pugs.blogs.com | pugscode.org | pugs.kwiki.org | paste: sial.org/pbot/perl6 | <stevan> Moose... it's the new Camel ":P | .pmc == PPI source filters!
Set by Alias_ on 16 March 2006.
00:17 Aankhen`` joined 00:26 Quell joined 00:27 linolium joined 00:28 linolium left 00:33 hexmode joined
svnbot6 r9758 | fglock++ | pX/PCR - new method 'perl5', returns a string that can be eval'ed into a new PCR object; fixed a test 00:42
00:45 fglock_ joined
fglock_ audreyt: PCR Makefile.PL has 'use inc::Module::Install' - is this ok? 00:54
svnbot6 r9759 | fglock++ | PCR - Pugs-Compiler-Rule-0.02-beta.tar.gz
fglock_ & 00:55
00:55 fglock_ left 01:11 mako132_ joined 01:23 vel joined 01:33 ColARes joined
ColARes hy 01:33
01:36 stevan__ joined 01:38 arcady joined 01:50 weinig joined 01:53 theorbtwo joined 01:58 GabrielVieira joined 02:51 justatheory joined 03:25 vel joined 03:28 azurot1 joined 03:36 christo joined 03:37 theorbtwo joined
svnbot6 r9760 | lwall++ | Various builtin clarifications such as s/undef/undefine/. 03:58
04:00 qu1j0t3 joined 04:35 Khisanth joined, cmarcelo joined 04:56 bewest joined
bewest coignard: greetings from myarray :-) 05:07
coignard: PM?
coignard: myarray said I should chat with you 05:14
05:20 cmarcelo left 05:23 FurnaceBoy_ joined 05:46 GeJ joined
wolverian hm, I forgot how lazy lists work. how would I define @fibs lazily? 06:33
06:38 KingDiamond joined 06:49 p5evalbot joined 07:17 drrho joined 07:30 nnunley joined 07:55 azuroth joined, marmic joined 08:03 FurnaceBoy_ joined 08:06 dduncan joined
dduncan has anyone gotten this error from svk when trying to sync with the pugs repository: 08:10
"Mirror source already removed."
I have a fresh install, and I decided to start with downloading just the revisions from Jan 31st 2006 and after, essentially the last cpan release and after 08:11
it got the first 200 or so, then stopped with that error
geoffb dduncan, I pulled a few 2-3 hours ago successfully (no errors); but I was only updating an existing wc, not starting a new one
Oh sorry, I got confuzzled, it was the AES repo I pulled cleanly from 08:12
didn't pull pugs at that point
dduncan attempts to resume give that error immediately following "retrieving log information ..."
geoffb though might as well, just to test ...
dduncan my initial request was: Retrieving log information from 8869 to 9760 08:13
geoffb attempting pull ...
dduncan it got up to 9152
geoffb Syncing svn.openfoundry.org/pugs
Retrieving log information from 9555 to 9760
It appears to be committing revs
dduncan will try again ...
same error 08:14
makes me wonder if openfoundry now has a limiter on it to prevent too many downloads?
eg, "you got 200+ revs, so wait awhile"
geoffb possible, though seems kinda silly
mirroring open SVN repos is a fundamental goodness 08:15
dduncan I did get 8869 thru 9152 before it stopped
geoffb That's not a round number.
dduncan I also posted on the svk list first, in case they know what that cryptic error message means
yes ... it may be by data size though
geoffb nodnod
dduncan however, I can still get to the site through web browser 08:16
anyway, its midnight; I'll try to resume tomorrow
thanks for replying
geoffb np, I'm (very) rarely here,
So it seemed the least I could do when I am. :-) 08:17
08:22 mj41_ joined 08:50 bsb joined
drrho quick question: where can I find more information about this Tree Grammar Engine (TGE)? 09:01
and: can it be used without the rest of the parrot tool chain? 09:04
09:05 iblechbot joined 09:11 chris2 joined 09:12 nnunley joined
bsb drrho: parrot/compilers/tge parrot/t/compilers/tge, no idea about the rest 09:22
09:26 jiing_mIRC joined, jiing_mIRC left
drrho bsb: thx found it, will investigate 09:35
TimToady this is silly--the only way I can talk to audreyt across the table in Tokyo is to send messages through California and Taiwan... 10:25
broquaint That's the beauty of the internet - you can talk to someone in the same room as if they were on the other side of the world. 10:27
10:37 elmex joined 10:45 amv joined 10:46 nothingmuch joined 10:57 KingDiamond joined 11:26 b_jonas joined 11:32 orafu joined 11:34 trym joined 11:52 bsb joined 11:54 kane_ joined 11:56 kane_ is now known as kane-xs 12:05 mj41_ joined 12:26 weinig joined 12:38 whiteg left 12:44 nothingmuch joined 12:47 ajani joined, broquaint joined, anatoly joined
gaal anatoly: }:) 12:49
12:57 fglock joined 12:58 Aragone is now known as Arathorn 13:11 theorbtwo joined
svnbot6 r9761 | fglock++ | PCR - added placeholder file 'lib/Pugs/Grammar/Category.pm' 13:11
13:22 xX_rep_Xx is now known as xX[ReP]Xx
svnbot6 r9762 | fglock++ | PCR - added test t/07-category.t 13:26
13:29 Limbic_Region joined 13:45 chris2 joined
Limbic_Region integral ping 13:45
svnbot6 r9763 | fglock++ | PCR/category - more tests 13:53
14:06 Qiang joined 14:07 ruz joined
svnbot6 r9764 | fglock++ | PCR/category - added more docs 14:07
14:10 kisu joined
integral Limbic_Region: pong 14:28
svnbot6 r9765 | fglock++ | PCR - removed opp.pl
Limbic_Region integral - the command to allow unregistered users to /msg me - is that a 1 shot deal or do I need to do that everytime I log in? 14:43
and what was the command again?
integral Limbic_Region: it's "/msg chanserv set unfiltered on". The setting is stored against your nickserv account, so it takes effect everytime
Limbic_Region great - thanks 14:44
integral no problem :-)
14:47 nnunley_ joined 15:02 eric256 left 15:07 justatheory joined 15:19 hexmode joined 15:20 miyagawa joined
fglock what happens if you declare 'proto xxx:<a>'? does it create a category 'xxx'? 15:41
Juerd proto? 15:50
fglock Juerd: sub prototype. almost the same as 'sub xxx:<a> {...}' - I want to know what happens when you use a new category name 15:53
Juerd Okay. Right. 15:56
What happen?
Prototype? :)
16:12 Limbic_Region joined 16:13 b_jonas joined
b_jonas join #lua 16:13
argh 16:14
16:29 Arathorn is now known as Ara4n
svnbot6 r9766 | fglock++ | PCR/category - added a failing test 16:53
17:12 bernhard joined 17:18 GeJ joined 17:21 DesreveR joined
svnbot6 r9767 | fglock++ | PCR/category - more tests 17:23
17:24 Ara4n is now known as Aragone 17:46 siosiosios joined 18:14 dduncan joined 18:17 larsen joined
svnbot6 r9768 | fglock++ | PCR - initial implementation of ':p' flag; category tests pass 18:17
dduncan okay, now that it's the next day, I'm going to try syncing with the pugs repo 18:18
still failed with "Mirror source already removed." 18:19
will try something else ...
that won't work as tried 18:21
18:23 ruz joined
pasteling "fglock" at 200.17.89.80 pasted "PCR/category - pseudo-code for usage inside a grammar" (8 lines, 209B) at sial.org/pbot/16518 18:29
18:31 fglock joined 18:37 ghenry joined
svnbot6 r9769 | fglock++ | PCR/category - cleanup, more docs 18:51
dduncan it seems that error is reported in SVN::Mirror, which isn't svk specifically, and is spat out of an exception is generated containing 'no item' when trying to download the revision deltas 18:52
fglock audreyt: ping
dduncan I tried something different, and it seems to work, though not sure why ... maybe bad data?
<break> darren-duncans-power-mac-g4:~ darrenduncan$ svk sync //pugs_try2/mirror 18:53
<break> Syncing svn.openfoundry.org/pugs
<break> Retrieving log information from 9153 to 9768
<break> Mirror source already removed.
<break> darren-duncans-power-mac-g4:~ darrenduncan$ svk sync //pugs_try2/mirror -s 9153
<break> Syncing svn.openfoundry.org/pugs
<break> Retrieving log information from 9153 to 9768 18:54
<break> Committed revision 334 from revision 9153.
<break>
so it seems using the explicit -s for the next revision, rather than it going there implicitly, makes it move
I think the difference here is that perhaps a revision had data that svn couldn't handle, and this approach made the 9152..9153 into a skip rather than a delta, or something, so different code was run than that which had a problem ... this is speculation 18:56
okay, now I'm up to date 19:05
19:09 pmichaud joined 19:14 FurnaceBoy_ is now known as FB|afk
svnbot6 r9770 | fglock++ | PCR - TODO update 19:18
fglock pmichaud: ping 19:20
pmichaud hello, fglock
fglock I'm trying to find out how to integrate grammars and categories - see sial.org/pbot/16518 19:21
pmichaud you mean things like infix:, statement_modifier:, etc? 19:22
19:22 hexmode joined
fglock the idea is that when you define a new sub/macro they will add themselves to the category 19:22
yes
pmichaud I was actually going to do that semantically when new subs/macros are defined with a special test, as opposed to trying to make it quite *that* automatic 19:23
i.e., if the sub name has a grammatical category in it, then figure out where it should go
buu Hrm. What is the status on p6-evalbot anyway? 19:24
pmichaud I haven't been thinking of grammatical categories as objects in their own right, except maybe there would be a correspondence to hashes somewhere 19:25
(it's possible that's flawed thinking on my part -- I just haven't gotten to that binding yet)
fglock I'm following audreyt idea to implement a 'fat' grammar
pmichaud I haven't seen that idea yet -- what's a 'fat' grammar?
fglock I'm thinking of grammar categories as a special kind of rule object 19:26
that's my words :)
the idea is to return an AST from the match
pmichaud I'm not sure the grammatical categories generalize quite that far. 19:28
fglock such that you can match->emit, with optional steps in the middle
pmichaud I'm getting ready to send an announcement to p6c, but I've checked in the PGE version of a Perl6 parser into languages/perl6 in the Parrot repository 19:29
svn.perl.org/parrot/trunk/languages/perl6/
fglock yes - I saw it
19:34 Corion joined, Corion left 19:44 ColARes joined 19:45 nnunley joined 19:57 davidra joined, davidra is now known as autark 20:07 larsen joined 20:08 penk left
svnbot6 r9771 | Darren_Duncan++ | ext/Rosetta/ : updated the Language.pod section on transactions 20:12
r9772 | fglock++ | PCR/category - first steps to return an AST in the category match 20:18
ColARes . 20:22
20:25 elmex joined 20:27 larsen_ joined 20:48 larsen joined
svnbot6 r9773 | fglock++ | PCR/category - partially reverted last change; no AST yet 20:48
20:49 kanru joined 20:51 SamB joined 20:52 FB|afk is now known as FurnaceBoy
svnbot6 r9774 | fglock++ | PCR/category status: current tests pass; uncomment line 117 of 21:06
r9774 | fglock++ | 07-category.t for a match-tree dump; the idea is to transform
r9774 | fglock++ | (internally) this tree into an AST; see SYNOPSIS for how this module
r9774 | fglock++ | could be used inside a Grammar
TimToady sub shower { atomic {... 21:12
audreyt "sub shower is atomic" is the current S17 form
fglock audreyt: hi! 21:13
audreyt fglock: hey! I saw your amazing work :)
fglock do you think the usage makes sense? 21:14
audreyt aye.
21:14 evalbot_9774 joined
audreyt also, have you tackled listfix aka 1 Y 2 Y 3? 21:15
(I've got that in the Hs version of OpTable)
listfix = infix+assoc<list>
fglock I think so - I've not tested yet :) - see the templates table
audreyt also, you need to commit eagerly for token-level parsing 21:16
iirc they never backtrack
21:17 justatheory joined
fglock 'infix_list' => '<tight> [ <op> <tight> ]+' 21:17
ok
audreyt well, not quite
that is the chain one
listfix requires the same op 21:18
fglock it is the same op - this is just the template
audreyt <tight> <op> [<tight> $<op>]*
fglock it is compiled to $op:=(<'...'>) 21:19
audreyt (where $<op> means something like \1)
oh ok... still, you need to allow for trailing comma
<tight> <op> [<tight> <op>]*
is still correct (I think all assoc<list> allows trailing.) 21:20
fglock comma is special? or can you have '1 Y 2 Y' ?
audreyt I think you could... there is no mention about Y in the canon, but trailing semi is allowed as well 21:21
pmichaud PGE allows comma (and semicolons) to have null terms. I dunno about '1 Y 2 Y'
fglock any problems are easy to fix, I think - just change the template (and pray)
audreyt I think it's better for consistency for all listfix to trail
otherwise we need another trait (shudder)
pmichaud I already think it's somewhat per-operator
PerlJam a trailing Y would be weird though
pmichaud is it possible to have something like (from S09) $a[3;;4] ? 21:22
audreyt PerlJam: true, though it wouldn't hurt either, or would it
PerlJam audreyt: I don't know, but I get the feeling that it may matter
I think comma and semi should be special
audreyt pmichaud: doesn't it just mean $a[3;4] ?
(both p5 and pugs currently collapses [1,,,3] to mean [1,2]) 21:23
fglock: so maybe the template is really
<tight> <op>+ [<tight> <op>+]* <tight>? 21:24
pmichaud I really think it needs to be per-op, unless we want trailing &, |, etc also
i.e., I don't think a trailing op is automatic simply because something is "listfix"
PerlJam audreyt: [1,,,3] seems like it would be syntactically useful as something other than [1,3]
fglock audreyt: I think so - I'll add some tests and try 21:25
PerlJam same for ;
audreyt PerlJam: I don't neccessarily disagree, I'm just describing current behaviour :)
pmichaud loop ($a;;$b) { ... } comes to mind
(depending on how one decides to parse that) 21:26
audreyt it's a special form via macro already...
it's not arbitrary statements* inside that loop ()
so not sure that applies
pmichaud s/statements/expressions?
semicolons are both expression separators (e.g., from S09) and statement terminators 21:27
audreyt expression. you are correct
but, my point is that
TimToady }}; shower()
audreyt loop ($a;$b;$c;$d)
is illegal
which means we are not using the general expresison prec parser inside the () 21:28
pmichaud depending on where we decide to trap the illegality, yes
audreyt s/using/rlelying on/
aye
pmichaud we don't have to necessarily trap it syntactically :-)
audreyt I'd prefer that, but sure :)
but I'm find with another trait to describe trailingness. it's just I can't quite see a case where implicit trailingness is going to hurt. 21:29
svnbot6 r9775 | fglock++ | PCR/TODO - added some examples to test later
pmichaud I'm uncomfortable with Y, &, and |
fglock needs to go :(
pmichaud although I can see arguments in favor of them all
fglock &
21:29 fglock left
audreyt my $choice = ( 21:30
$foo |
$bar |
$baz |
21:30 FurnaceBoy is now known as FB|afk
audreyt ); 21:30
pmichaud interesting, Perl 6 rules explicitly disallowed this sort of thing :-)
PerlJam different context though :) 21:31
pmichaud however I agree with you that having the null trailing op is a sane choice
I'm just not sure we want to commit it to all listfix ops :-)
anyway, not terribly important now 21:32
audreyt pmichaud: p6 rules allow it the other way though
rule choice {
| foo
| bar
| baz
}
pmichaud I thought that leading | was illegal
audreyt (as discussed in YAPC::Israel - still need to run it by you and commit to S05) 21:33
b_jonas there's a function form, isn't there?
audreyt yeah, Migo raised the issue in OSDC.il and lwall thinks it's sane
b_jonas my $choice = one (
$foo,
$bar,
$baz,
);
audreyt b_jonas: there is (any())
b_jonas and surely a trailing comma must work
pmichaud b_jonas: yes, you're correct there
b_jonas oh, it's any, not one
PerlJam is starting to think that everyone is insane
audreyt pmichaud: leading | makes things line up "correctly" :) 21:34
pmichaud audreyt: so, leading | is like a non-op?
audreyt pmichaud: right, in an alternation group, leading | is discarded
pmichaud same with leading & ?
audreyt sure. 21:35
pmichaud yes, we need to update S05
audreyt *nod*
ok, I_just_ got my translated slide now
PerlJam Would that it worked like that for listops ... Y @a Y @b Y :-)
audreyt and I need to finish it and leave for the conf in... oh, less than <2 hrs
so, I guess bbiab :)
pmichaud see ya 21:36
svnbot6 r9776 | Darren_Duncan++ | ext/Rosetta/ : incremented main version numbers following the v0.723.0 CPAN release 21:44
nothingmuch pasta.woobling.org:3000/smoke/list 21:49
smokeserver with comparison
tag support coming tomorrow
integral nothingmuch++
nothingmuch hopefully we can generalize runcores/platforms/whatever to tags
21:50 b_jonas left
nothingmuch the results inside it are C::P::Session with some dummy fails 21:50
get it here: dev.catalyst.perl.org/repos/Catalys...okeServer/
if anybody with mad html skillzz wants to fix it up a bit, please do 21:51
in the future i hope it'll be a two pane view thing
with test result lists on the left 21:52
the >> link popping a matrix on the right
and a + link to add any report to the matrix as a comparison
21:54 pdcawley_ joined
nothingmuch anyway, does anybody have any feedback/complaints/praise about the existing server that I should know about? 22:02
22:10 ruz joined 22:16 xerox_ joined 22:17 dduncan left 22:33 hcarty joined 22:34 hcarty left 22:36 theorbtwo joined, ruoso joined
audreyt flock: feel free to upload perlcabal.org/~autrijus/tmp/Pugs-Co....02.tar.gz 22:47
TimToady s/flock/fglock/ 22:55
nothingmuch good night 22:56
oh, audreyt: got any smokeserver wishes?
stevan__ hey audreyt, TimToady how is .jp 22:57
22:57 stevan__ is now known as stevan
TimToady As long as you're here, fglock, xxx:<+> has to be considered just a funny looking name. It's the grammar's responsibility (somehow) to pull in any existing xxx and newly created xxx and combine them into any rules or %hash that references them. 22:57
Today .jp is windy but clear, which is a refreshing change from the haze of the last two days. 22:58
pmichaud: see Timtoady[-2]. 22:59
stevan TimToady: cool, are you just all at the hackathon or the conf?
TimToady stevan: just about to go to the conf in 10 minutes.
supposing a grammatical category shows up in %xxx, then we need two ways to deal with it. 23:00
first, if we want one category to hide another, you can get away with a mixin style of 23:01
rule { %xxx | %yyy | %zzz }
but there are some syntactic categories that have to be magically combined like compile-time roles:
rule { %xxx_or_yyy_or_zzz } 23:02
that is, the longest-token rule is applies in parallel across all the categories simultaneously.
that's why the magic hash was invented (or more accurately, is scheduled to be invented :)
nothingmuch =) 23:03
i like that
TimToady well, I'd better go finish abluting before minnasama wants to go. 23:04
23:07 rafl joined
TimToady j 23:13
s/j//
svnbot6 r9777 | ruoso++ | closer to make lrep work again... 23:14
23:54 Shabble joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!