Camelia, the Perl 6 bug

IRC log for #parrot, 2008-08-07

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:09 AndyA joined #parrot
00:10 Whiteknight joined #parrot
00:11 kid51 joined #parrot
00:19 AndyA joined #parrot
00:20 slavorg joined #parrot
01:05 petdance joined #parrot
01:30 dalek r30075 | jkeenan++ | scriptconfigure:
01:30 dalek : [configure] Refactor code from within
01:30 dalek : Parrot::Configure::Options::process_options() into more testable internal
01:30 dalek : subroutines _process_options_components() and _initial_pass().  Add tests for
01:30 dalek : these subs in t/configure/001-options.t.
01:30 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30075
01:33 Ademan joined #parrot
01:44 Whiteknight RT#57690. Anybody here an expert with headerizer?
02:08 dalek r30076 | jkeenan++ | scriptconfigure:
02:08 dalek : Begin to write tests for Parrot::Configure::Options::Conf::File.
02:08 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30076
02:08 kid51 Perhaps if you ping petdance ??
02:08 petdance what what what
02:08 petdance OH THE HEADERIZER
02:08 purl rumour has it the headerizer is making me happy
02:08 petdance MY BABY
02:08 kid51 Whiteknight requests assistance with headerizer
02:09 Whiteknight don't ask me, i dont know anything about it
02:10 Whiteknight I mean, I can pretend to help, commit a "fix" and then undo it in my next commit because it breaks everything
02:11 * kid51 pulls petdance's coat to http://rt.perl.org/rt3/Tic​ket/Display.html?id=57690
02:11 AndyA joined #parrot
02:13 davidfetter joined #parrot
02:22 petdance ok what am I looking at?
02:24 petdance Whiteknight: What are you looking for?
02:29 dalek r30077 | jkeenan++ | scriptconfigure:
02:29 dalek : Add more tests for internals of Parrot::Configure::Options::Conf::File.
02:29 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30077
02:31 nopaste "kid51" at 70.107.15.43 pasted "my results: make realclean; svn update; perl Configure.pl; make headerizer && make" (29 lines) at http://nopaste.snit.ch/13753
02:32 kid51 Whiteknight:  Does your failure resemble mine?
02:33 kid51 This appears to be simply a failure in 'make headerizer' independent of whether it's followed by 'make' or not.
02:33 kid51 Confirmed.
02:34 kid51 Of course, I don't really know what tools/build/headerizer.pl is supposed to be doing, having never attempted to use it before.
02:40 Theory joined #parrot
02:46 petdance It sounds like someone did something manually that the headerizer stomps on.
02:47 Whiteknight kid51, that's not the same error that I get
02:48 nopaste "Whiteknight" at 71.230.33.251 pasted "Build failure" (39 lines) at http://nopaste.snit.ch/13754
02:54 Whiteknight i've got to go though. goodnight
03:26 tetragon I now have a stacktrace for RT#57536, although my 32-bit build (which doesn't crash on the input the 64-bit one does) didn't hit the breakpoint my 64-bit build needed for that trace
03:32 justin joined #parrot
03:45 dalek r30078 | infinoid++ | pdd13pbc:
03:45 dalek : [pdd13] Apply API changes from pdd25cx merge to packfile PMCs.
03:45 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30078
03:49 dalek r30079 | infinoid++ | pdd13pbc:
03:49 dalek : [pdd13] Fix packfile PMCs to pass t/codingstd/c_parens.t.
03:49 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30079
03:56 tetragon Hrm... looks like Parrot_memalign is trying to shove 4294967296 into a function that takes what appears to be the difference between two pointers on a 64-bit binary
04:13 dalek r30080 | infinoid++ | pdd13pbc:
04:13 dalek : [pdd13] Merge changes from trunk (-30063:30079) into pdd13pbc branch.
04:13 dalek : This gets all the tests passing again.
04:13 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30080
04:14 tetragon joined #parrot
04:33 dalek r30081 | infinoid++ | trunk:
04:33 dalek : [docs/pdds/pdd13] Add the ability to delete a segment from a PackfileDirectory,
04:33 dalek : per my discussion with jonathan++ on this subject.
04:33 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30081
04:37 masak joined #parrot
04:40 tetragon Okay, so a few bits are not making it through an address calculation.  Thanks unsigned for the bogus address calculations, but I think a type with the correct width should be used.
04:41 cotto_home does anyone know how non-core types (i.e. not INTVAL, PMC*, etc) work as ATTRs in pdd17 PMCs, because it looks like they don't atm
04:48 Infinoid I honestly don't have a clue.  but isn't our general strategy to just wrap PMCs around any non-core types?
04:52 tetragon So, the shootout tests pass on OS X 10.5 64-bit Intel build once a stack of unsigned are changed to unsigned long in config/gen/platform/darwin/memalign.c
04:54 Infinoid ooh, nice
04:55 tetragon Before, they were giving rather scary looking crash reports
04:58 cotto_home src/pmc/exception.pmc makes it work
05:38 Psyche^ joined #parrot
05:39 masak re adding rules to rakudo: is it somehow an advantage that they already work in PGE?
06:02 UltraDM joined #parrot
06:37 slightlyoff_ joined #parrot
06:38 slightlyoff_ left #parrot
06:38 uniejo joined #parrot
06:50 masak joined #parrot
06:53 Ademan joined #parrot
07:09 cotto-work joined #parrot
07:18 barney joined #parrot
07:54 jan joined #parrot
08:22 cosimo joined #parrot
08:57 purl joined #parrot
09:02 dalek r30082 | kjs++ | trunk:
09:02 dalek : [pirc/new] more work on built-in constant folding.
09:02 dalek : Might need a bit refactoring.
09:02 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30082
09:15 kj joined #parrot
09:24 vhold joined #parrot
09:25 iblechbot joined #parrot
09:44 jonathan morning all
09:44 purl evening, jonathan
09:44 jan_ joined #parrot
09:44 jonathan purl: time fail!
09:44 purl jonathan: what?
09:50 Debolaz joined #parrot
09:58 kj morning jonathan
09:58 jonathan hi kj
10:32 * jonathan spots 57584 and sees that moritz beat him to tracking the issue down
10:36 dalek r30083 | kjs++ | trunk:
10:36 dalek : [pdd19] add description for binary relational operators in <var> = <var> <binop> <var> section.
10:36 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30083
10:46 dalek r30084 | julianalbo++ | trunk:
10:46 dalek : clean and headerize debug.c
10:46 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30084
11:00 dalek r30085 | julianalbo++ | trunk:
11:00 dalek : quick fix for a headerizer problem in include/parrot/debug.h
11:00 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30085
11:11 kid51 joined #parrot
11:25 dalek r30086 | jonathan++ | trunk:
11:25 dalek : [pct] The PAST spec says that using attribute scope should allow a child node to be supplied, which is used in place of self to supply the invocant, but it wasn't implemetned. This patch implements it.
11:25 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30086
11:25 dalek r30087 | jonathan++ | trunk:
11:25 dalek : [rakudo] Make use of attribute variables ($!foo) work in nested scopes, not just in the main method body itself.
11:25 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30087
11:59 dalek r30088 | jonathan++ | trunk:
11:59 dalek : [rakudo] Add a (currently wrong) implementation of time(). Sent a couple of mails to try and ascertain how we can make this More Correct; in the meantime, this allows you to do relative times OK (and will make my YAPC code examples run ;-)).
11:59 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30088
12:00 masak hm, spectest_regression passes 4/6 subtests in S02-polymorphic_types/subset-range, but tools/test_summary.pl passes 6/6. what gives?
12:00 jonathan Oddness. Not sure.
12:04 dalek r30089 | coke++ | trunk:
12:04 dalek : [docs] Fix grammar issue.
12:04 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30089
12:04 masak jonathan: another thing entirely: what I miss most these days with rakudo is s///. but I guess this is quite a big field which ties into rules and stuff. still, is there anything I can help with to speed it up?
12:05 ambs joined #parrot
12:05 jonathan Substitutions? Hmm.
12:06 masak right now I'm feeling the pain from having to reinvent them using string-stitching
12:06 masak ...which is fun, in a masochist kind of way
12:07 masak I just feel I could use that time to hack on rakudo to help make it work instead
12:08 jonathan Yeah
12:09 masak re the test thing: I think a Parrot segfault is behind it all. I'll submit a bug report
12:15 jonathan OK, thanks.
12:17 jonathan I'm staring at STD.pm's parsing of s/// and not quite getting it. :-)
12:17 masak :)
12:18 jonathan The .subst method on a string may just be easier.
12:20 masak even that would be a welcome addition
12:21 jonathan That may not be so hard to implement.
12:22 jonathan I think we don't have the :g modifier yet, so it's just grab the bit of the string before the match, grab the bit of the string after the match, and stick in the replacement.
12:22 jonathan Checking whether the replacement is Code or not and if so running it.
12:22 masak this is good. I'll have to write this down. :)
12:27 jonathan masak: In code? ;-)
12:27 masak jonathan: if you don't beat me to it, yes. :)
12:28 masak I'm tied to a normal working day today, but there's a night and a morning coming up after that
12:28 japhb joined #parrot
12:28 jonathan masak: If you plan to work on it, I'll leave it. :-)
12:28 masak jonathan: I'll give it a shot and get back to you in 24h, either stuck or done. :)
12:29 jonathan Cool!
12:30 masak ...yes, potentially
12:30 masak is it rakudo day today?
12:33 Whiteknight joined #parrot
12:35 jonathan masak: Yes.
12:35 jonathan Been checking my examples I plan to show at YAPC::EU and done a few things to get them running.
12:35 jonathan Just investigating some RT tickets at the moment.
12:35 masak nice.
12:35 jonathan And will make a branch and hack on lazy lists later on today.
12:36 masak awesome.
12:36 masak when's list assignment due?
12:36 * jonathan points at pm ;-)
12:36 masak haha
12:36 jonathan Hopefully with the hacking time available at YAPC::EU, he may get to it then. :-)
12:38 masak yes.
12:39 Ademan joined #parrot
12:39 masak did I mention that we'll be showing up a day early to allow for hackathon time?
12:39 * ambs is still thinking if he will take the macbook pro or the acer aspire one. This later is not that good to help on a hackathon.
12:39 jonathan I think you mentioned, yes. :-)
12:39 jonathan I don't know where the hackathon is, though!
12:39 jonathan I know I'm coming over to it form Malmo in the morning...
12:40 skv joined #parrot
12:41 masak there's plenty of time left for a location to reveal itself :)
12:48 jonathan masak: You filed #57398
12:48 jonathan But I can't see the difference between the two code examples.
12:48 jonathan sub a($x) { $x = 5 }; my $y = 7; a($y);
12:48 jonathan sub a($x) { $x = 5 }; my $y = 7; a($y);
12:49 masak hold on
12:49 masak checking
12:51 masak yes, they're the same, unintentionally -- sometimes pasting goes awry
12:51 masak I'll try to replicate what I did
12:51 masak I think it has something to do with is rw
12:53 masak ah -- I think the second one should be ./perl6 -e 'sub a($x) { $x = 5 }; a(7)'
12:53 masak I'll amend that to the bug
12:59 masak there.
12:59 masak also, I'm not sure that the name of the bug describes the bug very well.
13:04 petdance joined #parrot
13:08 gryphon__ joined #parrot
13:09 jonathan OOh, I forgot to have lunch!
13:09 jonathan brb
13:09 masak "you know you're having a productive day when..."
13:12 iblechbot joined #parrot
13:30 dalek r30090 | jonathan++ | trunk:
13:30 dalek : [rakudo] Initialize attributes with @ and % sigils correctly. Fixes RT#57422.
13:30 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30090
13:31 jonathan masak: BTW, I know you're from somewhere in the Nordics, since you were at NPW, but which country?
13:43 dngor joined #parrot
13:44 workbench joined #parrot
13:46 masak jonathan: I live one city away from where the NPW was. Uppsala.
13:48 jonathan Aha, OK.
13:48 jonathan I'm flying into Stavska tomorrow, and then taking the bus to Linkoping and having Saturday there as vacation. :-)
13:49 jonathan Then in Malmo on Sun/Mon having some break too, but also meeting some friends there.
13:50 masak jonathan: Sweden is very moist right now.
13:58 jonathan masak: As in, rain?
13:58 jonathan Or humidity?
13:58 purl humidity is just a touch to high on average.
13:59 masak jonathan: rain.
14:00 masak occasionally large amounts of it.
14:01 jonathan Ugh.
14:02 jonathan But forecast says Saturday is sun with scattered clouds. :-)
14:02 masak ack -- and the rest rain :)
14:07 Whiteknight purl forget humidity
14:07 purl Whiteknight: I forgot humidity
14:09 Whiteknight purl, humidity is the worst
14:09 purl OK, Whiteknight.
14:11 Infinoid masak: I'll trade you!
14:11 Infinoid clear blue skies get boring after a while.
14:11 masak Infinoid: careful what you wish for.
14:12 masak but yes, I'm not exactly miscontented
14:12 masak this is good indoor weather
14:13 Infinoid http://www.wunderground.com/​US/CA/South_Lake_Tahoe.html
14:13 Infinoid nooo, they's stealin my clouds
14:22 jonathan Some things that look easy to fix, so aren't...
14:25 masak jonathan: I can tell you're somewhere in the "difficult tasks [should be] possible" domain, then
14:26 jonathan masak: I'm in the, make (1+1).WHAT return Int without breaking a bunch of spectests domain. :-)
14:27 jonathan Current run through the spec tests has made it to the OO ones and is looking OK so far.
14:27 jonathan If it all passes, I can commit and close two RT tickets. :-)
14:27 gaz that's a rather specific domain :)
14:28 dalek r30091 | julianalbo++ | trunk:
14:28 dalek : fixed perl6multisub for c++ build
14:28 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30091
14:29 * jonathan wonders how he broke the C++ build...
14:31 Andy joined #parrot
14:35 jonathan Argh! It's *so* annoying when you break the _last_ spectest, then need to run the lot again to make sure the fix for that one didn't break anything elsewhere. :-)
14:36 jonathan masak: You're filing a lot of RT tickets that look very similar-ish (Rakudo gives an error message that it should but subsequently we get a segfault)
14:36 jonathan Are all of them double-frees?
14:36 masak jonathan: yes.
14:36 jonathan OK. I can't reproduce these on Win32, but I think that's because my memory library doesn't blow up if you double-free.
14:36 jonathan :-|
14:37 jonathan Do you know how to run it in gdb and get a backtrace?
14:37 masak jonathan: no, but feel free to give me some hints
14:37 masak I know I have gdb installed.
14:37 jonathan Hmm
14:37 jonathan OK, I think from the languages/perl6 you'd do
14:38 jonathan gdb ../../parrot
14:38 jonathan And then once it's started type
14:38 jonathan r perl6.pbc name_of_file.p6
14:38 jonathan or I guess you can do -e '...' too
14:38 masak ah
14:39 masak sounds easy -- I'll try immediately
14:41 masak I did the former -- no double free, no stacktrace
14:41 masak :/
14:42 jonathan Oh. :-S
14:42 jonathan If we can't trap it in the debugger, that may make it...harder...to find. :-(
14:43 masak I figured as much
14:44 masak I could try it for my other, similar bugs too
14:44 jonathan If you get chance, yes...a stack trace would help enormously.
14:45 masak I'll see what I can do.
14:45 jonathan Hurrah, fixed the bugs without breaking the spec tets.
14:46 masak \o/
14:50 dalek r30092 | jonathan++ | trunk:
14:50 dalek : [rakudo] Make various arithmetic operations return an Int when they are performed on Ints. However, to get the right semantics, we must also add some logic to do upgrades when needed, which is also in this patch. Resolves RT#56944 and RT##56748.
14:50 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30092
14:55 jonathan That unskips another spectest too. :-)
14:56 masak jonathan++
15:25 dalek r30093 | jonathan++ | trunk:
15:25 dalek : [rakudo] Fix a regression in the handles trait verb.
15:25 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30093
15:38 dalek r30094 | jonathan++ | trunk:
15:38 dalek : [rakudo] Fix bug in does operator, where it failed to work if you tried to apply multiple mixins to an object over time.
15:38 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30094
16:04 dalek r30095 | jonathan++ | trunk:
16:04 dalek : [rakudo] Add S12-attributes/delegation.t to spectest_regression, so our handles trait verb implementation gets some testing.
16:04 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30095
16:22 Andy joined #parrot
16:27 timbunce joined #parrot
16:39 contingencyplan joined #parrot
16:42 Theory joined #parrot
16:44 ruoso joined #parrot
16:53 dalek r30096 | jonathan++ | trunk:
16:53 dalek : [rakudo] Add S02-builtin_data_types/hash.t to spectest_regression - another 35 passing tests.
16:53 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30096
17:06 tewk joined #parrot
17:15 rurban joined #parrot
17:16 cotto-work kj, ping
17:25 dalek r30097 | kjs++ | trunk:
17:25 dalek : [pirc/new] implement optimization for x += 1, so that this uses the inc instruction instead of add. Same for dec/sub.
17:25 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30097
17:26 kj cotto-work: poing
17:27 cotto-work if I wanted to help out with pirc, what'd be a good way to get familiar with the code while making myself useful?
17:27 kj aha! helpers :-)
17:27 kj ehm
17:27 kj let me think
17:27 kj be back in 2 min. (thinking)
17:27 dalek r30098 | kjs++ | trunk:
17:27 dalek : [pirc/new] allow for x += x. Forgotten in prev. commit.
17:27 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30098
17:27 cotto-work take your time.
17:28 cotto-work I can't write any easily-contributable code at work anyway. ;)
17:28 rurban pirc/README.pod eg
17:30 kj well. One Q from me would be: do you know bison and flex?
17:30 kj that's kindof important :-)
17:30 cotto-work no.  How hard are they to get into if I already know C?
17:31 kj not so hard, I guess.
17:31 kj there's 2 things to it:
17:31 rurban harder than C. see the both .l and .y files
17:31 kj writing/maintaing the grammar
17:31 kj maintaining
17:31 purl maintaining is for mere maintainance programmers. lathos is a pioneer! An inventor! A pathfinder! A shithead.
17:31 cotto-work yes
17:31 kj and the 2nd thing is the nitty-gritty of setting up a parser using bison/flex
17:31 kj the good thing is
17:31 kj I've done both already :-)
17:32 kj that means, writing a parser that is reentrant, that's a bit tricky, but doable
17:32 cotto-work that's a big plus
17:32 kj maintaining the grammar can be difficult for starters, because you'll encounter weird messages such as 'shift/reduce' annd 'reduce/reduce' conflicts
17:33 kj so you need to understand what these mean
17:33 cotto-work ok.
17:33 kj I remember when I started working on pirc/new (in the 'new' directory, not the recursive-descent version), I learned a lot from the documentation on bison
17:33 kj making it re-entrant
17:34 kj So, learning bison is best done through the online documentation
17:34 kj there's extensive material on that
17:34 cotto-work good to know
17:34 kj pirc's backend should be refactored though
17:35 kj so there's more work to do there, but the trouble now is, I don't know exactly how to generate PBC. It might not be as easy as it seems, because one should consider the :immediate and :init flags and stuff like that
17:36 kj so what I'm doing right now is generate almost-pure-pasm
17:36 rurban checking for builtin op's e.g is a major todo
17:37 kj that's not so difficult, in theory
17:37 kj in practice, well, that's a matter of linking the right thing, but I haven't succeeded in that yet
17:37 kj but yes, that'd be nice to have
17:37 kj because now there's only 5 instructions or so that you can use
17:37 rurban get the libs into the right order?
17:37 kj yeah, well, IMCC uses a function is_op() or whatever
17:38 kj and it pokes into the guts of parrot
17:38 kj checking whether the string is an op
17:38 kj don't know how that's done. It confused me
17:38 rurban Parrot_is_builtin() it is
17:38 kj is that it? strange name.
17:39 rurban problem is also that the current ldflags, libs order ... is broken
17:39 purl okay, rurban.
17:39 kj ehm, when doing what?
17:39 rurban linking with stricter compilers, linking to installables
17:39 cotto-work no, problem is <reply>
17:39 purl okay, cotto-work.
17:39 kj ok, but at this point pirc has no problem
17:39 rurban linking with already installed libparrot shared
17:40 Whiteknight I want to get in on helping with PIRC too, when I have the tuits to spare
17:40 kj woohoo! more help :-)
17:40 kj well, the best thing I can say is, just check out the code
17:40 particle problem is foo
17:40 rurban I'm working a largish linking patch right now. #39742
17:41 particle problem?
17:41 rurban me?
17:41 purl i think rurban is on Win XP sp2
17:41 particle ah. that's nice.
17:41 rurban now; I'm on cygwin
17:41 rurban linking is almost solved now.
17:41 cotto-work kj, thanks.  I'll bone up on those topics (reentrancy, yacc, bison and pbc) and go from there
17:42 rurban have to test on more stricter linkers now.
17:42 Whiteknight me?
17:42 purl well, Whiteknight is updating the book. It talks about .pragma fastcall, pushing and popping arguments onto the user stack, etc
17:42 particle rurban: try solaris.
17:42 kj cotto-work: you're very welcome
17:42 Whiteknight yeah, that's a little outdated....
17:42 rurban good.
17:42 rurban and maybe intel's icc
17:43 cotto-work icc is fairly easy to get working on Linux
17:43 rurban yeah. I have it home, but right now I'm in a hotel in germany
17:44 rurban solaris also later
17:46 rurban ok, have to leave.
17:54 pmichaud Aloha
17:55 particle hi-o
17:58 jonathan hi
17:58 purl bonjour, jonathan.
18:01 jonathan pmichaud: Didn't make it to the lazy lists today yet...been doing lots of other little fixes, adding more tests, and so forth.
18:01 jonathan Got all of the code examples in my YAPC Perl 6 OO talk running. :-)
18:01 pmichaud jonathan: no problem.  We can work on it at yapc::eu
18:01 jonathan And added tests to cover the things that didn't work.
18:01 jonathan I may start hacking on it tonight.
18:02 pmichaud afaict it's not as pressing an issue as, say, list assignment or signatures
18:02 jonathan Right now I'm trying to get hash composers { ... } to work.
18:02 jonathan Since they seem like a wanted, popular thing.
18:02 pmichaud oh, hash composers should just analyze the PAST tree for a block
18:02 jonathan Yeah
18:02 pmichaud I think it belongs in circumfix:<{ }>
18:02 jonathan my $x = { x => 42 }; $x.WHAT # Hash :-)
18:03 jonathan Yes, it's in circumfix:<{ }>
18:03 jonathan Amusingly, I then started failing things in the text suite like this:dies_ok { test2() }, "return() inside a try{}-block should cause following exceptions to really die";
18:04 pmichaud it detected the {} as a hash composer?
18:04 jonathan Because it saw the {} in there, it's an empty block so it winds up being a hash constructor as per the spec, and then the quote_term rule was just assuming any circumfix was a block.
18:04 pmichaud heh
18:04 pmichaud oh, in the double quotes
18:04 jonathan Yes
18:04 pmichaud yes, I've had that happen a few times
18:04 jonathan The bit of the spec says: {...} is always a block. However, if it is completely empty or consists of a single list, the first element of which is either a hash or a pair, it is executed immediately to compose a Hash object.
18:05 pmichaud right.  The test (message) is wrong there. :-)
18:05 jonathan OK, good. But we shouldn't be bombing out, right?
18:05 jonathan It's like trying to interpolate a hash?
18:05 pmichaud it should probably result in an empty string
18:05 jonathan I think so too.
18:06 pmichaud (yes, interpolate a hash)
18:07 jonathan Well, waiting to see how spectest_regression works out with the tweak in quote_term. :-)
18:07 pmichaud ...quote_term was tweaked..?
18:07 jonathan It had
18:07 jonathan elsif ($key eq 'circumfix') {
18:07 cotto-work seen chromatic
18:07 purl chromatic was last seen on #parrot 1 days, 15 hours, 31 minutes and 52 seconds ago, saying: JITted thunks?  [Aug  6 02:35:52 2008]
18:07 jonathan Followed by
18:07 jonathan $past.blocktype('immediate');
18:07 cotto-work seen allison
18:07 purl allison was last seen on #parrot 36 days, 20 hours, 19 minutes and 24 seconds ago, saying: whiteknight: sounds good  [Jul  1 21:48:28 2008]
18:08 pmichaud oh, you mean in actions.pm
18:08 jonathan If you've got a hash composer, that didn't work out.
18:08 jonathan Yes.
18:08 jonathan Sorry, should have been more clear I meant the action, not the rule.
18:08 pmichaud np
18:08 pmichaud time for breakfast here (and we're on a tight time schedule)
18:08 pmichaud bbl
18:08 pmichaud (maybe)
18:08 jonathan ok
18:09 jonathan enjoy
18:16 dalek r30099 | kjs++ | trunk:
18:16 dalek : [pirc/new] comments, refactoring, removal of .pragma n_operators
18:16 dalek : + add a temp. fix for segfaulting when doing keyaccess (set x, x[1])
18:16 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30099
18:25 dalek r30100 | kjs++ | trunk:
18:25 dalek : [pirc] update README a bit.
18:25 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30100
18:36 jonathan OK, going for dinner...hash thingy patch nearly there, just one more test fail to hunt down...
18:44 dalek r30101 | fperrad++ | trunk:
18:44 dalek : [Lua] libraries
18:44 dalek : - major refactor : now lua_register does the job
18:44 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30101
19:35 dalek r30102 | fperrad++ | opengl4lua:
19:35 dalek : [Lua] merge from trunk (29871 -> 30101)
19:35 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30102
19:35 cjfields joined #parrot
19:48 cotto-work is someone bringing up Parrot HLL interop at YAPC::EU?
19:51 cjfields seen jonathan?
19:51 purl jonathan was last seen on #parrot 1 hours, 15 minutes and 10 seconds ago, saying: OK, going for dinner...hash thingy patch nearly there, just one more test fail to hunt down...
19:56 jonathan cjfields: I'm here, just back from dinner.
19:57 jonathan cotto-work: There'll be a load of things discussed at YAPC::EU, I'm sure. And I'd be surprised if that wasn't one of them.
19:57 cotto-work ok.  I look forward to hearing what comes out of the discussion.  It's an interesting problem.
19:59 cjfields jonathan: backlogged, saw a mention of .subst
19:59 jonathan cjfields: Yes. Are you actively working on this?
20:00 cjfields no, $job in the way
20:00 jonathan OK.
20:00 cjfields but it might be best to get .match working prior
20:00 jonathan Phew, I feared I'd suggested someone else look at something that you were already on with. :-)
20:01 jonathan I think that, because subst just returns a string, which just gives itself in item context, we won't run into the same issue.
20:01 cjfields agreed.
20:01 jonathan So I don't think one blocks the other. I haven't forgotten the match issue though.
20:02 jonathan Unfortunately, I haven't thought of a good solution either. :-(
20:02 jonathan And didn't see one posted. But various people who might have ideas/answers will all be at YAPC::EU, so I'm hoping we can get a resolution there.
20:02 cjfields I think this may be dealt with when  .HLL is addressed
20:03 jonathan I don't see it as a HLL issue. At least, not that bit.
20:03 jonathan We're returning a match object.
20:03 jonathan The problem is when we do an assignment to a scalar, it causes item context on the RHS of the =.
20:04 jonathan And a match object in item context just gives back its string bit.
20:04 cjfields My bad (was reading the RT ticket and saw .HLL mapping from Pm)
20:04 jonathan Whereas what we actually want is to assign the match object itself.
20:04 jonathan I think that might come into it too.
20:04 jonathan Like, making sure the strings we get out of the match object are Perl6Str.
20:05 jonathan But that's not the specific problem that I think is currently blocking match from working.
20:05 cjfields he mentioned this would be taken care of after mapping
20:05 jonathan OK
20:05 jonathan Hopefully that's not so far off.
20:06 cjfields I know that mapping fixes a LOT of issues (I think .split returns a String instead of a Perl6Str).
20:06 jonathan Yes.
20:06 jonathan It'll fix that one.
20:10 cjfields I think we decided to wait on a .match commit (RT#56970), correct?
20:10 dalek r30103 | fperrad++ | opengl4lua:
20:10 dalek : [Lua] merge\n refactor with lua_register
20:10 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30103
20:10 cjfields There was some tweaking to .trans I wanted to work on to get it closer to S05, but it requires .match.
20:13 jonathan Yes, let's see if there's a solution to this issue
20:13 jonathan I'm hopeful we'll have one within the next week.
20:13 cjfields OK
20:24 Whiteknight joined #parrot
20:36 cognominal joined #parrot
20:37 dalek r30104 | jonathan++ | trunk:
20:37 dalek : [rakudo] Make an empty block, a block containing just a pair or a block containing a list that starts with a pair or a hash construct a Hash, as per S06. Resolves RT##57340.
20:37 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30104
21:02 dalek r30105 | jonathan++ | trunk:
21:02 dalek : [rakudo] Make classes declared as Foo::Bar work (that is, more than one level deep namespace). Resolves RT#57388, and makes another 5 spectests pass.
21:02 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30105
21:23 cotto-work lart?
21:23 purl lart is LUSER ATTITUDE RECALIBRATION TOOL or In the collective mythos of scary devil monastery, this is an essential item in the toolkit of every BOFH. The LART classic is a 2x4 or other large billet of wood usable as a club, to be applied upside the head of spammers and other people who cause sysadmins more grief than just naturally goes with the job.  or http://pics.livejournal.co​m/elevenwords/pic/0011ebtz
21:25 Whiteknight haha, that picture is priceless
21:25 Whiteknight purl++
21:28 gryphon__ joined #parrot
21:29 jonathan ouch!
21:29 * jonathan hopes not to ever meet someone who owns one.
21:37 dalek r30106 | jonathan++ | trunk:
21:37 dalek : [rakudo] Remove completed task from the ROADMAP.
21:37 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30106
22:18 dalek r30107 | Whiteknight++ | gsoc_pdd09:
22:18 dalek : [gsoc_pdd09] update to trunk r30105
22:18 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30107
22:23 dalek r30108 | jonathan++ | trunk:
22:23 dalek : [rakudo] Implement unlink function.
22:23 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30108
22:26 dalek r30109 | jonathan++ | trunk:
22:26 dalek : [rakudo] Add a (now completely passing) IO spectest to spectest_regression target.
22:26 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30109
22:26 kid51 joined #parrot
22:29 * jonathan decides to call it a day
22:32 Infinoid day: you're a day
22:37 jonathan :-P
23:02 dalek r30110 | chromatic++ | trunk:
23:02 dalek : [config] Improved test parallelization by appending PID to all generated
23:02 dalek : test.c, test.cco, test.ldo, and test exectuable files.  This is a good target
23:02 dalek : for further refactoring:
23:02 dalek :     - there's too much duplication of generated file names
23:02 dalek :     - temporary names from File::Temp would be better
23:02 dalek :     - building all files in a temporary directory would be even better
23:02 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30110
23:03 * particle notes that similar treatment would benefit future config parallelization as well
23:19 cotto-work looks like there
23:19 purl looks like there is, like, no difference
23:20 cotto-work looks like there's a typo in that latest commit
23:20 cotto-work open( my $out_fh, '>', "test_$$.c" ) or die "Unable to open 'test_$$.cb: $@\n";
23:20 cotto-work no, looks like there is <reply>
23:20 purl okay, cotto-work.
23:21 cotto-work config/inter/progs.pm:174
23:22 jonathan BTW, will write Rakudo day report tomorrow, probably on the plane...currently trying to get stuff sorted out for travelling.
23:30 Tene purl: forget looks like there
23:30 purl Tene: I forgot looks like there
23:30 Tene purl: looks like there is <reply>
23:30 purl OK, Tene.
23:34 tetragon joined #parrot
23:39 ruoso joined #parrot
23:40 dalek r30111 | jkeenan++ | scriptconfigure:
23:40 dalek : P::C::O::Conf::File:  Eliminate some unreachable branches.  t/configure/062:  Add tests for previously uncovered branches.
23:40 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30111
23:46 tetragon Hrm... the patch in RT#57700 is for the same issue as RT#57536
23:54 cognominal joined #parrot
23:55 kid51 tetragon:  Perhaps so.  I've got both of them listed in the metaticket for Mac:  http://rt.perl.org/rt3/Tic​ket/Display.html?id=52904
23:55 jonathan cognominal: Coming to YAPC?
23:56 bacek joined #parrot
23:59 cognominal yes

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

Parrot | source cross referenced