Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6, 2008-10-02

Perl 6 | Reference Documentation | Rakudo

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:00 TimToady it doesn't seem necessary to be able to look up a block by its label, anyway
00:01 TimToady on the other hand, if all implementations end up making it easy, who knows...
00:02 TimToady but the p6 model doesn't require it
00:02 TimToady kinda like it doesn't actually require continuations, quite...
00:25 s1n left #perl6
00:26 wayland76 joined #perl6
00:31 pugs_svn r22482 | lwall++ | [Cursor] fix macro bug induced by implicit {*} reductions
01:13 jdv79 left #perl6
01:39 charsbar_ joined #perl6
02:13 s1n joined #perl6
02:15 PZt joined #perl6
02:21 Ontolog joined #perl6
02:41 felipe joined #perl6
02:45 Alias_ joined #perl6
02:50 wayland76 Can anyone see the problem with this? ...
02:50 wayland76 role    testrole { method infix:<==> (testrole $self, testrole $other --> bool) { print "foo\n";}}
02:50 wayland76 ???
02:51 wayland76 (It gives an error in parrot 0.7.0)?
02:57 pmichaud rakudo doesn't support defining/overloading operators yet.
02:57 pmichaud beyond that, I don't know that infix:<==> will be a method.
02:57 wayland76 Hmm, that would do it :).
02:58 wayland76 I don't understand the bit about infix:<==> not being a method
02:58 pmichaud it's a multisub
02:58 pmichaud multi sub infix:<==> ( $a, $b ) { ... }
02:58 wayland76 Ok, got it :)
02:59 wayland76 I had that first, but it wasn't working :)
02:59 pmichaud right -- see "rakudo doesn't support defining/overloading operators" above :-)
02:59 pmichaud it should come in the next month or so, but we've got to do quite a few parser improvements to get there
02:59 wayland76 I'll keep an eye on the Rakugo page, then :)
02:59 wayland76 Ok, great :)
03:02 wayland76 (I've seen the ROADMAP file, btw, but only had a cursory glance over it, so I missed this -- stupido (that's me :) ))
03:04 yewenbin joined #perl6
03:09 eternaleye_ joined #perl6
03:17 pugs_svn r22483 | cjfields++ | [t/spec] added tests for checking number of times closures are invoked, ready for fudge
04:06 sail0r joined #perl6
04:07 Bzek joined #perl6
04:11 sail0r left #perl6
04:18 pugs_svn r22484 | lwall++ | [Cursor] add_macro couldn't handle macro name containing single quotes
04:35 alester joined #perl6
04:54 iblechbot joined #perl6
04:55 [particle]1 joined #perl6
05:11 Psyche^ joined #perl6
05:12 alester_ joined #perl6
05:43 [particle] joined #perl6
06:28 azawawi rakudo: say "Hi";
06:28 p6eval rakudo 31561: OUTPUT[Hi␤]
06:30 ashizawa joined #perl6
06:35 ashizawa joined #perl6
06:43 literal rakudo: "Hi".WHAT
06:43 p6eval rakudo 31561: RESULT["Str"]
06:44 literal rakudo: "Hi".WHAT.WHAT
06:44 p6eval rakudo 31561: RESULT["Str"]
06:57 azawawi is feather.perl6.pl down?
06:58 moritz_ it .nl, not .pl
06:58 moritz_ and it's up
06:58 azawawi hmmm
06:59 azawawi cant seem to reach it
06:59 moritz_ then you have a network problem
07:10 azawawi yup im down for today... DNS problems...
07:10 azawawi joined #perl6
07:10 moritz_ feather.perl6.nl has address 193.200.132.135
07:12 azawawi pinging doesnt work
07:13 azawawi so i guess it is not dns; it is a routing problem in my ISP
07:14 kst`` joined #perl6
07:22 hanekomu joined #perl6
07:28 simcop2387 joined #perl6
07:31 penk joined #perl6
07:47 rakudo_svn r31562 | moritz++ | [rakudo] implement regex and closure handling in Str.trans.
07:47 rakudo_svn r31562 | moritz++ | Patch courtesy by Chris Fields, cjfields++. Closes RT #59446
07:49 pugs_svn r22485 | moritz++ | [t/spec] unfudged trans.t for rakudo
07:54 rakudo_svn r31563 | moritz++ | [rakudo] more transliteration tests for spectest_regression
08:11 tomyan joined #perl6
08:15 moritz_ up to 4414 passing tests now
08:18 elmex joined #perl6
08:19 elmex joined #perl6
08:31 Liempt joined #perl6
08:31 Liempt left #perl6
08:32 meteorjay joined #perl6
08:45 viklund joined #perl6
09:02 barney joined #perl6
09:27 mtve joined #perl6
09:40 dmq joined #perl6
09:40 dmq left #perl6
10:52 penk left #perl6
11:01 iblechbot joined #perl6
11:17 ruoso joined #perl6
11:20 azawawi joined #perl6
11:39 azawawi @seen moritz_
11:39 lambdabot moritz_ is in #perl6. I last heard moritz_ speak 3h 24m 38s ago.
12:01 azawawi @tell moritz_ http://rt.perl.org/rt3//Public/Bug/Display.html?id=59552
12:01 lambdabot Consider it noted.
12:17 kanru joined #perl6
12:46 masak joined #perl6
12:48 abra joined #perl6
13:03 drbean_ joined #perl6
13:08 azawawi @tell moritz http://rt.perl.org/rt3//Public/Bug/Display.html?id=59556
13:08 lambdabot Consider it noted.
13:11 Lorn joined #perl6
13:19 kst`` joined #perl6
13:24 pyrimidine joined #perl6
13:35 aindilis joined #perl6
13:37 cognominal joined #perl6
13:38 masak rakudo: for {next}, {nothing}, {next} { $_(); say "whoa!" }
13:38 p6eval rakudo 31564: OUTPUT[whoa!␤]
13:42 masak rakudo: class A { method x { self++ }; method increment { say "I can haz one bigger?" } }; A.new.x
13:42 p6eval rakudo 31564: OUTPUT[increment() not implemented in class 'A'␤current instr.: 'postfix:++' pc 15846 (src/gen_builtins.pir:9887)␤]
13:42 masak :(
13:43 masak should I consider that a bug, or is it just Parrot internals poking out that makes it look like a bug?
13:43 jhorwitz joined #perl6
13:47 drbean_ joined #perl6
13:50 * ruoso sees the issue of "tree grammar" appearing on the list again, and that makes him remeber YAPC::EU::2007...
13:52 drbean__ joined #perl6
13:54 eternaleye__ joined #perl6
13:55 masak ruoso: why? (I wasn't there)
13:56 ruoso we had discussed this issue and we pretty much found out that XPath and XQuery are good enough... and that we could have a XPath grammar
13:58 ruoso because XPath might look like being specific to XML, but it actually isn't
13:59 masak no
13:59 wayland76 Hmm
13:59 wayland76 That's interesting :)
14:00 wayland76 So does that mean that you're recommending that we implement something XPath-like as a sublanguage?
14:00 masak it will happen
14:01 ruoso maybe not even XPath-like, but actually XPath
14:01 ruoso I don't know how to compose the XPath expressions
14:01 ruoso hierarchically
14:01 wayland76 Hmm.  I think there's room for improvement in XPath :)
14:01 ruoso like tokens and rules can
14:02 ruoso btw... the definition is not "dimension", but "axis"
14:02 wayland76 Anyway, I'm happy if the whole thing provokes a discussion about the whole thing fits together :)
14:03 ruoso one of the "axes" of XPath is attribute, which in P6 mean a method call anyway...
14:04 drbean___ joined #perl6
14:06 * ruoso taking the output of STD5_dump_match to sketch some ideas about that
14:09 drbean_ joined #perl6
14:10 ruoso ok... I just found out how to compose hierarchically XPath expressions...
14:10 ruoso <...> is not valid inside a XPath expression... that should mean a sub-path...
14:10 ruoso ;)
14:12 wayland76 I'm unsure what you mean by "hierarchically"
14:12 ruoso a xpath expression that requires another expression to be evaluated as part of that expressoin
14:13 wayland76 Oh, ok :)
14:13 ruoso and that might return an already transformed structure
14:13 wayland76 That's a new one :).
14:14 wayland76 One thing I'd like to see instead of XPath -- the filters in the [] part could be even more useful if they were in Perl :)
14:15 pasteling "ruoso" at 201.45.49.162 pasted "first sketch on what a xpath-grammar could look like" (6 lines, 131B) at http://sial.org/pbot/32425
14:15 ruoso in that example, the root element is dropped from the structure
14:15 ruoso and <statementlist> would be another token...
14:16 ruoso it would probably make sense to turn that expression into a relativepath...
14:18 wayland76 Not sure I understand either P6 grammars or what you're trying to achieve well enough to understand what's actually going on :)
14:18 ruoso let me sketch a little more
14:18 wayland76 But I was trying to implement an XPath variant in C recently, so I've got some grasp on its grammar
14:18 wayland76 :)
14:19 wayland76 (Btw, it's midnight here, so I probably won't be here for too much longer)
14:20 ruoso wayland76, please check the logs later then...
14:20 wayland76 Btw, ruoso, just got your e-mail, and will peruse :)
14:21 wayland76 (and my most recent message got crossed with yours in the ether, so I actually say some of the same things :) )
14:21 ruoso ;)
14:24 masak wayland76: what do you mean by XPath being too inflexible to be Perl?
14:25 drbean__ joined #perl6
14:25 masak I'm not trying to defend XPath, just curious
14:25 eternaleye_ joined #perl6
14:31 wayland76 inflexible?  Well, I guess the way I see it, each XML element represents an instance of a type...
14:33 wayland76 ...oops
14:33 wayland76 Sorry, I started talking about why XPath isn't the best for non-XML trees :)
14:33 ruoso is '[foo]?' the token for "match optionally the text foo without creating a capture?
14:34 wayland76 As far as inflexibility goes, it's basically just what I was saying above -- I want to be able to use Perl in the [] part of XPath, not just XPath expressions
14:34 pyrimidine joined #perl6
14:35 pasteling "ruoso" at 201.45.49.162 pasted "a little more ellaborated version of the sketch" (27 lines, 843B) at http://sial.org/pbot/32426
14:35 wayland76 Anyway, bedtime all, so gotta go.  Will try to remember to check the logs tomorrow or Monday, whenever I get back here
14:36 wayland76 ...ok, *after* I've looked at the pasteling :)
14:37 wayland76 ruoso: still unsure what you're demonstrating, but let me clarify my questions...
14:37 drbean___ joined #perl6
14:37 wayland76 it could be...
14:37 ruoso wayland76, that is based on the actual output of the STD match
14:37 wayland76 a) a proposal to be able to use XPaths in grammar
14:38 ruoso it would be able to turn the STD parse tree into other hierarchy of other types
14:38 ruoso in this case, turning the semantic-less parse tree into semantic "Module", "Statement" types
14:38 wayland76 ruoso: ok, I'll keep reading with that extra info in mind
14:39 wayland76 ...so we're talking about the transformation of a parse tree?
14:39 ruoso yes...
14:39 ruoso (btw... the par that lists <statement> should have a + modifier after it
14:39 wayland76 Is the parse tree just an example of a type of tree?
14:39 ruoso yes
14:40 wayland76 (just to check), so we could've been talking about file trees, or the tree of an XML document instead?
14:40 wayland76 (sorry, filesystem trees)
14:40 ruoso yes
14:45 sri_work joined #perl6
14:47 wayland76 I'm not sure about the syntax of some parts (eg. <<   text() = module >>, but I think I'm starting to get the idea.
14:48 wayland76 Just so I understand, am I right in understanding that the parts with the XPath syntax are a proposed syntax, rather than something that's possible now?
14:48 pasteling "ruoso" at 201.45.49.162 pasted "even more ellaborated version of the sketch" (29 lines, 1K) at http://sial.org/pbot/32427
14:49 ruoso wayland76, yes... it's a sketch of a proposed syntax
14:49 wayland76 Ok.
14:50 pasteling "ruoso" at 201.45.49.162 pasted "the last wasn't quite right" (29 lines, 1K) at http://sial.org/pbot/32428
14:51 wayland76 So how do we implement that?
14:51 wayland76 ie. how do we distinguish between an XPath and the division operator?
14:51 moritz_ I think / as an operator is already occupied too much
14:51 lambdabot moritz_: You have 1 new message. '/msg lambdabot @messages' to read it.
14:51 ruoso that's the purpose of the xtoken thingy
14:52 ruoso it defines a different syntax inside that block
14:52 wayland76 Ok, I see
14:52 moritz_ the general, friendly form of introducing a DSL is q:xpath{ .... } or something along these lines
14:52 ruoso hmmm... token doesn't have a q: in the front
14:52 wayland76 moritz_: We're proposing that it be used as a path separator as well as a division ... is that new?
14:53 ruoso moritz_, my idea is that the entire 'xgrammar' block is parsed by a different grammar...
14:53 moritz_ wayland76: it's in general a bad idea...
14:54 moritz_ wayland76: because in perl, the operators determine the context of the operands...
14:54 moritz_ wayland76: so current / always implies numeric context
14:54 ruoso moritz_, it's not Perl at all, so it's not the same operator
14:54 ruoso moritz_, it's not even the same grammar
14:54 moritz_ ruoso: ok, then... why not?
14:54 ruoso why not the same grammar?
14:54 moritz_ in lexical scopes everything is fair ;)
14:55 moritz_ that was a rhetorical question
14:55 ruoso ah
14:55 ruoso ok
14:55 ruoso ;)
14:55 wayland76 moritz_: The way I was origincally talking about (on the mailing list), the / would only be used as a path operator when it followed a Plex or Plex::Node
14:56 wayland76 (or array thereof)
14:56 ruoso wayland76, but it doesn't need to be seen that way... because the places that expect XPath won't be expecting Perl code
14:56 ruoso so there's no confusion
14:57 moritz_ so in $x = Plex.new; $x / 4; # will $x be numified? or 5 be plexified?
14:57 wayland76 Yeah, as long as we keep discussing the best way to do trees, I'm happy :)
14:57 moritz_ it sounds a bit like having + both for addition and string concatenation, which is a bit ugly, IMHO
14:57 yves joined #perl6
14:57 wayland76 moritz_: I would say that the left operator would determine what it was
14:59 moritz_ why?
14:59 ruoso does this discussion have a point? we're already considering a different grammar....
14:59 * ruoso runs
14:59 PerlJam he's a fan of arbitrary distinctions
14:59 wayland76 ruoso: I guess I'd still like to be able to use Perl in the filtering stage of XPath :)
14:59 ruoso wayland76, it's just a matter of defining that inside the [ ] it's Perl code
15:00 ruoso and that's all
15:00 ruoso problem solved
15:00 wayland76 ruoso: Ok, fine by me :)
15:00 ruoso no need to mix the syntaxes
15:00 wayland76 But then it's not XPath any more :)
15:00 ruoso fine fine...
15:00 ruoso XPath doesn't support the <subexpression> also
15:00 ruoso so, yes... it's not XPath anymore ;)
15:01 wayland76 moritz_: I'm not sure I agree, but ruoso has kindly invalidated any point I saw to the argument :)
15:01 wayland76 (ie. any point I saw in continuing the argument)
15:01 moritz_ well, all the hooks are in place
15:01 wayland76 in place for what?
15:02 moritz_ go write a library, and as soon as user defined operators and derived grammars work, profit
15:02 wayland76 ruoso's idea?
15:02 moritz_ wayland76: for all the stuff we've discussed so far
15:02 wayland76 ok
15:02 moritz_ wayland76: at least in the specs the hooks are in place
15:02 wayland76 I'm already working on a Plex library which models the objects
15:02 moritz_ extending existing types (like Hash, Array), adding DSLs, new operators...
15:02 wayland76 But it's pretty much useless now :)
15:03 ruoso wayland76, still, defining the semantics and syntax would be of great help...
15:03 wayland76 (ie. my code is useless, due to there not being much)
15:03 wayland76 ruoso: Agreed
15:03 alester joined #perl6
15:03 ruoso wayland76, would you be interested in continuing from where I stopped?
15:03 ruoso wayland76, I don't really have the time to work in that
15:04 ruoso and there are several semantic issues on how it's going to be parsed
15:04 ruoso and specially, in which node it is when evaluating which part of the expression...
15:04 wayland76 ruoso: well, not right now, but maybe monday or whenever I get back to Perl6 :)
15:04 ruoso one thing I'm quite confident is that every part of the path expression can be seen as method call
15:05 wayland76 Quite possibly
15:05 pyrimidine moritz_: (not to interrupt) thx for the .trans commit
15:06 ruoso wayland76, but my major concern is when to use the method in list context and when to use in scalar context...
15:06 ruoso i.e the 'statementlist' xtoken
15:06 moritz_ pyrimidine: thank you for the patch ;)
15:07 pyrimidine np
15:07 drbean___ joined #perl6
15:08 ruoso anyway...
15:08 * ruoso lunch &
15:09 ruoso wayland76, please consider extending this idea :) it will be much helpful in very short term
15:10 pyrimidine moritz_: btw, did we ever sort out how we would handle :g and other regex modifiers?  Didn't see it come up on the list, but maybe pmichaud asked in the weekly perl6 meeting...
15:10 wayland76 ruoso: Well, ok, I'll do that, but it may be a few days :)
15:11 [particle] yes, :g will be on the op, not inside the regex (for now)
15:11 [particle] m:g/.../ not m/:g.../
15:12 [particle] you'll see the design meeting discussion when chromatic posts it
15:12 pyrimidine that def. makes it easier to implement this in .match, .subst (i.e. pass named arguments)
15:13 moritz_ [particle]: so what will $str.subst(rx:g/.../) translate to on the side of the callee?
15:14 drbean_ joined #perl6
15:15 [particle] :g is an adverb on the operator
15:15 [particle] so, $str.subst(:g rx/.../) or something
15:16 moritz_ in inside the sub subst { ... } I'll just ask if $regex ~~ :g ?
15:17 [particle] :g is not and adverb on the regex
15:17 [particle] so you can't ask $regex ~~ :g
15:17 [particle] you have to ask .subst
15:17 moritz_ so subst has to ask self?
15:17 [particle] how do you introspect an op to tell what attributes have been set?
15:17 * moritz_ finds that very confusing
15:18 moritz_ [particle]: what about subs that can take multiple regexes?
15:18 moritz_ sub compare_matches($str, $re1, $re2) { ... }
15:18 moritz_ how will that know which adverbs are set on $re1 and $re2, if it's not stored in the regexes?
15:19 pyrimidine This is an example from S05: $target.subst(rx:g[pattern], { $() op expr })
15:20 [particle] i don't know. i was a fly on the wall in the discussion. larry and patrick are the two to ask for final answers, but i'm afraid it's still rattling around in larry's head
15:20 pyrimidine which to me indicates :g is part of the regex (not a named arg)
15:20 pyrimidine [particle]: ok, so it may change a bit
15:20 [particle] it may change entirely
15:21 [particle] i think the way to move forward is to implement something, and see what problems that causes/solves
15:21 moritz_ well, I'm sure that Larry will do a sane thing, and if not, we can convince him that insanity isn't appropriate here ;)
15:22 pyrimidine agreed.  It's wouldn't be hard to implement :g as a named argument in .subst
15:23 pyrimidine as for .match, I assume :g would get back a (lazy?) list of Matches
15:23 PerlJam It seems to me that this is a case where things that are different should look different.  we can use :i on the inside of a regex because that's tied to the regex.  But "global match" isn't a function of the regex, but how you apply it.
15:24 moritz_ aye
15:24 pyrimidine yes
15:24 moritz_ likewise :c
15:24 PerlJam I.e., having / :i foo /  work would make me expect  / :g foo / to work
15:24 moritz_ which looks kinda useless
15:24 PerlJam (I don't know what different should look like but I think they should be different syntactically somehow)
15:25 bennymack joined #perl6
15:25 drbean__ joined #perl6
15:25 PerlJam Luckily @Larry has said that non-alphanumerics are de facto "special", so maybe we could appropriate one of them for this use
15:31 drbean___ joined #perl6
15:37 hercynium joined #perl6
15:51 kanru joined #perl6
15:51 justatheory joined #perl6
16:02 azawawi rakudo: my $os = $*EXECUTABLE_NAME;
16:02 kst``` joined #perl6
16:02 p6eval rakudo 31565: RESULT[undef]
16:03 azawawi pugs: my $os = $*EXECUTABLE_NAME;
16:03 p6eval pugs: RESULT[\undef]
16:04 azawawi pugs: my $os = $*EXECUTABLE_NAME; say $os;
16:04 p6eval pugs: OUTPUT[␤]
16:27 azawawi is feather a bit slow today?
16:29 azawawi feather (ONE): tclsh -> 2478m VIRT...
16:30 azawawi @tell juerd feather (ONE) is a bit slow... i suspect tclsh (2478m VIRT) is the culprit...
16:30 lambdabot Consider it noted.
16:35 justatheory joined #perl6
16:36 azawawi does anyone have root access on feather1?
16:41 moritz_ I don't see such a job anymore
16:42 azawawi moritz: top and then see non-idle
16:42 azawawi pid: 10510
16:43 azawawi 92.9% MEM and growing
16:43 moritz_ killed
16:43 azawawi cool but after 8 hours it grew to 2.2GB
16:44 azawawi and the new tclsh is like 9.1% and growing
16:47 drbean_ joined #perl6
16:53 apeiron_ joined #perl6
16:55 Exodist joined #perl6
16:58 drbean___ joined #perl6
16:59 ab5tract joined #perl6
17:06 * azawawi wonders why tclsh is still taking 2.2g on feather1
17:11 exodist_ joined #perl6
17:11 apeiron joined #perl6
17:20 pugs_svn r22486 | lwall++ | [STD] more ACTIONS infrastructure, can now set $ACTIONS to a class in which
17:20 pugs_svn r22486 | lwall++ | methods are called according to reduction tag.
17:22 pmichaud I'd like to set up a mailing list to report pugs svn commits.  Would this be better done (1) via Google groups, (2) on my private server, or (3) other.  And once I set up the mailing list, would anyone be able to get svn on pugs to forward commit details to that list?
17:24 moritz_ I think it doesn't matter where the list is
17:24 kst``` joined #perl6
17:26 pasteling "azawawi" at 212.38.136.170 pasted "$*EXECUTABLE_NAME & $*OS globals.pir patch" (27 lines, 820B) at http://sial.org/pbot/32433
17:27 azawawi i can only get it to work with $*OS but not with $?OS
17:27 moritz_ they could, in principle, have different values
17:28 azawawi the thing is that with $?OS i simply get the PAST:VAR out of scope error
17:28 * [particle] prefers google groups
17:29 [particle] it's straightforward to set up svn to forward commits to an ml
17:29 * azawawi thinks that google groups is a good idea also
17:29 moritz_ pmichaud: I have root privileges on feather, so I could set up a mailing script
17:43 TimToady note to treematching folks: it is envisaged that signatures in a rule will match nodes in a tree
17:46 [particle] t\spec\S02-builtin_data_types\anon_block..................Could not find non-existent sub plan
17:46 [particle] pmichaud: any ideas?
17:59 drbean__ joined #perl6
18:09 ewilhelm azawawi, I found a 5.10.0-ism '//' in pugs's Makefile.PL
18:11 [particle] we can get rid of v6-alpha everywhere, yes?
18:16 zamolxes joined #perl6
18:19 drbean__ joined #perl6
18:34 ludan joined #perl6
18:37 pmichaud pugs-commits@googlegroups.com
18:38 pmichaud if someone can get svn to forward commits to that address, we can announce it
18:40 drbean__ joined #perl6
18:40 pmichaud azawawi:  $?OS is a compile-time variable, so it's not exactly the same as $*OS
18:40 pmichaud at present rakudo isn't doing much in the way of compile-time variables, although that will likely change soon.
18:41 azawawi i know but most tests in S02-magicals are using $?
18:41 azawawi pmichaud: how do i define a $? global variable?
18:41 pmichaud it has to be defined in the compiler -- probably as part of actions.pm
18:42 pmichaud basically we have to use the value at compile-time
18:42 kst``` joined #perl6
18:45 pmichaud so, if I have    my $x = $?OS;     then the compiler needs to switch that to be something like    my $x = 'Linux ...';
18:46 azawawi i see
18:46 pmichaud (except that $?OS should probably be some sort of object, not just a string.  We'll have to think about that a bit.)
18:47 azawawi so the change wont be as easy as i first thought...
18:47 pmichaud we could *fake* $?OS to act like $*OS for now, though, assuming that the compile-time value will be the same as the runtime value.
18:47 pmichaud But that sounds like a serious cheat.
18:49 azawawi the thing that i dont understand is how the PAST compiler sees $?OS as $OS but not $*OS
18:59 pmichaud I don't understand the question.
19:00 justatheory joined #perl6
19:06 ksem joined #perl6
19:06 ksem Evening.
19:07 ksem left #perl6
19:11 azawawi pmichaud: r u there?
19:11 pmichaud here.
19:11 azawawi im sorry i wasnt clear in my last question
19:12 pmichaud no problem.  :-)  Perl has lots of murky areas sometimes.
19:13 azawawi set_hll_global("$*OS",...) works perfectly but $?OS results in a scope error and it sees it as $OS
19:13 [particle] is the ? twigil implemented?
19:13 pmichaud I don't think we implement the ? twigil yet.
19:13 azawawi i dont know... that's why im asking ;-)
19:13 [particle] there you go.
19:14 azawawi cool
19:14 pmichaud I'm thinking we might cheat for now and make the ? twigil act like *, though.
19:14 [particle] works for me
19:14 pmichaud but that feels like an awfully big cheat.  :-|
19:14 [particle] until we're failing tests because of the cheat, all's fair ;)
19:15 azawawi we're failing a lot of tests because we dont support ? twigil
19:16 azawawi btw, where do we implement the ? twigil ??
19:22 Lorn joined #perl6
19:22 [particle] needs to be parsed in src/parser/grammar.pg, if it's not already
19:22 pmichaud I'm pretty sure it's parsed in grammar.pg already
19:22 [particle] then probably the variable method in src/parser/actions.pm needs modification
19:23 pmichaud right.  but ?-twigil variables are a little special.
19:23 pmichaud for example, $?LINE needs to be able to query the line number from the match object.
19:23 [particle] well, if you're copying *, then it's in variable, no?
19:23 pmichaud yes, if we just treat ? the same as *, then it's easy.
19:24 pmichaud but that doesn't work for many ? variables
19:24 pmichaud for example, there's not a $*PACKAGE global.
19:25 drbean__ joined #perl6
19:26 [particle] true
19:26 pmichaud so, things like $?CLASS, $?MODULE, $?LINE, etc. aren't represented in PAST as normal PAST::Var lookups.  Generally they become PAST::Val nodes of some sort.
19:27 pmichaud at runtime, $?LINE is a constant.
19:28 pmichaud anyone know what needs to be done to get pugs svn commit to post to the mailing list I just created?
19:28 [particle] need to add a post-commit hook to the pugs svn repo
19:29 [particle] google 'svn post commit email'
19:41 Exodist joined #perl6
19:41 rakudo_svn r31569 | particle++ | [rakudo] accept apostrophe (' and -) in identifiers
19:53 pugs_svn r22487 | particle++ | [t] remove use of external module
19:56 drbean joined #perl6
19:57 kst``` joined #perl6
20:02 drbean_ joined #perl6
20:12 drbean___ joined #perl6
20:17 pyrimidine joined #perl6
20:28 drbean__ joined #perl6
20:39 pmurias joined #perl6
20:39 pugs_svn r22488 | moritz++ | [t/spec] unfudge passing transliteration tests
20:41 pugs_svn r22489 | particle++ | [t] improve sub names in export tests
20:44 * pmurias is back home for the weekend so has internet access again ;)
20:44 moritz_ pmurias: did you notice that STD.pm got reduce actions?
20:46 pmurias looking
20:46 pmurias moritz_: hi
20:46 moritz_ hi pmurias ;)
20:48 * pmurias needs "This week in #perl6" ;)
20:50 drbean__1 joined #perl6
20:51 pmurias moritz_: having that builting instead of inheriting and using nasty hacks is nice
20:53 * moritz_ reports 4424 passing spec tests for rakudo
20:55 azawawi moritz_++
20:56 drbean_ joined #perl6
20:56 * moritz_ hates RT
20:56 moritz_ why does it redirect me to the public interface after successful login?
20:57 azawawi if you use the public url, it does...
20:58 iblechbot joined #perl6
20:59 moritz_ azawawi: could please send your patches to rt as attachments? either your mail client or RT wraps  the lines of inline patches
20:59 azawawi moritz_: hmmm let me check
20:59 pmurias what would be a good name for a STD parse tree to mold compiler?
21:01 [particle] mildew is mold growing on fabric
21:01 azawawi moritz_: gmail is the culprit...
21:02 azawawi moritz_: would u like a final any-str.pir with all patches applied?
21:02 moritz_ azawawi: no, one by one is fine
21:03 moritz_ azawawi: with p5cho(m?)p I'd like to wait until pmichaud has refactored the scalars/list/container stuff
21:03 moritz_ I'm reviewing/testing chomp() right now
21:04 pmurias [particle]: mildew seems a nice word
21:04 moritz_ uhm, wtf is that .newline method in S29-str/chomp.t?
21:05 moritz_ I've never seen a reference to it anywhere
21:05 azawawi moritz_: yeah
21:05 azawawi moritz_: i was surprised when i saw it....
21:06 moritz_ that's... weird. unspecced. evil.
21:06 drbean__ joined #perl6
21:09 azawawi moritz_: the patches are on RT as attachments
21:09 pugs_svn r22490 | moritz++ | [t/spec] chomp.t: deleted tests involving unspecced .newline method.
21:09 pugs_svn r22490 | moritz++ | Fudged for rakudo
21:11 drbean___ joined #perl6
21:14 rakudo_svn r31573 | moritz++ | [rakudo] implement Str.chomp, and add test to spectest_regression.
21:14 rakudo_svn r31573 | moritz++ | Patch courtesy by Ahmad Zawawi <ahmad dot zawawi at gmail.com>, azawawi++
21:14 rakudo_svn r31573 | moritz++ | Closes RT #59556
21:15 azawawi moritz_: thx
21:15 kst``` joined #perl6
21:17 * moritz_ wants a post-commit hook that automatically closes tickets when I say 'closes RT #d+' in the commit message
21:18 wknight8111 joined #perl6
21:27 pyrimidine moritz_: only place I could find .newline is in pugs/docs/Perl6/Overview/Data.pod
21:27 pyrimidine and it's def. unspecced (and quite possibly evil)
21:30 pyrimidine line 18:        .chomp    # Returns chomped string with a .newline property giving the newline removed
21:30 pasteling "azawawi" at 212.38.136.170 pasted "'make spectest' output" (161 lines, 10.9K) at http://sial.org/pbot/32441
21:33 drbean__ joined #perl6
21:35 azawawi moritz_: This bug can be closed http://rt.perl.org/rt3//Public/Bug/Display.html?id=57014
21:35 lambdabot Title: #57014: ./perl6 -e '1' gives error and segfaults
21:37 moritz_ azawawi: thanks, closing it now
21:39 zamolxes joined #perl6
21:39 moritz_ azawawi: if you want to get information about tests that are not in spectest_regression, you can also run 'perl tools/update_passing_test_data.pl'
21:40 azawawi moritz_: cool
21:40 * azawawi thinks about working on RT #49586
21:40 azawawi http://rt.perl.org/rt3//Public/Bug/Display.html?id=49586
21:40 lambdabot Title: #49586: [TODO] Write tests for perl6 -c/-v/-e/-h
21:43 moritz_ the chomp.t assume that there's a perl 5 like List.chomp form
21:44 moritz_ *tests
21:44 azawawi moritz_: Can't locate TAP/Harness.pm in @INC when running it on feather1
21:45 azawawi i noticed... let me check
21:45 * moritz_ iinstalls TAP::Harness
21:46 azawawi moritz_: you're right; only one signature 'our Str multi method chomp ( Str $string: ) is export'
21:46 moritz_ azawawi: please try again
21:47 * azawawi thinks it a very cool script
21:47 azawawi moritz_++
21:48 azawawi moritz_: one note though, it does not go through them in order
21:49 moritz_ azawawi: right, it just goes through them in the order in which they are found in the dir tree
21:50 azawawi hmmm actually it is a bit random... e.g. S16
21:51 Limbic_Region joined #perl6
21:52 moritz_ "found in the dir tree" doesn't mean "as ls returns them", but rather in the order that the inodes are stored on disk
21:52 azawawi ;-)
21:54 rakudo_svn r31574 | moritz++ | [rakudo] in tools/update_passing_test_data.pl, try to print out the number of
21:54 rakudo_svn r31574 | moritz++ | planned tests instead of the number of ran tests. That's only possible for
21:54 rakudo_svn r31574 | moritz++ | very new TAP::Harness versions, so be nice if it's not new enough.
21:55 jferrero joined #perl6
21:56 azawawi moritz_: t/spec/S29-str/chop.t has also unspecced stuff at the end...
21:56 Exodist joined #perl6
21:58 moritz_ azawawi: yes, that's rather obscure
21:58 moritz_ azawawi: especially the last two tests seem to assume that chop mutates the values in-place
21:59 * azawawi checks svn log
21:59 drbean__ joined #perl6
22:02 pugs_svn r22491 | moritz++ | [t/spec] remove two very obscure chop.t tests that seemed to assume in-place
22:02 pugs_svn r22491 | moritz++ | mutating semantics. azawawi++ for noticing.
22:05 azawawi they seem to come from t/unspecced/chop.t
22:06 moritz_ freaky, t/sunspecced/sort.t actually implements its own sort routine
22:06 azawawi moritz_: to start working on RT #49586, i need RT #59570 to be committed
22:07 azawawi moritz_: so that $*OS and $*EXECUTABLE_NAME work...
22:13 moritz_ azawawi: where is  $*EXECUTABLE_NAME specced?
22:13 azawawi Variable.pod
22:14 azawawi moritz_: docs/Perl6/Overview/Variable.pod
22:14 pasteling "moritz_" at 89.13.235.192 pasted "$*EXECUTABLE_NAME misbehaviour?" (5 lines, 153B) at http://sial.org/pbot/32442
22:15 moritz_ I'd expect that to print out './foo.pl'
22:16 azawawi hmmm but you ran it using #!./perl6
22:16 moritz_ compare with perl 5
22:17 kst``` joined #perl6
22:18 pasteling "moritz_" at 89.13.235.192 pasted "perl 5 equivalent" (5 lines, 103B) at http://sial.org/pbot/32443
22:19 azawawi $0 is not $EXECUTABLE_NAME (use English)
22:20 moritz_ right, that's $*PROGRAM_NAME
22:21 azawawi moritz_: anyway, i only found the following in parrot
22:21 moritz_ azawawi: if you go by perl 5 analogy, $*EXECUTABLE_NAME should be an absolute path
22:22 azawawi http://www.parrotcode.org/docs/ops/core.html
22:22 lambdabot Title: Parrot Core Ops - parrotcode
22:22 azawawi moritz_: search for interpinfo
22:24 moritz_ I'm half sleeping alredy, and shouldn't try to think about such things right now ;)
22:24 azawawi me 2 ;-)
22:25 pmichaud it's possible that parrot doesn't have a convenient hook to that information yet.
22:25 [particle] moritz_: i hope by the time you wake up, 'is export' will work in rakudo :)
22:26 azawawi maybe we can get it like _config() in perl6.pir
22:26 moritz_ [particle]: that would be great ;)
22:26 pmichaud well, that would provide the parrot config information, but that probably wouldn't hold the address of the perl6 executable
22:26 pmichaud but it might.
22:27 pmichaud okay, I see how to enable post-commit email messages in subversion; how does one get admin access to the svn repository for pugs?
22:27 pugs_svn r22492 | azawawi++ | [t/spec] removed comment about evil unspecced chomp.newline
22:27 [particle] you need root on feather
22:28 pmichaud I have root on feather
22:28 moritz_ pmichaud: either ask Juerd for access, or tell me what to do ;)
22:28 moritz_ oh
22:28 moritz_ you just need file-level access
22:28 pmichaud is feather the pugscode svn repo?
22:28 moritz_ yes
22:28 pmichaud ah.
22:28 moritz_ in /data/svn/
22:28 pmichaud /data/svn/pugs/ ?
22:28 moritz_ yes
22:29 ab5tract1 joined #perl6
22:30 moritz_ this host has a surprising number of host names
22:30 moritz_ perlcabal.org, parrotvm.org, pugscode.org, feather.perl6.nl
22:30 pmichaud okay, going for "forgiveness rather than permission"  :-)
22:30 [particle] well, parrotvm.org is now property of parrot foundation
22:31 [particle] we'll keep it pointed at feather until we decide what to point it to instead :)
22:31 azawawi [particle]: any idea how to get $*PID for perl6/parrot process?
22:33 [particle] hrmm...
22:34 pmichaud okay, pugs-commit should be enabled.  Now to do a commit and see if it works.  :-)
22:35 [particle] azawawi: not possible afaics
22:35 pugs_svn r22493 | moritz++ | [util/perl6.vim] coro is gone, macro rules!
22:35 tristes_tigres joined #perl6
22:35 tristes_tigres hi
22:37 moritz_ hi
22:37 moritz_ pmichaud: what's the google group?
22:37 pmichaud just a sec
22:38 [particle] pugs-commits@ iirc
22:38 [particle] pugs-commits@googlegroups.com
22:39 pmichaud yeah, that's it.
22:39 pugs_svn r22494 | pmichaud++ | Fix my cpan id in AUTHORS.
22:40 ab5tract1 are there any plans to write a(nother, newer) book on parrot?
22:40 azawawi http://groups.google.com/group/pugs-commits
22:40 lambdabot Title: pugs-commits | Google Groups
22:41 moritz_ ab5tract1: Whiteknight is updating p6pe these days
22:42 justatheory joined #perl6
22:44 * azawawi feeling sleepy
22:44 azawawi good night
22:46 pmichaud what's a good return address to use for the svn commit messages?
22:46 pmichaud (From: address, that is.)
22:47 ab5tract1 p6pe?
22:47 ab5tract1 this link is broken on the smop wiki page: http://people.debian.org/~ruoso/SMOP.pdf
22:47 pmichaud ab5tract1: p6pe is "Perl 6 and Parrot Essentials"
22:47 pmichaud an O'Reilly book
22:47 ab5tract1 okay. so there will be a new version soon-ish?
22:47 pmichaud I don't know.
22:48 moritz_ pmichaud: so far I haven't got a mail from pugs-commits
22:48 [particle] we're wearing our hacker hats, not our author hats lately
22:48 pmichaud moritz_: I still have to subscribe the svn commit bot to the group (or else open the group up to "anyone can post")
22:49 [particle] &
22:51 eternaleye_ joined #perl6
22:54 ab5tract1 [particle]: i dig it. i'd rather see parrot 1.0.0 than p6pe 2
22:54 ab5tract1 though i am definitely looking forward to the 4th edition of the camel book
22:54 moritz_ ok, first mail arrived
22:55 drbean___ joined #perl6
22:58 pmichaud I sent that mail manually -- now we need to see if it happens automatically upon a commit.
22:58 pugs_svn r22495 | pmichaud++ | [t/spec]  Add a note about Junction.values to TASKS.
22:59 pmichaud yay, it works!
23:00 moritz_ pmichaud++
23:00 drbean_ joined #perl6
23:06 pmichaud now to get a good procmail filter on it so I only see the good stuff.  :-)
23:10 justatheory joined #perl6
23:11 drbean__ joined #perl6
23:14 eternaleye_ joined #perl6
23:18 kanru joined #perl6
23:28 Exodist joined #perl6
23:31 drbean__ joined #perl6
23:32 ab5tract1 how does one leave a message? i saw it happen earlier
23:35 ab5tract1 wait, nevermind, its a wiki i can change the link myself :)
23:35 kst``` joined #perl6
23:37 drbean___ joined #perl6
23:43 Exodist joined #perl6
23:43 kanru joined #perl6
23:50 kanru3 joined #perl6
23:51 Khisanth joined #perl6
23:52 drbean__ joined #perl6
23:58 eternaleye_ joined #perl6

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

Perl 6 | Reference Documentation | Rakudo