Camelia, the Perl 6 bug

IRC log for #parrot, 2008-12-08

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:05 Alias_ joined #parrot
00:14 Whiteknight allison++
00:18 tak joined #parrot
00:23 Coke we going to sneak it in before release?
00:24 * Coke tests out tcl on the branch.
00:26 kid51 joined #parrot
00:29 baest_ joined #parrot
00:35 nopaste joined #parrot
00:41 Whiteknight we've got more then a week before the release, I'm fine with merging it in if it happens soonish
00:41 Whiteknight insofar as my fineness with it is a limiting factor
00:42 stockwellb joined #parrot
00:46 dalek r33639 | allison++ | pdd22io_part2:
00:46 dalek : [pdd22io] Clean up several more instances of values modified but not set
00:46 dalek : in the FileHandle object.
00:46 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33639
00:47 Tene allison: can you explain what I need to do for the pdd24 review again?
00:47 Tene Just check for sanity, NYI, and other differences between spec and current reality?
00:47 allison Tene: read through it, and check for reality
00:47 allison Tene: yes
00:48 Tene Okay.  Thanks.  I'll have a writeup soon.
00:48 Whiteknight allison, ETA on the IO branch? soonish?
00:48 allison cool, thanks!
00:48 Tene There are also some parts in there that still need to be specced.
00:48 allison Whiteknight: just sent an email to parrot-dev
00:48 allison Whiteknight: as soon as I get some language/platform reports back
00:48 allison Whiteknight: hopefully tomorrow
00:49 Coke allison: tcl go boom.
00:49 Tene Also, is there a detailed list anywhere for "exceptions, debugging, exception types, resumable exceptions"?  I think this is mostly done, unless there are specific bugs that need debugging, and depending on what "Exception types" is asking for.
00:49 allison Coke: how boom?
00:49 Coke [01:48:59] t/cmd_apply....................Method 'slurp' not found for invocant of class 'FileHandle'
00:50 allison Coke: change call to 'slurp' to 'readall'
00:50 purl allison: that doesn't look right
00:52 Coke [01:52:09] t/cmd_catch....................Cannot readall on a new file from an already open filehandle
00:52 Whiteknight if the branch lands tomorrow, will you lend me a hand figuring out the calling_convention nonsense before the release?
00:52 * Coke is digging.
00:53 allison Whiteknight: yes, will do that next
00:53 allison Whiteknight: has the calling_conventions branch been recently updated from trunk?
00:54 * Tene drive home.
00:54 allison Coke: that means you first opened the file handle (with a filename), and then tried to call 'readall' passing it another file name
00:55 allison Coke: you can close the filehandle and then call 'readall', or just call 'readall' without any parameters (to use the existing open filehandle)
00:55 Coke which is how slurp used to work. =-)
00:55 * Coke cleans up.
00:57 allison Coke: yeah, slurp was a little inconsistent, and didn't have much error checking
00:58 allison (the 'slurp' method, that is)
00:58 Coke allison: is open 'foo', '<' still allowed?
00:58 Coke or do I need to use 'r' ?
00:58 allison there seems to be something amiss with my checkout of the Rakudo spectests
00:58 allison Coke: I left the old flags working for now
00:59 allison Coke: but they will be deprecated
00:59 allison Coke: so, doesn't hurt to change to 'r'
01:00 allison Coke: but make sure not to write to a filehandle opened read-only (the old I/O system allowed this, for some strange reason)
01:02 * allison back in 20 min or so
01:07 stockwellb joined #parrot
01:10 Coke allison: I had had to do slurp('') before. (or at least, I did it. I presume I had to at some point). changing to readall() did it.
01:11 Coke (mostly)
01:17 Whiteknight i hate to keep griping about this branch, but i just can't get my mind wrapped around it
01:19 Coke which branch?
01:19 purl which branch is that you want to merge?
01:19 Coke purl, forget which branch.
01:19 purl Coke: I forgot which branch
01:19 ChrisDavaz joined #parrot
01:19 TiMBuS did something happen recently with the exception handlers in parrot? my language suddenly broke =/
01:19 Coke since the last release, I think so. Is your language in svn?
01:20 TiMBuS no i didnt add it
01:20 TiMBuS i have code confidence issues ;[
01:21 TiMBuS but seriously uh, what changed? my programs now go into an infinite loop and eat ram
01:22 Coke I believe it's that exceptions are no longer automatically disabled when caught.
01:22 Whiteknight Coke: the calling_conventions branch.
01:22 purl the calling_conventions branch is very conservative, we've broken the changes down into small stages
01:23 Whiteknight it was working beautifully, then I waited a few days until after the last release, and when I tried to merge again it was broken
01:23 Coke TiMBuS: if you have a small sample that shows the behavior, someone can probably give youa  fix.
01:23 Whiteknight and by that time, so many other branches had merged that it was impossible to track down where the problems were from
01:23 TiMBuS ill see what i can cut it down to
01:23 Coke Whiteknight: so do what pmichaud does. create a new branch from trunk *NOW* and merge to that.
01:24 Coke if it's just a question of merging, i can probably do that.
01:25 bacek joined #parrot
01:28 TiMBuS oh, i accidentally fixed it trying to pin it down.
01:29 TiMBuS i was using 'die' inside an exception handler, now i use rethrow. seems to fix it.
01:29 Whiteknight Coke, I tried that. Created a new branch, tried to merge it, and got more errors then when I started
01:29 Whiteknight it was a total mess
01:41 dalek r33640 | jhorwitz++ | trunk:
01:41 dalek : [NCI] add iVppp signature for mod_parrot
01:41 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33640
01:42 Coke Whiteknight: I'll take a look if time permits.
01:43 Coke (but not just now.)
01:43 Whiteknight don't worry, we're not in a hurry anymore
01:43 Whiteknight I'd be happy for any help at any time
01:44 Whiteknight okay, i'm off to bed now. Goodnight!
01:57 silug joined #parrot
02:27 allison Coke: ah, the old slurp(”) is now readall() with no parameters
02:28 allison Coke: I just couldn't see parsing a string to see if it was empty when we have optional parameters now
02:34 pmichaud seen tewk
02:34 purl tewk was last seen on #parrot 22 hours, 11 minutes and 50 seconds ago, saying: Doesn't it build unstripped by default
02:35 pmichaud tewk: ping
02:45 nopaste "pmichaud" at 72.181.176.220 pasted ":subid fails with :multi (for tewk)" (23 lines) at http://nopaste.snit.ch/14848
02:55 dalek r33641 | jkeenan++ | testparrottest:
02:55 dalek : File used in testing case where file extension is missing.
02:55 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33641
03:16 adu joined #parrot
03:23 ChrisDavaz joined #parrot
03:40 jimmy joined #parrot
03:47 dalek r33642 | jkeenan++ | testparrottest:
03:47 dalek : Remove commented-out debugging code.
03:47 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33642
03:47 dalek r33643 | pmichaud++ | assign:
03:47 dalek : [rakudo]:  Initial assignment refactors.
03:47 dalek : * set infix:= to be a :multi.
03:47 dalek : * clean up array assignment.
03:47 dalek : * clean up lots of list/array methods.
03:47 dalek : * move Protoobject methods into a separate .pir file
03:47 dalek : * reorder class methods for some classes
03:47 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33643
03:48 TonyC joined #parrot
03:51 * kid51 must sleep
03:51 purl $kid51->sleep(8 * 3600);
03:54 Andy joined #parrot
04:08 dalek r33644 | allison++ | pdd22io_part2:
04:08 dalek : [pdd22io] Bringing the pdd22io_part2 branch up-to-date with trunk r33643.
04:08 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33644
04:29 tewk pmichaud: pong
05:06 allison hmmm t/pmc/key.t #7 is failing now in the I/O branch, but only on Ubuntu
05:15 pmichaud tewk:  left you a nopaste with a failing :subid case
05:22 allison doesn't appear to be related to I/O, but isn't happening on trunk
05:38 dalek r33645 | pmichaud++ | assign:
05:38 dalek : [rakudo]:  initial slice implementation -- still more to do
05:38 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33645
05:52 ChrisDavaz joined #parrot
05:56 cotto allison, make realclean fixes it for me
06:00 allison cotto: mmm... I tried realclean once, will try again
06:09 GeJ All tests successful, 14 tests and 618 subtests skipped.  (on FreeBSD 7, freshly co'ed pdd22io_part2 branch)
06:09 GeJ testing rakudo now.
06:11 Theory joined #parrot
06:31 ChrisDavaz joined #parrot
06:37 dalek r33646 | pmichaud++ | assign:
06:37 dalek : [rakudo]:  First cut at list assignment.
06:37 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33646
06:37 pmichaud time for sleep -- will finish up the assign branch tomorrow morning.
06:38 dalek r33647 | allison++ | calling_conventions:
06:38 dalek : [calling_conventions] Bringing the calling_conventions branch up-to-date
06:38 dalek : with trunk r33644.
06:38 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33647
06:43 Hadi joined #parrot
06:45 Theory joined #parrot
06:48 Alias joined #parrot
06:55 dalek r33648 | moritz++ | trunk:
06:55 dalek : [rakudo] add a file to spectest.data which the lex refactor made pass
06:55 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33648
06:57 Alias \o
06:58 * Alias succesfully create Vanilla Perl 6 alpha, printed it onto CD, and presented it to TimToady at the OSDC conference :)
06:58 Alias created
06:58 moritz Alias++
06:58 Alias Of course, it leaks like a motherfucker :)
06:59 Alias Any flags I should pass to Configure.pl to start stripping out development/debugging stuff?
06:59 Alias I want to slowly start polishing it into something a bit more real than one dll and a .exe
07:00 moritz Alias: (unrelated) http://www.perlmonks.org/?node_id=728739
07:00 moritz afk
07:01 Alias Thanks for the link
07:06 Hadi left #parrot
07:06 Alias moritz: http://www.perlmonks.org/?node_id=728869 :)
07:16 uniejo joined #parrot
07:33 stockwellb joined #parrot
07:46 Hadi joined #parrot
07:46 Hadi left #parrot
07:52 Theory joined #parrot
08:55 ChrisDavaz joined #parrot
09:05 Zaba_ joined #parrot
09:11 Zaba joined #parrot
09:20 particle joined #parrot
09:21 tomyan joined #parrot
09:29 iblechbot joined #parrot
09:34 bacek joined #parrot
09:47 Theory joined #parrot
10:05 alvar joined #parrot
10:22 register joined #parrot
10:31 Theory joined #parrot
10:40 alvar joined #parrot
10:52 kj joined #parrot
10:55 gaz joined #parrot
10:58 Theory joined #parrot
11:13 mj41 joined #parrot
11:24 riffraff joined #parrot
11:28 riffraff hi
11:28 purl privet, riffraff.
11:29 riffraff privet?
11:29 purl privet is the bane of Shoalhaven or an aggressive environmental weed requiring control to halt its spread and prevent further degradation
11:29 riffraff I don't think so
11:30 riffraff privet is also Ligustrum Vulgare, often grown as a hedge
11:30 purl okay, riffraff.
11:33 jonathan It's also Russian for "hi" :)
11:46 riffraff ah, now everything clicked, thanks you a lot :)
11:47 riffraff s/thanks/thank
11:48 jonathan Pazhaloosta. ;-)
11:59 bacek jonathan: 'Pozhalujsta" little bit close to russian pronunciation :)
12:00 bacek (hi everybody)
12:01 nopaste "bacek" at 87.250.243.86 pasted "pmichaud, is it proper implementation of Match::ACCEPTS ?" (18 lines) at http://nopaste.snit.ch/14857
12:02 bacek pmichaud: (it passed my simple testcase from #61136)
12:02 jonathan bacek: Transliteration is a rather inprecise art. :-)
12:03 * jonathan much prefers to just see the stuff in cyrillic, alas his IRC setup fails at unicode :-(
12:03 bacek jonathan: indeed :)
12:03 bacek jonathan: what is your IRC client?
12:03 jonathan Well, the client is irssi running on feather
12:04 jonathan It's not the IRC client itself that's the problem.
12:04 jonathan I suspect it's something in my SSH client/terminal setup/something else.
12:04 bacek jonathan: probably 'export LANG=en_US.UTF8' will help
12:04 jonathan Where should I stick that?
12:05 jonathan .bash_rc?
12:05 bacek hm... ~/.bashrc
12:05 bacek or you can just invoke it from shell before running irssi
12:06 bacek (.bash_profile is better than .bashrc)
12:06 jonathan Ah, I'd need to exit irssi and screen?
12:06 bacek (for LANG)
12:07 bacek just run another irssi client :)
12:07 jonathan OK, thanks, will experiment later. :-)
12:07 bacek Ctrl-A; export LANG=en_US.UTF8; irssi
12:08 bacek Ctrl-A-N probably... Last time I run screen few year ago...
12:27 ruoso joined #parrot
12:31 Lorn joined #parrot
12:37 ff-wonko joined #parrot
12:38 ruoso joined #parrot
12:39 bacek hmm... What is proper way to dumping PGE::Match objects?
12:39 ff-wonko joined #parrot
12:40 Alias joined #parrot
12:46 apeiron_ joined #parrot
12:55 kj bacek: have a look at languages/json
12:55 kj (note: languages/json, not compilers/json)
12:55 dalek r33649 | util++ | trunk:
12:55 dalek : Typo - s/trigonmetric/trigonometric/
12:55 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33649
12:56 bacek kj: grep Match . -r in json doesn't produce any results...
12:56 bacek any particular file?
12:57 kj actions.pm
12:57 purl somebody said actions.pm was getting kinda...big. What would you think to trying to refactor some of the larger actions a bit to use various subs, and putting those into a separate .pm file or files? It's just that it takes a little while to compile.
12:57 kj ISTR dumper is used there
12:57 kj see TOP method
12:58 bacek kj: o! thanks.
12:58 kj you can try to run some input through langauges/json, then you'll see what it does
13:00 bacek kj: it doesn't help... '_dumper' is for PAST classes AFAIU. I've got Match objects in Rakudo.
13:00 kj aah yes
13:00 kj eh, --target=parse?
13:00 pmichaud _dumper works for Match objects also.
13:01 pmichaud That's how --target=parse works.
13:01 bacek Method '_dumper' not found for invocant of class 'PGE;Match'
13:01 kj maybe it's just a sub?
13:01 bacek rakudo: token a { 'a' }; 'a' ~~ / <a> /; say $/.WHAT; say $/._dumper()
13:01 polyglotbot OUTPUT[Match␤Method '_dumper' not found for invocant of class 'PGE;Match'␤current instr.: '_block11' pc 70 (EVAL_15:25)␤called from Sub 'parrot;PCT;HLLCompiler;eval' pc 892 (src/PCT/HLLCompiler.pir:508)␤called from Sub 'parrot;PCT;HLLCompiler;evalfiles' pc 1227 (src/PCT/HLLCompiler.pir:674)␤called
13:01 polyglotbot ..from Sub 'parrot;PCT;HLLCompiler;command_line' p...
13:02 kj not a method on pge
13:02 pmichaud _dumper($/)
13:02 bacek pmichaud: O! Thanks.
13:06 pmichaud bacek (Match::ACCEPTS):  what should smart matching against a match do...?
13:06 pmichaud that seems a bit... weird.
13:07 jonathan rakudo: say [max] 1,4,3;
13:07 polyglotbot OUTPUT[Statement not terminated properly at line 1, near "1,4,3;"␤␤current instr.: 'parrot;PGE;Util;die' pc 129 (runtime/parrot/library/PGE/Util.pir:83)␤called from Sub 'parrot;Perl6;Grammar;eat_terminator' pc 26237 (src/gen_grammar.pir:3363)␤called from Sub 'parrot;Perl6;Grammar;statementlist' pc
13:07 polyglotbot ..24893 (src/gen_grammar.pir:2830)␤called from Sub...
13:07 bacek pmichaud: check 'exist self[topic]' probably.
13:08 jonathan rakudo: say [+] 1,4,3;
13:08 polyglotbot OUTPUT[8␤]
13:08 jonathan rakudo: say 1 max 4 max 3;
13:08 polyglotbot OUTPUT[4␤]
13:08 bacek pmichaud: I just trying to understand better way to handle something like token 'Foo { <Bar>|<Baz>|<AnotherBar>}'
13:09 bacek <moritz_> let me check the S03 smart match table first...
13:09 bacek yes, Any .foo - method truth
13:09 ruoso joined #parrot
13:09 bacek (second line by moritz_ aswell)
13:13 pmichaud We don't have smart match against method implemented yet
13:13 pmichaud that needs some syntactic magic
13:15 pmichaud haven't implemented [max] yet.
13:16 moritz also the [\+] etc. meta ops are missing
13:16 bacek pmichaud: ok.
13:16 pmichaud we need to refactor quite a few of the listops that are in List.pir
13:16 pmichaud (started some of that in the assign branch)
13:16 riffraff joined #parrot
13:16 pmichaud yes, the only metaop we have so far are the op=  and [op]
13:17 pmichaud and there, only for the ops listed in build/gen_metaops_pir.pl
13:18 pmichaud I might give a shot at »op« or XopX at some point
13:18 pmichaud these are all pretty big cheats at the moment, though.
13:19 pmichaud rakudo:  say max 1, 4, 3;
13:19 polyglotbot OUTPUT[4␤]
13:19 Alias pmichaud: Is [op] in 0.8.1?
13:19 pmichaud Alias: no, I just added it yesterday morning.
13:19 Alias ah, ok
13:20 Alias Because we were playing with Vanilla Perl 6 drunk at the pub after OSDC, and we weren't sure if it was not implemented or the fact none of us know Perl 6 yet :)
13:20 Alias Funny to see a bunch of serious Perl geeks with iPhones out trying to find Perl 6 tutorials :)
13:20 Alias Totally confirmed for me the effect of having a real distro too :)
13:20 pmichaud agreed.
13:21 Alias Had similar "Holy shit, it's real now?" experience work coworkers
13:21 Alias at work with coworkers
13:21 pmichaud that's why I'm also pushing for parrot to have its 'make install' house in order so we can do similar things there.
13:21 Alias Well, I had a reasonably easy time of that
13:21 Alias I just had to be insistant
13:21 Alias make reallyinstall
13:22 Alias http://svn.ali.as/cpan/trunk/Perl-D​ist-Parrot/lib/Perl/Dist/Parrot.pm
13:22 Alias That's about all the work I had to do
13:24 Alias It would be really nice if the next parrot release passed tests on Windows... pretty please
13:24 Alias (to whoever can arrange that)
13:28 pmichaud I probably need to set up a windows devel environment soon
13:32 Alias You have a CPAN login right?
13:32 Alias I'll be doing the rounds once I get the keys to this new Microsoft VM environment
13:32 Alias To get people that need it the most access
13:37 pmichaud yes, I have CPAN
13:38 sjn itym PAUSE?
13:38 pmichaud I also have Windows running in vmware, so it's just a matter of doing the development install (and having tuits for that)
13:41 Alias ah
13:42 Alias You probably need Vanilla...
13:42 Alias Well, maybe not
13:45 riffraff pmichaud, did you implement [] as a proper metaoperator? cause I noticed some special casing for [+] and others
13:45 pmichaud it's not a proper metaoperator in the sense that the grammar recognizes '[' and ']' as metaoperator tokens
13:46 pmichaud for that we need LTM
13:46 riffraff ah I see
13:46 pmichaud but I figured the cheat was "good enough" for now to get the popular reduce operators in place
13:46 pmichaud and the spec *does* say that this approach is valid
13:47 riffraff still missing [~] maybe?
13:47 pmichaud I thought I put [~] in there
13:47 pmichaud rakudo:  say [~]  <a b c d>;
13:47 polyglotbot OUTPUT[abcd␤]
13:47 riffraff og I didn't see it in gramar-oper.pg
13:48 pmichaud oh, the only ones that appear in grammar-oper.pg are the ones that are already listops
13:48 pmichaud for example,  [&]  translates to all()
13:48 pmichaud there's no need to do a reduction
13:48 pmichaud the rest are in build/gen_metaops_pir.pl
13:49 riffraff yeah just found them thanks
13:50 pmichaud ...why can I not find anything about Vanilla Perl 6 on the web?
13:50 kj Vanilla Perl 6?
13:51 kj purl doesn't know it..
13:51 purl kj: what?
13:54 riffraff mh, I think this should fail, but possibly not like this
13:54 riffraff rakudo: [~>](1,2,3)
13:54 polyglotbot OUTPUT[Null PMC access in invoke()␤current instr.: '!REDUCEMETAOP' pc 11692 (src/builtins/assign.pir:99)␤called from Sub '_block11' pc 51 (EVAL_13:20)␤called from Sub 'parrot;PCT;HLLCompiler;eval' pc 892 (src/PCT/HLLCompiler.pir:508)␤called from Sub 'parrot;PCT;HLLCompiler;evalfiles' pc 1227
13:54 polyglotbot ..(src/PCT/HLLCompiler.pir:674)␤called from Sub 'parrot;...
13:54 pmichaud looks like we haven't implemented ~> yet.
13:55 pmichaud (infix:«~>», that is.)
13:56 ruoso joined #parrot
14:00 dalek r33650 | pmichaud++ | trunk:
14:00 dalek : [rakudo]: spectest-progress.csv update: 233 files, 4779 passing, 0 failing
14:01 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33650
14:01 pmichaud yay.  +62 passing from yesterday.
14:01 pmichaud +90 passing from last week
14:02 jonathan Maybe we can do 5,000 before the year is out.
14:02 pmichaud I'm thinking maybe by dec release.
14:02 jonathan pmichaud: Maybe I'll take a stab at the hypers on Rakudo day. :-)
14:02 pmichaud I was only going to do the infix hypers
14:02 jonathan That's fine. :-)
14:02 pmichaud but yes, it shouldn't be too difficult to do
14:03 pmichaud just follow the same model as the reductions -- have a standard function that does the work
14:03 jonathan *nod*
14:03 jonathan Yes, I planned to see how you'd done it.
14:03 jonathan And follow the lead.
14:03 moritz with slices we might be able to get 5k tests
14:03 pmichaud anyway, I have the core of list slices and list assignment "working" in the assign branch, so we could strike 5k tests soon
14:03 moritz (assuming that I can spare more tuits for writing and/or moving test)
14:04 pmichaud ("working" in quotes because I haven't gone back to clean up the rough edges that cause spectests to regress)
14:04 jonathan Nice progress, though.
14:04 jonathan Just as rvalues for now?
14:04 pmichaud and yes, that is really *list assignment*
14:04 jonathan Or lvalues too?
14:04 pmichaud lvalues too, but it means we auto-extend arrays on access
14:05 jonathan OK.
14:05 jonathan Nice work!
14:05 pmichaud i.e.:   my @a;  my $b = @a[1000];   say @a.elems;   # 1000
14:05 jonathan pmichaud++
14:05 jonathan Ah.
14:05 pmichaud I figure we can live with that until proxies are in place
14:05 jonathan Yes.
14:05 pmichaud I'd rather have slices and list assignment than make sure my arrays are the right length
14:06 moritz pmichaud: are you sure your example is correct?
14:06 pmichaud my point is that my example is incorrect
14:06 moritz ah
14:06 * moritz is relieved
14:06 pmichaud I'm going to go ahead and implement slicing with that incorrect feature for now, and clean it up a bit later.
14:07 jonathan Agree - it's a step in the right direction.
14:08 pmichaud I figure we'll pass far more tests than we suddenly fail.
14:08 jonathan pmichaud: Might do Rakudo day tomorrow. :-)
14:08 jonathan Work for you?
14:08 pmichaud that would be good.  I'm scheduled to be out of town Wed and Thu.
14:08 jonathan I've seen the flood of masakbugs!
14:08 gryphon joined #parrot
14:08 moritz jonathan: yes, masak has a new project (a text UI game)
14:09 pmichaud > my @a = 1..5;   @a[7,6,5] = @a[1,2,3];  say ~@a;
14:09 pmichaud 1 2 3 4 5 4 3 2
14:09 moritz jonathan: new features, new requirements, new bug reports.
14:09 pmichaud it even works if lhs elements are on the rhs:
14:09 pmichaud > my @a = 1..5;   @a[0,1] = @a[1,0];  say ~@a;
14:09 pmichaud 2 1 3 4 5
14:10 ruoso joined #parrot
14:11 jonathan Bug reports good. :-)
14:12 ruoso joined #parrot
14:14 jhorwitz joined #parrot
14:16 riffraff joined #parrot
14:17 mberends joined #parrot
14:19 particle joined #parrot
14:26 gmansi joined #parrot
14:35 jhorwitz joined #parrot
14:36 jimmy joined #parrot
14:38 jimmy Is Whitenight usually not here?
14:38 jimmy Is Whiteknight usually not here?
14:39 moritz he is, from time to time
14:42 dalek r33651 | pmichaud++ | assign:
14:42 dalek : [pct]:  Allow 'lvalue' attribute on all PAST::Nodes.
14:42 dalek : So far, lvalue is still only meaningful for PAST::Var nodes,
14:42 dalek : but at least we don't throw an exception for other types.
14:42 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33651
14:44 jimmy moritz: I have not seen him for a long time.
14:44 moritz seen Whiteknight
14:44 purl Whiteknight was last seen on #parrot 13 hours, 2 seconds ago, saying: okay, i'm off to bed now. Goodnight!
14:44 dalek r33652 | pmichaud++ | assign:
14:44 dalek : [rakudo]:  clean up lvalue and list() semantics
14:44 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33652
14:44 Infinoid good morning
14:45 jimmy morning
14:45 ruoso joined #parrot
14:46 iblechbot joined #parrot
14:47 PerlJam good morning #parrot
14:48 pmichaud good morning, PerlJam
14:48 pmichaud nice work on :nth
14:49 moritz ... and there's already a ticket asking for more :-)
14:49 moritz and tests in the appropriate places
14:49 pmichaud is it specced how :nth is supposed to work with :x and/or :g?
14:50 moritz no
14:50 jimmy moritz: can you take a look at tarc #20
14:50 moritz jimmy: not now, sorry
14:50 jimmy moritz: nothing
14:50 ruoso joined #parrot
14:51 pmichaud moritz: the problem with the 'expectation' you give in RT #61130 is that :nth doesn't have to be an integer.
14:52 pmichaud .subst( $pat, $rep, :x(2), :nth({ .is_prime }) )
14:52 moritz ouch
14:53 pmichaud or
14:53 pmichaud .subst( $pat, $rep, :x(2), :nth(1,2,3,5) )
14:53 moritz so what I thought was :nth(2) would actually be written as :nth(2, 4, 6 ... *), right?
14:53 pmichaud that's what I would guess.
14:54 * Infinoid applies jimmy's (trivial, documentation) patch
14:54 PerlJam :nth(2..*:by(2))  ?
14:54 pmichaud PerlJam: I'm curious about that one.
14:54 pmichaud in particular, I don't know what happens with    $x ~~ 2..*:by(2)
14:54 PerlJam 2:by(2) would be an interesting shortcut if allowed  :)
14:55 pmichaud and 2:by(4) would be lumber.  :-)
14:55 jonathan I'm not sure the range constructor is smart enough to infer :by from the previous things.
14:55 PerlJam It plays right into the religious references that show up in perl occasionally
14:55 pmichaud jonathan: infix:<...>  isn't a range constructor
14:55 pmichaud It's the "series operator"
14:55 moritz PerlJam: I think that adverbs always go on ops, so 2:by(2) wouldn't work
14:56 pmichaud moritz: correct
14:56 PerlJam moritz: right.
14:57 pmichaud (and yes, infix:<...> does infer :by from the previous things if it can)
14:57 PerlJam that sure is a load of meaning hinged on a single .  (the difference between .. and ...)
14:57 jonathan Oh, three dots.
14:58 dalek r33653 | infinoid++ | trunk:
14:58 dalek : [doc] Apply documentation patch from jimmy++ (trac #20).
14:58 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33653
15:00 Coke why did we just switch parrot.org to parrotcode.org in that patch?
15:00 Coke ah, because. nevermind.
15:06 Lorn_ joined #parrot
15:12 ruoso joined #parrot
15:15 ff-wonko joined #parrot
15:22 particle i can't build JSON in the pdd2210 branch, due to a tge error
15:22 particle Null PMC access in get_pmc_keyed_int()
15:22 particle current instr.: 'parrot;TGE;__onload' pc 19 (TGE.pir:94)
15:23 ff-wonko joined #parrot
15:23 particle why would i get get_pmc_keyed_int() failure with:     load_bytecode 'compilers/tge/TGE/Rule.pbc'
15:25 particle nevermind...
15:25 dalek r33654 | pmichaud++ | assign:
15:25 dalek : [rakudo]:  more code adjustments
15:25 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33654
15:25 particle cd compilers/tge && nmake clean all # fixes it
15:26 particle looks like there's a bad/missing dependency in tge makefile
15:27 pmichaud yes, istr running into that also
15:32 allison joined #parrot
15:36 Tene pmichaud: I'll likely work on adding the EH stuff to PAST::Block compiler bits today.  If you could work out how to get .include include/except_types into a prelude somewhere, that would be great.
15:36 pmichaud Tene: okay, will do.
15:37 particle allison: lots of pdd22io failures for me. tests are almost done here, will report
15:37 pmichaud I'm working on pdd22io also.
15:37 allison particle: thanks
15:38 nopaste "particle" at 76.121.106.245 pasted "pdd22io test failure report" (2570 lines) at http://nopaste.snit.ch/14858
15:41 allison particle: lots of "Cannot write to a filehandle not opened for write"
15:42 particle allison: yep, but there's a surprising dynpmc error...
15:43 particle "The procedure entry point Parrot_default_type_keyed_str could not be located in the dynamic link library libparrot.dll."
15:44 particle PARROT_EXPORT must be missing somewhere
15:45 pmichaud how old is the version of rakudo that is in the branch?
15:45 allison particle: probably so, but not on Parrot_default_type_keyed_str. maybe on one of the I/O api functions
15:46 allison pmichaud: updated last night
15:46 pmichaud okay.
15:46 particle allison: right. i've just run headerizer, and am inspecting the diff now
15:47 particle string_free needs PARROT_EXPORT. lemme rebuild...
15:47 particle oh, sorry, that's PARROT_INLINE :(
15:50 pmichaud afk for a bit while spectest runs
15:51 dalek r33655 | fperrad++ | trunk:
15:51 dalek : [WMLScript] box
15:51 dalek : refactor with opcode 'box' :
15:51 dalek : - code generation
15:51 dalek : - standard libraries
15:51 dalek : - test PMC
15:51 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33655
15:52 dalek r33656 | allison++ | pdd22io_part2:
15:52 dalek : [pdd22io] Actually set filehandle flags when opening an already existing
15:52 dalek : filehandle object.
15:52 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33656
15:52 allison particle: just committed a fix for the "filehandle not opened for write" problem, update and rerun when you get a chance
15:52 * allison biab
15:56 AndyA joined #parrot
16:00 dalek r33657 | fperrad++ | pdd22io_part2:
16:00 dalek : [pdd22io] Lua
16:00 dalek : - replace ParrotIO.slurp() by FileHandle.readall()
16:00 dalek : (TODO: error handling)
16:00 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33657
16:15 Andy joined #parrot
16:22 pmichaud allison:  the errors I'm getting in rakudo's "make spectest" appear to be the harness-related errors
16:22 pmichaud when I run the files directly from the command line, they work fine.
16:22 pmichaud oh, wait, no they don't.
16:22 pmichaud checking again.
16:23 pmichaud at a first glance, I'm thinking the ones that fail are the ones containing utf8 characters.
16:23 pmichaud which seems to tell me that the files aren't being read as utf8
16:25 masak joined #parrot
16:26 elmex joined #parrot
16:32 pmichaud seems to be confirmed... one sec
16:32 Infinoid Rafe's law: An Internet service cannot be considered truly successful until it has attracted spammers.
16:32 Infinoid regarding trac and spam, http://trac.edgewall.org/wiki/SpamFilter looks like a nice plugin
16:34 tewk Can we mark t/pmc/exception untodoed, its passing for me
16:35 particle in trunk?
16:35 tewk yep
16:35 nopaste "pmichaud" at 72.181.176.220 pasted "FileHandle 'readall' method doesn't return utf8 encoded string (for allison)" (19 lines) at http://nopaste.snit.ch/14860
16:35 tewk test 30 is passing
16:36 dalek r33658 | fperrad++ | trunk:
16:36 dalek : [Lua] box
16:36 dalek : - test the opcode 'box' with the HLL 'Lua'
16:36 dalek : - now, LuaNumber PMC maps Float & Integer
16:36 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33658
16:37 particle tewk: it's not passing with msvc
16:37 particle #   Failed (TODO) test 'catch ex from C-level MULTI function'
16:37 tewk ok, I've seen it passing for a week so I thought I would ask.
16:38 Limbic_Region joined #parrot
16:39 tewk Can spectest run in parallel?
16:39 jhorwitz tewk: fyi, the "V" NCI signature works beautifully in mod_parrot
16:39 Coke (undo'd) - someone reported it was failing on ppc/darwin
16:39 Coke s/undo/un-todo/
16:39 Coke so if you just skip it there, that's fine.
16:39 Coke s/skip/todo/
16:39 Infinoid it passes here on x86 but fails (as expected) on x86-64
16:40 Infinoid (linux)
16:40 particle tewk: yes, check rakudo's t/harness, set TEST_JOBS=9 or whatever
16:44 tewk jhorwitz: great
16:49 tewk pmichaud: subids for multis fixed.
16:50 dalek r33659 | tewk++ | trunk:
16:50 dalek : [subid] fixed multi lookup via subid
16:50 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33659
16:52 dalek r33660 | tewk++ | trunk:
16:52 dalek : [ncigen] [sqlite] almost generating SQLite3.pir
16:52 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33660
16:52 dalek r33661 | tewk++ | trunk:
16:52 dalek : [ncigen][sqlite] MANIFEST
16:52 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33661
16:58 ff-wonko joined #parrot
17:01 dalek r33662 | tewk++ | trunk:
17:01 dalek : Trying to be a good codeing standard robot
17:01 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33662
17:22 Infinoid what's the current status of "make install"?
17:22 pmichaud unchanged, I think.
17:22 Infinoid is anyone currently working on this, or is it still so broken that applying random patches from tickets won't hurt anything?
17:22 pmichaud at least, unchanged in trunk.
17:23 moritz Infinoid: rurban was working on this, but lost time/track
17:24 particle Infinoid: there's a branch for pdd30_install
17:24 particle rurban was working on it, and i believe he gave a recent list mail update on his and its status
17:25 Infinoid ok, I'll look at that.  I mainly just want to know if I can apply (and close) RT #60608, or if I'd be creating merge problems for anyone.
17:25 Infinoid thanks
17:25 moritz Infinoid: I think the pdd30 branch is so out-of-date that it doesn't really matter :/
17:26 Infinoid I see it on ParrotRoadmap for january's release, with no name assigned to it
17:27 particle yes, and that's a big risk item
17:27 cognominal joined #parrot
17:27 allison joined #parrot
17:27 particle there are many dependencies on install
17:28 pmichaud we really need someone to take over the 'make install' issue, if rurban won't be able to do it between now and Jan release
17:29 Infinoid okay.  distro packaging folks tend to know what they're doing, so I'm going to apply the patch for now.  it's a tiny tweak, and I figure we can back it out if it conflicts with the pdd30 branch contents
17:29 Infinoid I really need to start working on pdd13 again soon
17:31 jonathan Infinoid: I've got a branch where I'm doing the bytecode annotatiosn chunk of it.
17:32 Infinoid cool.  the stuff on my plate will happen in trunk, until its time to switch the rest of parrot over to using it
17:33 Infinoid but I'll make sure that I don't conflict with your branch
17:34 jonathan OK, great.
17:34 tomyan left #parrot
17:35 contingencyplan joined #parrot
17:36 dalek r33663 | infinoid++ | trunk:
17:36 dalek : [INSTALL] Apply patch from lu_zero++ to fix --libdir handling.
17:36 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33663
17:42 Infinoid sigh.  Infinoid-- for forgetting to mention the ticket number in the commit message
18:33 allison joined #parrot
18:38 allison joined #parrot
18:42 cotto seen kj
18:42 purl kj was last seen on #parrot 4 hours, 50 minutes and 59 seconds ago, saying: purl doesn't know it..
18:43 kj cotto: hi
18:43 kj looking for me?
18:48 * Coke finally completes a todo item that's been on his list for a week.
18:53 * kj wonders what todo that was, if it took a week :-)
18:57 * Coke was just lazy. (The email I just sent out.)
18:59 particle joined #parrot
19:20 Hadi joined #parrot
19:20 Hadi left #parrot
19:25 AndyA joined #parrot
19:28 rdice joined #parrot
19:38 chromatic joined #parrot
19:40 * Coke throws an attack kitten at chroamtic.
19:41 Coke ... lucky for you I spelled it wrong!
19:42 chromatic No kidding.  They're sharp.
19:42 Coke so are we going to try to get IO into this release?
19:43 Coke I think if we merge by parrotsketch tomorrow we'll be in good shape before the release.
19:48 ff-wonko joined #parrot
19:49 dalek r33664 | allison++ | pdd22io_part2:
19:49 dalek : [pdd22io] Keep the encoding set on temporary filehandle for 'readall' on an
19:49 dalek : existing FileHandle object.
19:49 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33664
19:51 dalek r33665 | allison++ | pdd22io_part2:
19:51 dalek : [pdd22io] Tests for encoding on 'readall'.
19:51 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33665
19:54 ffwonko joined #parrot
20:01 particle coke: there's still a number of win32 failures for pdd22io branch
20:05 dngor joined #parrot
20:08 dalek r33666 | particle++ | pdd22io_part2:
20:08 dalek : [pmc] fix C89-o
20:08 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33666
20:09 allison joined #parrot
20:10 workbench joined #parrot
20:12 donaldh joined #parrot
20:13 stockwellb joined #parrot
20:13 GeJ Good morning everyone
20:14 Tene Hi!
20:17 * Coke takes a nap waiting for a parrot checkout on windows.
20:19 Tene rakudo: my @a = 1,2,3; say 'omg' if @a[1|2] == 2;
20:19 polyglotbot OUTPUT[get_integer() not implemented in class 'Junction'␤current instr.: '!MAKE_JUNCTION' pc 6681 (src/classes/Junction.pir:222)␤called from Sub '_block11' pc 78 (EVAL_12:27)␤called from Sub 'parrot;PCT;HLLCompiler;eval' pc 892 (src/PCT/HLLCompiler.pir:508)␤called from Sub
20:19 polyglotbot ..'parrot;PCT;HLLCompiler;evalfiles' pc 1227 (src/PCT/HLLCompiler.pir:674)␤...
20:19 Coke particle: build errors or test failures?
20:19 Coke (I assume you're still building with the MS tools)
20:20 dalek r33667 | particle++ | pdd22io_part2:
20:20 dalek : [src] headerizer changes
20:20 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33667
20:21 allison particle: ping
20:21 particle coke: fixed the build error, now test failures
20:21 particle pong, allison :)
20:22 allison particle: do we have any remote windows debugging solution?
20:22 particle calling_conventions branch has build errors, too
20:22 stockwellb joined #parrot
20:22 particle allison: i'll try to get you remote access to my vm...
20:23 Coke and all of us with PAUSE ids will have access to the msoft build stuff shortly.
20:23 allison Coke: that'll be useful
20:23 particle indeed
20:24 nopaste "coke" at 72.228.52.192 pasted "pdd branch failures on strawberry." (17 lines) at http://nopaste.snit.ch/14863
20:24 allison Coke: this is pdd22?
20:24 allison Coke: that's progress from this morning...
20:25 Coke yes.
20:25 Coke URL: http://svn.perl.org/parrot/branches/pdd22io_part2
20:25 Coke Revision: 33666
20:25 allison looks like the "filehandle not opened for write" bug is resolved...
20:26 Coke ah. the File::Spec stuff is probably a missed 'global' update from trunk.
20:26 chromatic particle, what's with removing the parameter decorations in r33667?  Did they not work on Windows?
20:26 Coke (that or it's something that was never fixed in trunk because I don't normally test on windows.)
20:26 particle chromatic: i ran make headerizer, then committed
20:30 chromatic We have ping-pong commits there, then.
20:30 chromatic If someone runs make headerizer on 32-bit x86 Linux and commits, it'll reapply my changes there.
20:30 chromatic That means we need to decorate all of the functions under src/jit/ appropriately.
20:31 allison do 'todo' tests work on windows? the 'record_separator' test is failing, but it's marked as 'todo'
20:31 allison (at least, was failing in KJS's last report from windows)
20:32 particle don't some older T::H versions report a passing todo as a failure?
20:33 allison well, the error was "# Method 'record_separator' not found for invocant of class 'FileHandle'"
20:34 allison but, then he didn't include the summary, so may not be reported as a failure
20:34 allison I'll trust Coke's latest report for that
20:34 particle i'm running make test now
20:34 particle did you get my privmsgs?
20:42 Coke allison: created a tcl branch to track changes necessary here; getting the "tests fail when run with prove but work when run with ../../parrot tcl.pbc) mode.
20:42 Coke er, failure mode.
20:42 Coke https://partcl.googlecode.​com/svn/branches/parrot_io
20:43 allison Coke: is that a problem you've encountered before?
20:43 Coke recently? no. rakduo has, though.
20:45 cotto chromatic, ping
20:45 chromatic pong
20:45 ruoso joined #parrot
20:45 cotto do you happen to have that patch you mentioned in the email about the OrderedHash bug?
20:46 chromatic No, but it's really easy.  Just add an INTVAL to the end of the Hash struct declaration in parrot/hash.h.
20:46 allison Coke: I think with Rakudo it's often a GC thing because the tests run with different options than running the code straight
20:47 pmichaud we've actually narrowed it down to an environment variable.
20:47 cotto You mean _hash?
20:47 chromatic Yes.
20:47 cotto I tried that, but t/pmc/orderedhash.t was fine.  trying again...
20:49 cotto same result
20:49 Coke allison: entirely possible, esp. given that I have GC failures in trunk.
20:49 cotto There's some programming by coincidence somewhere in that code.
20:49 Coke ah, my test harness still supports TEST_PROG_ARGS. Yup.
20:50 Coke tests start passing.
20:50 Coke Of course, they soon eat up an insane amount of memory and swap.
20:51 chromatic cotto, I misremembered.  I added an INTVAL at the end of the HashBucket struct.
20:52 allison Coke: as long as it's only the usual insane amount of memory and swap...
20:52 cotto chromatic, thanks.  trying to reproduce...
20:53 cotto while I'm here, is there any reason that pmc2c.pl needs to do --dump and --c separately?
20:54 chromatic I'm not sure anyone knows.
20:54 particle is --dump for the string constants?
20:55 pmichaud allison: pdd22io branch now has only one failing rakudo spectest
20:55 pmichaud t/spec/S16-filehandles/io_in_whil    1   256    13   26 200.00%  1-13
20:55 chromatic It shouldn't be; c2str should handle that the same way it does regular .c files.
20:55 pmichaud I'm happy to regress that one test, as we know we need to re-do Rakudo's I/O anyway.
20:56 allison pmichaud: excellent (that it's only one test)
20:56 tewk cotto: I think it help do incremental builds.
20:56 tewk you change a child, you don't have to reparse the parent.
20:57 allison pmichaud: and regressing one is a tolerable loss for a revamped I/O system. Let me know if you need anything for Rakudo's I/O when you update
20:57 pmichaud will do.  jonathan++ did most of the I/O work, so perhaps he'll take a crack on it again with the new stuff.  If not, then I'll probably get to it later in the weeke.
20:57 pmichaud *week
20:57 jonathan chromatic: ping
20:57 cotto chromatic, you suggestion causes that test to segfault.  It's even the same test that segfaults with a different hash seed.
20:57 pmichaud and yes, the revamped I/O is a big plus.
20:58 cotto tewk, I don't think so.  When I touch a .pmc, both --dump and --c are run.
20:59 nopaste "particle" at 76.121.106.245 pasted "pdd22io test failure report" (2056 lines) at http://nopaste.snit.ch/14864
20:59 tewk right, but you don't have to redump the parents.
20:59 particle touch a parent, and see if the children re-run dump
21:00 cotto particle, they do
21:00 particle there you go, tewk++ is correct.
21:01 cotto the children also rerun --c
21:01 chromatic cotto, then we've found the source of some flakiness.  It looks like hash thawing.
21:01 particle that's what i'd expect
21:02 tewk dumping is priming the cache.
21:02 jonathan pmichaud: If it's only one test we regress on, I can very likely look into it tomorrow.
21:02 allison particle: got it, thanks! will look after lunch
21:02 tewk once a parent dumps its children can use the dump rather than reparse the parent
21:03 jonathan allison: Is the IO PDD in line with the implementation now, so far as you know?
21:03 jonathan As in, can I read the PDD and have a good idea of what to expect?
21:03 particle jonathan: yes
21:03 jonathan OK, good.
21:03 particle and there's even tests :)
21:04 jonathan pmichaud: I/O not sucking in Rakudo partly seems to depend on laziness issues, unfortunately.
21:04 pmichaud I still don't understand why that's a huge issue.
21:04 jonathan Until then, for =$fh -> $line { ... } stuff ain't going to play ball.
21:04 pmichaud =$fh is evaluated in list context, so just go ahead and read all of the lines of the file at that point
21:05 pmichaud we'll know that because its .list() method gets executed.
21:05 pmichaud (i.e., this may be fixed in the assign branch)
21:05 jonathan pmichaud: We *can* do that, but it won't play well with say =$*IN?
21:05 pmichaud =$*IN is evaluated in list context, so read all of the lines.....
21:06 jonathan for =$*IN -> $line { say $line; say "enter next thingy"; }
21:06 jonathan Should this not keep looping?
21:06 jonathan Or do you have to write while =$*IN { ... } for that?
21:06 particle for, not while
21:07 pmichaud while would probably work here as well, but yes, 'for' is canonical
21:08 Coke hurm. when run with -G, my test suite hangs on like the 5th test.
21:08 pmichaud jonathan: okay, I see your point.  We might indeed have issues with =$*IN and laziness, yes.
21:08 pmichaud but other "non-interactive" forms of I/O ought to be okay for now.
21:09 pmichaud I'm not too sure I want to blaze too many trails on laziness until S07 lands.
21:09 jonathan OK. If we put that aside, I agree we can make it work.
21:09 pmichaud or at least gets closer.
21:09 jonathan No, me either.
21:09 Coke (in the paired parrot and tcl io branches)
21:09 * Coke writes a perl script to verify his answer to this week's puzzler.
21:09 jonathan I already did blaze one trail on it... :-)
21:09 * Coke includes the script for extra credit.
21:09 jonathan pmichaud++ # nice subject line on p6l
21:10 pmichaud I also thought about adding "Perl's gone wild" in there somewhere, but decided to save that for my next post.  :-P
21:13 cotto chromatic, in both OrderedHash segfaults, it looks like src/hash.c:1240 is somehow adding a bogus key to the OrderedHash
21:13 Tene pmichaud: You said PCT prelude coming today?
21:13 cotto clone segfaults in the test because there's no valid value associated with the key
21:13 pmichaud Tene: still working on that.  I'm trying to decide if it deserve a special POST node or what.
21:15 allison joined #parrot
21:17 tewk w
21:17 chromatic 1240?
21:18 Tene I can't start on it until after I finish perparing for class tomorrow anyway.
21:19 allison resolving test failures in the lunch line++
21:19 cotto yes, the return statement
21:19 purl the return statement is 2 instructions I think
21:19 dalek r33668 | allison++ | pdd22io_part2:
21:19 dalek : [pdd22io] Setting filehandle flags in 'fdopen' also.
21:19 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33668
21:19 allison particle: try t/pmc/io.t now
21:20 chromatic is that src/pmc/hash.c or src/hash.c?  :1240 in the latter is POD, for me.
21:21 nopaste "cotto" at 96.26.202.243 pasted "backtrace on OrderedHash when changing the hash key to 3794" (28 lines) at http://nopaste.snit.ch/14865
21:21 cotto src/hash.c
21:22 cotto (I added a method to OrderedHash that prints all keys and values, which is what the breakpoing at 603 is)
21:23 particle allison: failed 4-5
21:23 particle # Cannot write to a closed filehandle
21:24 chromatic Doesn't make sense to me at the moment.
21:26 PerlJam chromatic: why do you ask the 5-devs question on use.perl?
21:27 particle $$$
21:27 PerlJam chromatic: did you win the lottery?
21:32 chromatic The only way to win is not to play.
21:33 chromatic I was just noodling with an idea.
21:33 * Infinoid would make some real AI and take everything over, matrix style.
21:33 PerlJam Infinoid: money doesn't solve such problems.
21:33 Infinoid no, but 5 developers would make a big dent in it
21:35 Infinoid but with such an open-ended question, I think the answers would be utterly random.
21:36 PerlJam My answer would be:  I'd have them work on Parrot and Perl6  until I found something better for them to do.
21:38 dalek r33669 | pmichaud++ | assign:
21:38 dalek : [rakudo]:  Fix forgotten namespace directive.
21:38 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33669
21:38 chromatic Five FTE is an order of magnitude more than Perl 6 and Parrot have.
21:38 Coke FTE?
21:38 purl hmmm... FTE is made by the guy that also coded icewm BTW :) or full time employee aka perm
21:38 Coke danke, purl.
21:38 purl bitte
21:40 Tene Also, there certainly exist developers who would take a pay cut to work on something they love.
21:40 * Infinoid raises his hand enthusiastically
21:40 Coke I, however, am not one of them.
21:40 chromatic Sure.  $120k/year in salary and benefits for telecommuting is very attractive.
21:41 * Coke 's ears perk up.
21:41 * Coke suspects that's not a pay cut for 99% of our committer base. =-)
21:41 Tene I'm currently taking a significant pay cut to do work that I enjoy.
21:42 * Infinoid does work he enjoys for free
21:42 pmichaud I took a big pay cut to be able to do this, but (so far) it hasn't been nearly as big as I might've feared and I have significantly fewer $job stresses
21:43 pmichaud if course, if we measure "actual pay cut" versus "what I'd be making today if I had remained in my former position", the size of the cut is _much_ bigger.
21:44 Coke executable memory: http://rt.perl.org/rt3/Tic​ket/Display.html?id=43146
21:44 chromatic $120k is a rough estimate of what it might take to tempt me.
21:45 chromatic ... being, uh, not 21, reasonably experienced, and used to health insurance and 401(k).
21:46 Infinoid Coke: I'd say the existence of PARROT_HAS_EXEC_PROTECT means that's been covered, at least, from the perspective of config/detection
21:47 PerlJam chromatic: you're not 21?  Hey, I'm not 21 too!  Is today your unbirthday?
21:49 pmichaud otoh, for some bizarre reason I continue to be employed by my "former" employer.  19 years next month.
21:49 PerlJam pm: hey, in 13 months you can "retire" :)
21:49 chromatic Today is my 25% birthday.
21:49 pmichaud PerlJam: well, I could if all of those years had been fte's  :-)
21:50 Coke Infinoid: feel free to grab and close the ticket. the subject line just caught my eye based on recent commits.
21:51 * Infinoid goes and reads the recent commit log
21:51 pmichaud 14.75 of the years were full time, the other 5.25 years are 25% time.  I guess that comes out to around 16 years of service.
21:51 pmichaud also, I'm not in TRS, so the normal "retirement" calculations don't apply to me.
21:51 Tene TRS?
21:52 pmichaud "texas retirement system" or "teacher retirement system" or whatever the standard retirement program is for educators in Texas
21:52 PerlJam pm: I bet you'll still be employed by TAMUCC when you're 85.  They'll never let you go.
21:52 dalek r33670 | coke++ | trunk:
21:52 dalek : Fix example PIR against recent PIR syntax cleanup
21:52 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33670
21:52 Infinoid "svn log" for the last 3 weeks is 2900 lines.  You guys are incredible.
21:53 jonathan (long commit messages)++ ;-)
21:53 Infinoid that's significantly more than I would expect from 5 FTE's
21:54 pmichaud I think chromatic was referring to "paid FTEs", not just FTEs working on parrot
21:54 chromatic I write long commit messages anyway.
21:55 Infinoid stop trying to argue yourselves out of being incredible. :)
21:58 chromatic I'm trying to argue that we could be more incredible.
21:58 particle here's what gets me: http://www.perlfoundation.org/parrot/index.cgi/​parrot-devel-0.7.0.png?action=attachments_downl​oad;page_name=parrot;id=20080823041511-0-7546
21:58 particle our velocity is *still increasing*
21:59 Tene particle: could that be updated to current?
21:59 Tene Looks like that's from Jan 08
21:59 particle it's from august 08
22:00 particle r30369
22:00 Tene Ah.
22:00 Tene Oh, I just can't read.
22:01 PerlJam hey, is tomorrow a release day?
22:01 particle no
22:01 particle tomorrow's the second tuesday
22:01 * PerlJam doesn't keep track of time too well.
22:02 PerlJam Although I should since it's my son's birthday on Wednesday.
22:02 PerlJam I hope I remember on that day
22:03 Coke rant: trac's wiki doesn't show what pages link to the page you're on.
22:04 * jonathan stares at calling_sets.t and starts clearing it up
22:05 Whiteknight joined #parrot
22:12 * Coke cleans up the repository slightly.
22:12 * Coke waits to see if the logbot can deal with that. =-)
22:12 Coke dalek?
22:12 purl dalek is, like, dha's distributed perl community badger badger badger attack or http://ourworld.compuserve.com/hom​epages/grahamwalters/dalek_fr.htm or 'Dalek' for the language, and 'dalek' for the program or http://www.daleklinks.co.uk/ or a Dr Who baddie or {see: dalek meme} or at http://www.deviantart.com/deviation/20016573/ or http://www.asciiartfarts.com/20020615.html or http://xrl.us/2doh
22:13 dalek r33671 | coke++ | RELEASE_0_:
22:13 dalek : Remove some old branches; the result of these releases are logged as tags.
22:13 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33671
22:13 Coke nice. there's a dalek bug for someone. =-)
22:15 Infinoid uhhoh, I wrote that code
22:15 Coke doesn't deal with cross-branch commits.
22:15 Coke well, it did, sort of.
22:16 Infinoid we hacked SVN::Web to include the branch name in the RSS data.  the bug is probably in that.
22:16 Coke pmichaud: is the branch for RT 59410 obsolete?
22:17 Coke (you resolved the ticket on 01OCT08)
22:17 Coke ->
22:18 mj41 coke TapTinder has also branch list ... http://tt.ro.vutbr.cz/report/pr-Parrot/
22:18 pmichaud which branch is the one for RT 59410?
22:21 gmansi joined #parrot
22:29 PerlJam particle: are you sure that graph shows an increasing velocity?
22:29 Tene PerlJam: eh?
22:29 purl Speak up, sonny!
22:30 PerlJam without any distinguishing marks, it looks like it's just a plot of time versus commit.  Which as long as there are *any* commits, the graph will continue increasing.
22:31 Tene PerlJam: look at the slope of the graph
22:31 Tene PerlJam: the number of commits in each unit of time is greater than in the previous unit.
22:31 Tene If the commit rate remained constant, the graph would be linear.
22:33 gmansi joined #parrot
22:33 PerlJam Tene: If you look at the graph from Jan 01, 2007 to now, it looks fairly linear.
22:34 Tene Huh.  You're right.
22:35 PerlJam What's the graph of the first differences look like?
22:35 PerlJam (where's the data?  or do I have to troll the logs for it?)
22:40 chromatic http://www.perlfoundation.org/parrot/index.cgi has commit counts between releases.
22:42 dalek r33672 | moritz++ | trunk:
22:42 dalek : [rakudo] add tests for meta reduction ops to t/spectest.data
22:42 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33672
22:44 jonathan That must give us a nice few more. :-)
22:44 moritz jonathan: not so many, actually
22:44 chromatic Coke, you'll like this: http://www.forta.com/blog/index.cfm/2008/12​/8/Gartner-Recommended-WHAT-Over-ColdFusion
22:44 PerlJam chromatic: from that data I guess I could infer a generally increasing trend line
22:45 chromatic PerlJam, there's likely to be chart junk in there however.
22:45 chromatic What's the size of a commit?  What's the impact of a commit?
22:45 moritz jonathan: but that file is full of TODOs of which I suspect that many are easy for you
22:46 jonathan moritz: Aha.
22:46 jonathan moritz: Found we weren't testing .+, .* and .? at all.
22:46 jonathan moritz: And .* had got broken.
22:46 jonathan So, have got tests for the single dispatch case written.
22:47 PerlJam chromatic: Have you perlbuzzed that article?
22:47 jonathan And now doing some basic ones for the multi case.
22:47 jonathan And then will review the more complex ones etc.
22:47 jonathan Finding/fixing bugs as I go. :-)
22:47 jonathan Should give us another at least 20 tests...
22:47 jonathan (Passing ones)
22:49 chromatic Andy's offline, sadly.
22:57 pmichaud rakudo's spectest-progress.csv file has a regular list of commit numbers
22:57 pmichaud they're snapshots at the same time each day, so it's easy to produce a graph
22:58 pmichaud or even to reason about what's happening.  For example, 06-01 to 07-01 was r27986 to r28858, so we had approx ~900 commits in june
22:59 pmichaud 11-01 to 12-01 was r32261 to r33411, so there were 1200 commits in november
22:59 chromatic Good thinking.
22:59 pmichaud I've thought about putting a commit count line on the spectest-progress graph
23:00 PerlJam too bad it only starts in May :)
23:00 pmichaud I have another file that has commit numbers on regular monthly intervals going back a couple of years
23:00 pmichaud 33411-27986
23:00 purl 5425
23:00 pmichaud 5425 commits from jun through nov
23:00 pmichaud 5425/9
23:00 purl 602.777777777778
23:01 pmichaud 5425/6
23:01 purl 904.166666666667
23:01 pmichaud 904 commits per month, avg
23:01 jonathan Around 30 a day. Not bad momentum. :-)
23:03 chromatic Give us 5 FTE and double or triple that.
23:03 moritz FTE?
23:03 purl FTE is made by the guy that also coded icewm BTW :) or full time employee aka perm
23:03 PerlJam chromatic: depends on the developers
23:03 moritz well, give us the right 5 FTE ;-)
23:03 PerlJam chromatic: some are good at implementation, others are good at thinking.  Sometimes it's the thinking that's bottleneck.
23:04 PerlJam yeah, 5 FTE would be good no matter what :-)
23:06 pmichaud I have the commit numbers for 2007/2008, just a sec
23:12 moritz 4788 passing tests in rakudo right nwo
23:12 pmichaud so far the assign branch gives me 22 unexpected passing todo's
23:12 pmichaud (of course we will pass a lot more skips as well)
23:13 pmichaud how to save an OpenOffice graph as gif/jpg/png/whatever?
23:13 moritz and how many new failures? ;-)
23:13 Infinoid take a screenshot? :)
23:13 moritz pmichaud: maybe export as html, then it has to be converted to something like that
23:14 * pmichaud tries copy/paste into photoshop/windows
23:15 pmichaud I really should learn gimp someday.
23:18 Infinoid on a gnome desktop, the "PrtSc" key saves a .png of the current window
23:20 pmichaud aha, OpenOffice Draw can do it.
23:24 pmichaud http://www.pmichaud.com/perl6/commits-by-month.gif
23:24 pmichaud the trend line shows an increase in average commits by month over the 2-year period
23:26 * moritz can recommend the plotting tool xmgrace, though it's a bit tricky with date/time
23:34 TiMBuS joined #parrot
23:41 allison joined #parrot
23:53 mj41 pmichaud TortoiseSVN has own statistics  ... http://perl6.cz/wiki/Parrot_commits
23:59 dalek r33673 | jonathan++ | trunk:
23:59 dalek : [core] Implement remove_method on roles.
23:59 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=33673

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

Parrot | source cross referenced