Regex support
diff: www.parrotvm.org/svn/parrot/revision?rev=30736
r30737 | tene++ | trunk: 21:15
: MANIFEST and metadata fixes for recent cardinal files.
diff: www.parrotvm.org/svn/parrot/revision?rev=30737
21:20 chromatic joined
chromatic In Rakudo, .exists doesn't exist on Hash yet? 21:20
rurban hi chromatic: do you know about my cla?
chromatic I didn't see it the other day; I'll check again tomorrow. 21:21
rurban thanks.
moritz chromatic: I think it does 21:22
chromatic I see a multi for it, but November's CGI.pm has a comment: 21:23
(Hash.:exists not
# implemented yet)
moritz uhm.
you're right
at least the "normal" form %hash.exists('item') could be implemented, I think 21:24
dalek r30738 | allison++ | pdd27mmd:
: [pdd27mmd] Shifting 'pow', 'bor', 'band', and 'bxor' over to regular opcodes.
diff: www.parrotvm.org/svn/parrot/revision?rev=30738
chromatic I can define a method in Hash.pir; is that all?
moritz yes 21:25
(at least for a start)
chromatic and then run t/spec/S29-hash/exists.t ?
moritz I think that also tests 'exsits %hash{"item"}', which isn't specced anymore 21:26
chromatic It looks like it needs fudging anyway.
moritz oh, and it uses list assignment and hash slices
which both aren't implemented
or in other words, at the moment it's useless to us :( 21:27
chromatic I figured, but I've spent a lot more time fixing bugs in C than Rakduo.
I can't even spell Rakudo, apparently!
moritz I can add a few basic tests that are easier to work with
chromatic I'll make a basic sanity test file for now, but I won't check it in. 21:29
(also there's a typo in the word "existance")
moritz I'll fix that too, thanks 21:31
updated the test. 21:32
nopaste "chromatic" at 63.105.17.30 pasted "exists method for Perl6Hash" (16 lines) at nopaste.snit.ch/13949
chromatic Seems like that's all it takes. 21:33
21:33 toddr joined, toddr left
moritz chromatic: go right ahead and commit it, I'll make the tests runnable 21:33
chromatic I can't seem to find the %params.:exists{$key} syntax that November wants, but %params.exists{$key} seems to work
moritz that .:exists syntax is still being discussed 21:34
chromatic This ought to help them slightly though.
dalek r30739 | chromatic++ | trunk:
: [Rakudo] Added .exists method for Hash.
diff: www.parrotvm.org/svn/parrot/revision?rev=30739
pmichaud I think we could get .:exists to work without too much difficulty 21:39
the biggest "problem" is that STD.pm has substantially refactored since we wrote that part of rakudo 21:40
chromatic I just couldn't imagine writing many programs without .exist.
moritz chromatic_away: well, there's only one program in Perl 6 yet that runs on Rakudo... 21:41
(or maybe two)
dalek r30740 | moritz++ | trunk: 21:44
: [rakudo] added S29-hash/exists.t to spectest_regression, chromatic++
diff: www.parrotvm.org/svn/parrot/revision?rev=30740
cotto_work particle, ping 21:52
dalek r30741 | julianalbo++ | trunk: 22:00
: add some keyboard handling to example xlib.pir
diff: www.parrotvm.org/svn/parrot/revision?rev=30741
r30742 | tene++ | trunk:
: [cardinal]
: * Few more sequence points
: * Parse next,break,redo, but throw a more-specific error
diff: www.parrotvm.org/svn/parrot/revision?rev=30742
chromatic_away moritz: also, the irclog HTML has a typo in "suggetions" in the footer. 22:01
22:07 grim_fandango joined, slightlyoff joined 22:10 idemal joined
dalek r30743 | allison++ | pdd27mmd: 22:12
: [pdd27mmd] Shift 'bors', 'bands', and 'bxors' to real opcodes.
diff: www.parrotvm.org/svn/parrot/revision?rev=30743 22:13
r30744 | julianalbo++ | trunk: 22:15
: disable some debugging output in examples xlib.pir 22:16
diff: www.parrotvm.org/svn/parrot/revision?rev=30744
cotto_work moritz, the URL for 'imprint' is also incorrect 22:50
22:57 tetragon joined
s1n chromatic: you around? 23:08
chromatic Yes.
s1n i replied to RT#56712 23:10
chromatic Thanks!
s1n if you don't have access to a 64bit system, i can do the debugging for you if you can help me understand some of that code
i'm getting some strange memory problems with parrot from time to time but i've never been sure if it was just rakudo or parrot, i can work through those too if you want my help 23:11
chromatic I don't have access. Did you see the GC problem debugging guidelines on the Wiki?
s1n uhh no, link?
i don't really know how to read that C code for the PMC stuff and i do have experience with C heh 23:12
chromatic perlbuzz.com/project-hum/2007/11/ho...arrot.html
www.perlfoundation.org/parrot/index...ng_gc_bugs 23:13
s1n hmm ok, lots to read 23:14
i'll give this a try
chromatic The last one is the best. 23:15
s1n to the point, im going out for a run, i'll read it when i get back