Please test pre0: pugs.blogs.com/pre0/Perl6-Pugs-6.2.12.tar.gz | paste: sial.org/pbot/perl6 | pugs.blogs.com | pugscode.org | pugs.kwiki.org
Set by audreyt on 12 June 2006.
00:38 nnunley joined 00:45 clkao joined, helix6 joined 00:49 nnunley joined 00:54 helix7 joined, helix7 left, helix7 joined 00:57 helix7 left, helix7 joined, helix7 left 00:58 helix7 joined, helix7 left, helix7 joined 00:59 helix7 left, helix7 joined, helix7 left, helix7 joined, helix7 left 01:36 nothingmuch joined 01:57 frederico joined 02:05 justatheory joined 02:32 buu joined 03:01 buu joined 03:05 ArthurBrazil joined, justatheory joined 03:10 buubot joined 03:20 buubot joined 03:32 buubot joined 04:02 buubot joined 04:17 SamB joined 04:43 buubot joined
svnbot6 r10726 | cmarcelo++ | * Judy.Map: fromList and toList 05:02
pasteling "cmarcelo" at 200.232.236.25 pasted "IHash using Judy.Map (not working yet)" (133 lines, 5.3K) at sial.org/pbot/17791 05:22
cmarcelo audreyt: take a look at the paste when you have time. Judy.Map is working already, I was playing with Pugs code but didnt get too far. About CollectionM/MapM stuff I'm working on that, I'm thinking about simplifying a little the classes (in particular, integrate the necessary methods from Foldable in original CollectionsM), at least for now. 05:26
audreyt k. I just woke up 05:29
give me some mins
cmarcelo ok 05:31
audreyt what original CollectionsM? 05:33
cmarcelo ops
s/M// 05:34
Collection is child of Foldable, i was thinking by now just defining CollectionM (and take necessary methods from foldable)..
05:35 buubot joined
audreyt maybe even just implement Foldable and Transversable? 05:36
hm, bad idea 05:37
need to have -M suffix and restrict to IO
so nvm
CollectionM sounds indeed correct
cmarcelo i wasn't clear: Collection is "child" of Foldable. I'm thinking of doing CollectionM as "orphan".. not defining FoldableM and TransversableM, at least for now. 05:40
05:42 amnesiac joined
audreyt yup. 05:44
but if you reuse method names such as mapM fmapM mapM_ 05:45
then that's going to clash
but maybe that's okay. not sure, my brain havn't woken up yet
cmarcelo hmm, original collections already clashes with prelude.. 05:47
well, there's always "import qualified".. 05:49
gaal morning all 06:05
audreyt hey gaal 06:06
food, be back in ~30min, then I'll look at releng related items such as emit_yaml
my plane leaves at 5:20pm so that gives me ~7hr of pugswork time 06:07
gaal your coming flight brings you to redmond? 06:09
audreyt to taiwan
I sleep one night there
then immediately fly to redmond.
jet lag is going to be fun
06:10 buu joined
gaal yeah, i can imagine :/ 06:10
06:12 buubot joined 06:52 bjoern_ joined
audreyt back from breakfast but going to nap a bit first... 07:09
cmarcelo audreyt: (take a nap first, read later) [context: Judy.Map as IHash impl.] there are places (reducePad in Eval.hs) that use newObject, which may create a new IHash (so it depends on IO monad), but reducePad somehow complains about this. maybe is the liftSTM before the call, dunno. [cf. my last paste in pbot, outdated code but gives the idea].. any idea how to get off this? or am i messing more code than needed? 07:20
i'm going to sleep now. (i read irc logs on web, so it's ok to reply even when i'm off).. 07:22
&
07:28 Aankhen`` joined 07:30 carbon joined
carbon i will now proceed to ask a question i'm sure is on every faq 07:31
*ahem*
any estimations on when perl6 will be ready and released?
audreyt by christmas. 07:35
carbon of this year?
audreyt that we don't know. :) 07:36
a zeroth estimation is something alpha or beta quality this year, and a more feature-complete by next christmas, but it varies depending primarily on how many help we get. 07:37
07:41 LCamel joined 07:44 Soga joined, tcliou joined 07:48 |mjk| joined 07:49 bernhard joined 08:26 lisppaste3 joined 08:37 trym_ joined 08:55 larsen joined 09:08 iblechbot joined 09:16 penk joined 09:19 Aankh|Clone joined 09:20 lisppaste3 joined 09:32 mj41_ joined 09:40 bjoern_ joined 09:55 diablo` joined 10:09 chemaja joined 10:27 nothingmuch joined 10:32 diablo`` joined, diablo`` left 10:34 chris2 joined 11:13 drbean_ is now known as drbean 11:31 buetow joined 11:54 weinig joined
audreyt ...and I slept well past checkout time, only have 1hr to work now 12:00
fixed PGE::Hs to support adverbs
and corrected captures; emit_yaml next 12:01
wolverian sleeping is good 12:03
12:05 DaGo joined 12:09 szbalint joined
audreyt :s and :sigspace added to PGE 12:24
now working on the pugs side bridge
12:28 penk joined
gaal I've been meaning to merge genParseYAML and genParseHsYAML a little more 12:30
audreyt oh cool 12:32
wolverian heh, chmod -x chmod # oops 12:33
audreyt perl -e 'chmod' 12:34
to the rescue
wolverian hehe :)
gaal lol 12:35
12:46 elmex joined
audreyt ok, gotta run to airport now. bbiab 12:47
svnbot6 r10727 | audreyt++ | * Pugs.Parser: start parsing for "rule" and "token" specific adverbs. 12:48
12:50 marmic joined 13:03 cognominal_ joined 13:06 dolmen joined
nothingmuch seen ingy? 13:10
jabbot nothingmuch: ingy was seen 18 hours 13 minutes 53 seconds ago
nothingmuch hmm
audreyt: can i bug you about Module::Compile?
gaal nothingmuch: 15:47 < audreyt> ok, gotta run to airport now. bbiab 13:11
13:11 dolmen joined
nothingmuch oops 13:11
gaal I need help loosening this signature (or rank-2ing it or something): 13:12
nothingmuch ingy: ping?
gaal doGenParseYAML :: YAML a => (a -> IO b) -> Eval Val
I'm getting the error: Couldn't match the rigid variable `a' against `CompUnit' `a' is bound by the type signature for `doGenParseYAML' 13:13
so without grokking what I'm doing, I try
doGenParseYAML :: forall a. YAML a => (a -> IO b) -> Eval Val
but that gives the simple
Not in scope: type variable `b'
The task is higher-ordering the nearly identical genParseHsYAML, genParseYAML 13:14
which only differ in toYamlNode / showYaml respectively
which weirdly are 13:15
YAML a => a -> IO YamlNode and
YAML a => a -> IO String
come to think of it how did the CompUnit -> IO VStr work in the original? 13:18
13:34 myavuzselim joined 13:35 myavuzselim left 13:52 visq joined 13:54 visq joined 14:08 FurnaceBoy joined
lumi gaal: try "forall a b. YAML a => etc. " ? 14:20
14:20 Ymmv joined
lumi Quantifying explicitly over b 14:20
14:22 jsiracusa joined
mj41_ hello 14:25
Compiling Pugs.Prim ( src/Pugs/Prim.hs, dist\build\src/Pugs/Prim.o )
ghc.exe: internal error: update_fwd: unknown/strange object 0
Please report this as a bug to [email@hidden.address]
or www.sourceforge.net/projects/ghc/
Preprocessing library Pugs-6.2.12...
lambdabot Title: "SourceForge.net: The Glasgow Haskell Compiler"
mj41_ Building Pugs-6.2.12... 14:26
Build failed: 65024 at util\build_pugs.pl line 96.
14:31 justatheory joined 14:36 Limbic_Region joined 14:41 weinig is now known as weinig|coffee
nothingmuch nothingmuch.woobling.org/Module-Compile-TT/ 14:44
lambdabot Title: "Index of /Module-Compile-TT"
gaal mj41_: how much RAM do you have? can you vi config.yml and increase heap? 14:45
mj41_: you need to perl Makefile.PL after you do that. 14:46
lumi: that doesn't help.
lumi: when I remove the sig altogether I get an error that makes complete sense to me
14:46 mako132_ joined
gaal saying that you can't take a compUnit and turn it into a VStr 14:47
if it said you can't take a YamlNode and turn it into a VStr I'd also agree.
15:03 ludan joined 15:10 weinig|coffee is now known as weinig
visq overview question: is it by purpose that some things like the triangular comma are mentioned in the synopsis but neither implemented nor test-cased ? 15:17
how to find out what behaviour is intended eventually ? 15:18
TimToady visq: by necessity, things are specced before they are implemented, and often poorly specced before they are tested to see what the actual interactions are. 15:35
and we're all just volunteers, so often things get specced speculatively long before they're tested or implemented. 15:36
with something like triangular comma, it's not at all clear what the exact semantics should be yet. It's just a syntax that falls out of the other triangular infixes. But the spec is just saying it's available to make it mean something useful maybe. 15:38
visq Ok. Is there another obvious link between the synopses and the implementation other than the L<S?? tags ?
Has someone tried to annotate the synopses with test results based on those tags ? 15:39
Found it a bit hard to find out what actually works right now
15:39 soisoisoi joined
TimToady have you seen m19s28.vlinux.de/cgi-bin/pugs-smokeserv.pl? 15:40
lambdabot Title: "Pugs Smoke Reports"
visq Yes, already did that and it's very nice. But not directly linked to the Spec. 15:42
I think
gaal actually, it is, though we could use help with making the links richer.
TimToady occasionally some of the tests link back to the spec, but there's nothing official
gaal look for example in undef.t
scroll until you see a link and follow it
TimToady part of it is simply that the specs aren't as well structured as they would be if we were designing, say, Ada. 15:43
part of it is simply that volunteers aren't generally into that kind of accountability. 15:44
and like POD, we make it easy for people to contribute, even if it's suboptimal in various ways.
but as gaal sez, we'll take help wherever it's offered :) 15:45
even if it's actually improving things. :)
gaal visq: specifically to make this better, the'res magic in place that lets you do it w/o needing to edit the synopses
put a L<S> link in the test file, even to arbitrary places in the middle of a paragraph of the spec 15:46
visq I think it wouldn't be that hard to inject the relevant test results into the synopses .pod files
gaal visq: commits very, very welcome. we can do something like FIT 15:47
TimToady well, I don't know how many people would actually be interested in reading that, but certainly we could check coverage by backlinking.
FIT? 15:48
15:48 bjoern_ joined
gaal it's a way of embedding tests inside spec docs 15:48
Ran Eilam gave a talk on doing it with Perl and MS Word in OSDC
TimToady seems like a good way to discourage people from reading the specs, but probably there can be multiple ways to view the specs. :) 15:49
15:49 penk joined
gaal see search.cpan.org/~ingy/Test-FIT-0.11...est/FIT.pm 15:49
lambdabot Title: "Test::FIT - A FIT Test Framework for Perl - search.cpan.org"
TimToady thanks
gaal well, it is one way to read the spec and know "is this implemented? is this tested?" 15:50
visq thanks too.
15:51 penk joined
gaal visq: is this something you could be interested in looking into? 15:51
look at the smoke target and util/catalog_tests.pl
visq ahh, I see 15:53
TimToady don't let what I said earlier discourage you. good tools will motivate people to improve their cross-linkages if they can see pretty colors. 15:54
which reminds me, someone should put a "key" into the TAP matrix that explains what the various colors actually mean... 15:56
and I think it's a bit disengenuous to make "todo" items dark green... 15:57
purple, maybe... 15:58
visq maybe you know: make smoke does not generate the linked files like $PUGS_HOME/t/builtins/strings/length.html
for me
gaal someone should. till it's there: disengenuous dark green = todo; light green = OK; red = fail; yellow = unexpected success
TimToady I'm still not entirely sure what white means... Untested on this architecture? 15:59
gaal skipped
visq: perl util/catalog_tests.pl > /dev/null 16:00
TimToady but in a real sense, the test suite is becoming the specification. The S* are really intended primarily to just be deltas from Perl 5, whereas the test suite needs to be complete. 16:03
that's why it makes more sense to do the links from the testing end. 16:04
visq gaal: Didn't finish successfully. Says it cannot find /home/eric256/../code.tmpl :)
Will look at it later
gaal TimToady: one concern I have with that now that pugs has a lot of language working already, the tests use pretty expressive Perl 16:05
TimToady yep
gaal a new implementation will have to work hard till it can reuse our tests.
TimToady nod 16:06
visq TimToday: A test suite is no full spec, right ? And reading the Haskell code is a bit of a burden
TimToady I remember writing the original TEST program in a very limited subset of Perl 1. :)
gaal anyone remember where CPANPLUS keeps its global config?
:)
TimToady The test suite *can* be a spec if it's complete enough, I think. We're nowhere near that yet, of course. 16:07
but I don't know what "full spec" means. a spec is not a container that be filled. :) 16:08
s/be/can be/
visq Well, no one wants a *formal* spec, right (a bit hard to do as well) 16:09
TimToady I think the government defines full specs to be ones that break the scales you weight them on.
the problem with writing a formal spec is that you have to be omniscient to write one.
visq Still, I like the test suite, but it is not the best way to get started (synopses on the other hand give a good overview picture) 16:10
TimToady we're going for a more evolutionary approach, so the test suite is as formal as it gets.
at least the spec is not the C code, like it is with Perl 5.
visq Well, the semantics of the haskell core, should you dare to think about it, is kind of a formal spec
TimToady I think thinking about that too hard would damage me as a "language designer for the people" :) 16:11
16:11 amnesiac joined
TimToady I rely on smarter people to do that. 16:11
I love all the semantic feedback I get from the Haskell implementation though. 16:12
and I don't mind other people bugging me about areas where I think fuzzily, since that's basically all of 'em. 16:13
visq concerning util/catalog_tests.pl: Has anyone run this besides erik256 ? His home directory is hard-coded ;) 16:15
gaal it's on feather.. fixing 16:16
TimToady that's what symlinks are for ;p
gaal or relative paths :) 16:17
TimToady them's is too hifalutin' fer the likes o' myself 16:18
cognominal_ WHO THERE triangular comma was close to googlewhack
gaal I wonder whatever did happen to eric 16:19
fixed. 16:27
svnbot6 r10728 | gaal++ | * unbreak util/catalog_tests.pl by keeping the start path around when looking for the templates
16:28 visq_in_lederhos joined
TimToady you fixed eric? 16:29
gaal ah fixed 'im, all right 16:30
hmm, it used to be that the cataloger painted lines in the tests' HTML according to the test result. I wonder what happened to that too
maybe it eloped with eric 16:31
visq_in_lederhos gaal: catalog_tests.pl needs to be rewritten, propably
gaal how come?
visq_in_lederhos 1. sorry for the nick, my desktop crashed again 16:32
2. the referenced svn repo now has another layout svn.perl.org/perl6/doc/trunk/
lambdabot Title: "Revision 9686: /doc/trunk"
visq_in_lederhos my fault 16:33
weird command line argument handling though 16:34
gaal we can fix the command line handling w/o needing to rewrite the whole thing :) 16:35
by all means, hack on adding Getopt::Long to that though
got a commit bit?
16:35 lisppaste3 joined
visq_in_lederhos Ahh, it worked now 16:36
gaal: commit bit ? 16:37
gaal can you check in patches to pugs?
16:38 lollan joined, nperez joined
visq_in_lederhos I don't think so 16:38
gaal please send me your email, that'll change in a jiffy
hmm, it's marking everything as 'non_test', that's why the colors are wrong. 16:39
visq_in_lederhos: didn't get an email yet, if you aren't registered in freenode you can't send /msgs 16:44
16:48 FurnaceBoy joined 16:54 visq_in_lederhos is now known as visq
visq \msg gaal :works now ? 16:58
17:02 shachaf joined 17:04 cmarcelo left
gaal visq: invitation on its way to you. welcome aboard, and please update AUTHORS with your name :) 17:04
hey has anyone been using spork lately? its support for images seems broken all of a sudden 17:06
I have {image: localfile.jpg} and it isn't generating HTML for it
I'm pretty sure this used to.
ingy about? 17:07
nothingmuch if you find him i need him too 17:08
search.cpan.org/~nuffin/Module-Comp...tt.pm#TODO 17:09
lambdabot Title: "tt - Preprocess Perl code with Template Toolkit and Module::Compile. - search.cp ..."
pasteling "cmarcelo" at 200.232.236.25 pasted "IHash using Judy.Map (not working yet)" (133 lines, 5.3K) at sial.org/pbot/17801
nothingmuch urf lambadbot
gaal actually I like that feature
lowers pain of context switches 17:10
nothingmuch i'm just not used to it yet, i guess
gaal mooses randomly
nothingmuch ciao guys, bbl
gaal bye nothingmuch 17:11
17:12 cmarcelo joined 17:15 trym_ is now known as trym
gaal ah, solved the local image problem 17:19
i was sillily copying to slides/ and not to slides/images/
17:31 Limbic_Region joined 17:51 weinig is now known as weinig|out
kane__ gaal: global config is in CPANPLUS::Config 18:01
alternately, where you set $ENV{PERL5_CPANPLUS_CONFIG} to point to 18:02
18:02 nperez joined 18:15 nothingmuch joined
svnbot6 r10729 | lwall++ | Slaughter of :w/:words (the rx:w kind, not the q:w kind). 18:15
18:20 weinig|out is now known as weinig 18:22 mako132_ joined
ingy nothingmuch, gaal: pong 18:33
18:33 xpeed joined
nothingmuch ingy: search.cpan.org/dist/Module-Compile-TT 18:37
lambdabot Title: "&#1497;&#1493;&#1489;&#1500; &#1511;&#1493;&#1490;'&#1502;&#1503; / Module-Compi ..."
nothingmuch please look at the TODO
and comment on how i can improve on those features 18:38
ingy: stevan suggested that I should ride bitch on your bike 18:43
18:53 Aankhen`` joined 19:05 penk left
gaal ingy: I fixed my problem, never mind. just moose then 19:05
kane__: thanks
19:17 bjoern_ joined 19:34 weinig joined
ingy nothingmuch: sure, you can be my bitch any time 19:56
nothingmuch ingy: i'm more worried about Module::Compile though 20:00
ingy nothingmuch: what about? 20:01
nothingmuch did you look at the link?
ingy what's the problem? 20:02
nothingmuch well
a. use foo var => 'bah' kinda sucks
you have to eval the string afterwords
and you can't use lexicals from that scope
which is unintuitive 20:03
ingy backup
nothingmuch and b. you pmc_compile doesn't know what file/package/from line/to line it's filter
ing
it would be nice if those were passed in $_[1] along side $_[1]{use}
ingy can you say this in the form of tests?
nothingmuch a. not really
b. yes, but those will be sucky tests 20:04
a. is more of a please think about it thing
ingy I really don't have the context at all
nothingmuch well
ingy or maybe I do
nothingmuch my $var;
ingy but I don't get it yet
nothingmuch scratch that
let's say i have a filter
any filter
use foo extra_files_filter_cares_about => glob("foo/bar/*.file"); 20:05
this is a nice to have thing
but since everything after 'foo ' is treated as a string, i have to eval it myself in my filter
which is not 100% safe
is that intentional?
is that avoidable?
? 20:06
ingy not sure 20:07
nothingmuch okay, please think about it and once you make a decision i promise tests
ingy best to send me some cogent failing test
nothingmuch urf
but i don't know the right behavior
nothingmuch really hates this "send me a failing test" thing
it only makes sense when it makes sense 20:08
ingy just send a piece of code containing rm -fr /
that makes me see wtf you are talking about
I'm planning on hacking M::C today
so the timing is good
nothingmuch well... uh 20:09
i seriously don't know how to make this into a test
ingy just send me some code that makes me see the problem
I'm just not getting it yet
but I haven't been in this code in months 20:10
nothingmuch okay, i'll try
but i can promise you it won't be clear
ingy just something like
use tt files => 'system("rm *")'; 20:11
nothingmuch heh
ingy is that what you are talking about?
nothingmuch not exactly 20:12
or, not so clearly
do i have to provide the code for the filter?
ingy dude, I'm just asking for a clear explanation of the problem
nothingmuch one sec 20:13
ingy pretend that I did't write M::C
nothingmuch i'm writing
ingy because at this point I don't even remember what it does
nothingmuch okay
i'm sorry, i may have assumed too much
ingy right
nothingmuch no jokes about makeup sex
ingy :P 20:14
they say that's the best kind, but I'm a virgin so I wouldn't really know
pasteling "nothingmuch" at 84.108.53.74 pasted "<3 ingy" (9 lines, 247B) at sial.org/pbot/17811 20:15
nothingmuch ingy: is that clearer? 20:16
ingy much
it's all coming back 20:17
I'll think more about it 20:18
nothingmuch okay
now, the second thing is that i'd like $params, in that example, to contain more keys
like $params->{file}, $params->{package}, $params->{from_line}, $params->{to_line}
ingy sub foo::pmc_string_args { 0 } 20:19
nothingmuch $params->{file} is the file that 'use foo' was in, $params->package is the package it was in, etc
ingy perhaps
nothingmuch i don't really mind what the interface is
perhaps even provide both
ingy thing is I might not have access to the runtime values
I think that the way it is working now is mostly correct
but I'll try to give you what you want... 20:20
if it's possible
ideas welcome
nothingmuch well 20:21
it's not runtime
i thin
k
these are only called once, right?
ingy right... 20:23
nothingmuch oh, also
you can't do:
use foo (
blah => "Baz",
);
which people sometimes do
ingy ok
I don't know how to even run the params as an expression 20:25
nothingmuch i did it using:
s/^\s*use\s+tt\s*//
and then eval
which *really* stinnks
but works for very simple things 20:26
Limbic_Region ingy - did you see perlmonks.org/?node_id=554561
lambdabot Title: "IO::All is sliced bread"
ingy perl only calls the first use
Limbic_Region if you have any unmarried sisters - I would look out
nothingmuch ingy: oooohhhhhhh
so use foo is source filtered
instead of being an ->import trick?
or rathedr
the first time it's an import trick
and every time after that it's a source filter? 20:27
ingy Limbic_Region: praise for ingy code!!??
20:27 FurnaceBoy_ joined
nothingmuch you have a sisteR? 20:28
ingy guy is obviously a moron
Limbic_Region ingy - yes
"P.S. No, Brian Ingerson didn't pay me to say that, and I'm not married to his sister."
ingy I need to redo that module sans Spiffy
Brian Ingerson is dead
Ingy 2.0 -- The T-Shirt 20:29
Limbic_Region heh 20:30
20:46 lde joined 21:06 visq joined 21:10 penk joined 21:21 penk joined 21:27 mako132_ joined
xpeed is away: Busy 21:39
xpeed is back (gone 00:05:38) 21:45
nothingmuch ingy: want to reply to perlmonks.org/?node=556022? 21:50
21:50 Ymmv joined
lambdabot Title: "Not found" 21:50
marcus_ is bored 21:51
nothingmuch err, perlmonks.org/?node_id=556022
lambdabot Title: "Re: Module::Compile::TT"
22:05 weinig is now known as weinig|bbq 22:15 dolmen joined 22:26 cognominal_ joined
svnbot6 r10730 | benedikth++ | * fixed command line handling (sanity checks, GetOpt-Long) 22:58
r10730 | benedikth++ | * added help message
23:12 mako132_ joined 23:23 justatheory 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!