pugscode.org/ | nopaste: sial.org/pbot/perl6 | pugs: [~] <m oo se> (or rakudo:, kp6:, elf: etc.) (or perl6: for all) | irclog: irc.pugscode.org/
Set by Tene on 29 July 2008.
ruoso needs a fresher brain to implement map... zZzZz 00:06
rjbs use.perl.org/~rjbs/journal/37357 00:12
lambdabot Title: Journal of rjbs (4671)
rjbs That's the entry I was asking questions in furtherance of.
literal "[...] multiple language. In explaining some of the issues I have with Ruby, both the scope of scope and the resolution of methods." 00:29
is that supposed to say "scope of scope"? 00:30
and also, multiple languageS
rjbs thanks :) 00:34
pugs_svn r22158 | s1n++ | [spec] updated bool to support the latest tests and Bool class. 02:48
s1n moritz_: go ahead and add S02-builtin_data_types/bool.t to the spectest_regression 02:56
rakudo_svn r30772 | chromatic++ | [Rakudo] Improved caller lookup algorithm in !OUTER; it's slightly more clear 06:17
r30772 | chromatic++ | this way. It also appears to clean up RT #58578 (reported by Moritz Lenz).
r30772 | chromatic++ | Neither Patrick nor I trust it completely though.
moritz_ rjbs: re use.perl.org, it's 'self', not '$self' 06:38
rjbs: in the Perl 6 example
rakudo_svn r30778 | duff++ | Make the expressions in a loop statement optional and refactor loop_statement 07:33
rjbs moritz_: thanks 11:20
rakudo_svn r30785 | pmichaud++ | [rakudo]: spectest-progress.csv update: 157 files, 3129 passing tests 12:35
moritz_ rakudo.de/rakudo-progress-nonstacked.png that looks nice ;) 12:36
rakudo_svn r30786 | coke++ | [cage] fix trailing whitespace causing smolder failures. 13:24
pugs_svn r22159 | moritz++ | [t/spec] add subset mmd tests 15:22
r22160 | moritz++ | [t/spec] fixed previous commit 15:24
Khisanth pugs: my @a=([1,2],[3,4]); @a[0;0..1]; 17:47
p6eval pugs: OUTPUT[*** ␤ Unexpected ";"␤ expecting "x", octal digit, "o", "b", fraction, exponent, term postfix, operator or "]"␤ at /tmp/xa4yI1WkYg line 1, column 27␤]
Khisanth o_O
moritz_ no multi dimensional arrays and indexes yet
and no slice context either
Khisanth but why is it expecting an oct? 17:48
or fraction or exponent those seem like strange things to be expecting there
or does p6 accept fractional array indices? :P 17:49
moritz_ when you write '@a[0', it's a valid continuation from that point to write an octal digit, I suppose
or an exponent 17:50
for example '@a[0**1]' would be a valid continuation with an exponent
TimToady it should not be expecting an octal digit unless it sees 0o 17:54
moritz_ that's probably a pugs anachronism
but given that all octal digits are also decimal digits...
pugs: say 010
p6eval pugs: OUTPUT[10␤]
moritz_ pugs: say 1+010 17:55
p6eval pugs: OUTPUT[11␤]
TimToady btw, I'm making a fudgable pge.t out of what you checked in 17:58
moritz_ cool 17:59
I already thought about adding #emit rakudo $test_number in these skip/todo junctions 18:00
erm, #?rakudo emit 18:02
TimToady I'm just translating it all to inline code with appropriate skip/todo directives
only comes out to about 2500 lines of code :) 18:04
including all the old data as comments
moritz_ so you're not even eval'ing? 18:05
TimToady I'm just using a variant of pge-tests.t that spits out code instead
it inserts evals where it needs them
(I hope)
some trickiness to get the quoting right 18:06
since the tests depend on being data rather than code in spots, sigh
already found a few bugs in STD with it :) 18:07
like it couldn't parse / a : b /
or <- [a..z]> 18:08
[particle]1 yay for tests! 18:09
moritz_ now I know why you want inline code ;)
TimToady and there are pervasive regex-as-string assumptions, like /foo # bar/ eating up final / 18:10
ruoso TimToady, do you plan to add Iterator (maybe a Role, instead of a Class) type to the builtin types? 18:12
TimToady oh, likely 18:13
rakudo_svn r30793 | pmichaud++ | [rakudo]: Update svn:ignore to ignore auto-generated 'Test.pir'
ruoso TimToady, right.. it's just because I'm assuming that in SMOP ;)
TimToady, btw... I'm assuming that an iterator in item context should return one item at a time, even if the effective "producer" (like a map block) returns two items in an iteration... it would only return the capture when in slice context... 18:15
pmichaud TimToady: are you using the rx_* files from parrot?
I'm not entirely sure of the status of the pge_tests file that is in t/spec 18:16
[particle] me neither, i haven't looked at it since moritz imported it
ruoso TimToady, actually... "an iterator in item context should return" means "my $iterator = something(); foo(=$iterator);"... which should probably be written as "an iterator in item context should have prefix:<=> returning one item at a time" 18:19
TimToady pmichaud: I don't know where moritz_ got pge_tests from, but that's what I'm translating 18:20
pmichaud if it came from the pugs repository, I suspect it's a bit out of date 18:21
[particle] what does svn log say?
pmichaud I can check in the rx_* files from parrot
[particle] updates his pugs wc 18:22
pmichaud oh, never mind -- the version I'm looking at now looks okay 18:23
TimToady *nod*
moritz_ yes, it was the freshly imported pge tests (sorry, should have made that clearer) 18:48
TimToady heh, /:s[]foo/ is no longer legal 18:53
pmichaud what is it now? 18:55
TimToady [] is not legal rx since it's null inside
pmichaud (I've been through too many syntax changes to remember accurately :-)
TimToady well, :s::foo is the next workaround in the tests 18:56
pmichaud ah, yes... at one time :s[] was meant to be similar to :s()
TimToady but I just changed that one to :s foo since it wasn't actually caring about the ws
bug :s() is presumably false
moritz_ pmichaud: if you implement Object.defined in trunk I can push some test changes that fix is ($stuff, undef,...) 18:57
TimToady *but
pmichaud moritz_: doing that now. I'm merging the failthrow branch back into trunk, and then I can easily do an Object.defined
moritz_ pmichaud: ok
TimToady yow, :perl5 as an internal modifier
TimToady maybe I'll end up with an entire P5 grammar embedded in STD :) 18:58
pmichaud I don't think that PGE ever implemented :perl5
I think we just skip it
TimToady STD gets upset when it sees things like \z
well, it was \Q in this case 18:59
ruoso have sketched some notes about "map" in www.perlfoundation.org/perl6/index....p_operator ... and appreciate very much any comment...
lambdabot Title: SMOP map operator / Perl 6
rakudo_svn r30797 | pmichaud++ | [rakudo]: Merge failthrow branch back into trunk.
TimToady funny it accepted \A, but I guess that's "not an alarm character" o_O 19:00
pmichaud ...nothing to be alarmed about. :-) 19:01
TimToady well, except that we should probably catch \A and warn about obsolete P5isms 19:01
TimToady (in the absence of :P5, of course) 19:02
well, I know how to implement :P5, I just don't wanna :)
PerlJam it's weird that we're parsing parts of a some other language just so we can warn.
moritz_ otoh it feels bad to reserve all these nice control sequences just to catch p5isms 19:03
TimToady could warn only if there's a subsequent \z
though they don't always occur together
but the \a match is relatively useless in any case 19:04
and the \A not-a-bell match is even uselesser
moritz_ aye 19:05
TimToady so I'll just remove \A as a P6ism, I think
pmichaud +1 19:07
perhaps get rid of \a also? 19:08
ruoso TimToady, could you please, when you have some free time, take a look at the wiki page I just posted above about map? I should get to implement it in the weekend, so it would be nice to have some sanity checking on it... ;)
PerlJam wonders how many times people have used 19:09
(I can't recall a single time I've used it)
ruoso cannot remember what is...
pmichaud I've used it a couple of times in strings, but I don't think I ever used it in a regex
PerlJam ruoso: and map { $^a + $^b } @stuff falls under the lazy slice rubric? 19:16
ruoso PerlJam, I forgot to mention that map will consume as many items of the .arity of the block signature... 19:17
ruoso PerlJam, but this specific example will take a iterator for @stuff in item context, and consume twice for each call to the block 19:18
pmichaud that's what rakudo's map is currently doing. 19:19
ruoso pmichaud, is rakudo using iterators already?
or you mean just by using two items from the list as arguments?
pmichaud rakudo doesn't make a strong distinction between lists and iterators 19:21
ruoso but it isn't implementing lazyness in map yet, right?
pmichaud no, but that's because we don't have a good gather/take model yet
and we don't really have lazy lists yet 19:22
ruoso pmichaud, in smop map lazyness won't be implemented in terms of gather/take...
using iterators makes that easier
pmichaud, have you seen the notes on the wiki page I've posted above?
pmichaud not yet -- been busy trying to get some stuff into rakudo for moritz
and I have to go pick up kids in about 3 mins
ruoso ok... but try to take a look at it later... it should help you in rakudo... 19:23
pmichaud looking now.... but conceptually I don't see how Iterator differs significantly from List 19:24
that is, if Iterator is a class. I could see List implementing an Iterator role, however.
ruoso pmichaud, that would mean that the state of the iterator would be stored in the List?
pmichaud no, it's simply that when we ask a List for an iterator, it returns a clone of itself. 19:25
ruoso oh... ok... 19:26
pmichaud unless, of course, the intention of the spec is that modifying a List also modifies any previously created iterators, in which case I'll have to rethink that
ruoso pmichaud, that's indeed an important point.. 19:27
pmichaud I haven't seen or heard any clear examples one way or the other
but it's a question of:
my @a = 1..5; for @a { .say; @a = (); }
...what gets output? 19:28
ruoso that's something spec related...
pmichaud exactly. I haven't seen it clearly laid out anywhere.
ruoso I mean... both ways could be valid...
pmichaud and the choice will decide the implementation of iterators in Rakudo :-)
ruoso pmichaud, but wait...
this example is not precise...
because you're replacing the list entirely 19:29
my @a = 1..5; for @a { .say; @a[3] = 1 }
that should be a better example
pmichaud sure, that's another possibility
ruoso pugs: my @a = 1..5; for @a { .say; @a[3] = 1 } 19:30
p6eval pugs: OUTPUT[1␤2␤3␤1␤5␤]
ruoso that actually looks natural to me...
pmichaud pugs: my @a = 1..5; for @a { .say; @a = (); }
p6eval pugs: OUTPUT[1␤2␤3␤4␤5␤]
ruoso the iterator points to the list which is no longer stored in the variable @a 19:31
pmichaud then why did @a[3] = 1; modify the iterator?
ruoso it didn't
it modified the list
pmichaud pugs: my @a = 1..5; for @a, 6 { .say; @a[3] = 1; }
ruoso as for is lazy...
p6eval pugs: OUTPUT[1 2 3 4 5␤6␤]
pmichaud pugs: my @a = 1..5; for (@a, 6 ){ .say; @a[3] = 1; } 19:32
p6eval pugs: OUTPUT[*** ␤ Unexpected ";"␤ expecting term postfix, operator or "}"␤ at /tmp/cO9OOho02P line 1, column 33␤]
pmichaud pugs: my @a = 1..5; for (@a, 6) { .say; @a[3] = 1; }
p6eval pugs: OUTPUT[1 2 3 4 5␤6␤]
Myoma pugs: my @a = 1..5; for @a, @a, 6 { .say; @a[3] = 1; }
p6eval pugs: OUTPUT[1 2 3 4 5␤1 2 3 1 5␤6␤]
pmichaud those all look wrongish to me.
ruoso pmichaud, @a,6 expanded the array at that time... not lazily 19:33
pmichaud no, it stringified @a
i.e., @a,6 was treated as a list with two elements
ruoso ah..ok
yeah...
that's simply wrong
pmichaud anyway, the question is whether the iterator takes a reference to the list (and modifications to the list modify what the iterator sees) or if it is a snapshot of the list at the time of creation 19:34
and with that I have to run.
ruoso indeed...
rakudo_svn r30799 | pmichaud++ | [rakudo]: Initial implementation of .defined for Object.
ruoso however, "snapshot" doesn't seem to fit well with lazy lists
pmichaud the snapshot can itself be lazy 19:35
it doesn't require eagerness
PerlJam but the other way seems action-at-a-distancey
pmichaud if I say my @b = (@a, 1..5, 8..10); @b can still be lazy with respect to @a
but changes to @a shouldn't cause @b to change
I'm late -- bbl 19:36
ruoso pmichaud, I can't really see how to make snapshots of a network flow... (considering a network flow can be seen as a list)
(which <== implies)
pmichaud, and that should apply to database-backed lists as welll 19:37
pmichaud, my @b = (@a, 1..5, 8..10) can only be lazy if you have COW... 19:38
pmichaud, which is not always possible... 19:39
PerlJam, by "other way" you mean which way? ;) 19:42
pmichaud anyway, I can argue it either way, so it probably needs clarification from p6l 20:04
(back, btw)
moritz_: Object.defined is now available 20:05
it might have some bugs... but that's what tests are for :-)
pugs_svn r22161 | moritz++ | [t/spec] correct test wrt undef 20:33
pasteling "moritz_" at 89.13.234.179 pasted "pmichaud: those tests fail wrt .defined" (37 lines, 2K) at sial.org/pbot/32095 20:35
moritz_ pmichaud: see paste
moritz_ goes to bed now 20:36
pmichaud looking 20:36
moritz_: yes, there's a problem with uninitialized scalars
I'll need jonathan to fix it
or to suggest ways to fix it
pugs_svn r22162 | lwall++ | [STD] bugs found by rx.t 21:23
r22162 | lwall++ | [rx.t] bugs found by STD
pugs_svn r22163 | lwall++ | [pge-tests] no longer needed but moritz++ anyway 21:28
pugs_svn r22164 | lwall++ | [rx.t] forgot to use Test and plan 21:32
Auzon TimToady++ # making bugs cancel each other out 21:33
s1n pmichaud: any work done to support smoke with those test results? 22:33
moritz_: unless you want me to make a bug, add S02-builtin_data_types/bool.t to spectest_regression 22:34
pugs_svn r22165 | Auzon++ | [spectest] tests for regexes and closures in .trans. Also one test to ensure that y/// is gone 22:36
Auzon Also, a ++ for whoever added the PGE tests. I assume that was the huge jump in tests today :) 22:38
s1n Auzon: i think that was moritz 23:04
www.parrotvm.org/svn/pugs/log/t?lim...;rev=22165
lambdabot Title: log: /pugs/t (Rev: HEAD, via SVN::Web)
Auzon moritz_++ then :) 23:05

Powered by Cro and the Raku® Programming Language.

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

Thank you!