Summer of Perl | 6.2.12 released! | paste: sial.org/pbot/perl6 | pugs.blogs.com | pugscode.org | pugs.kwiki.org
Set by ingy on 4 July 2006.
00:03 sanug left
svnbot6 r11165 | putter++ | welcome.p6: Revert r11157, which no doubt broke the livecd. 00:06
putter clkao: It's been a while since I've looked at the p5 stuff, but I'll explore. Congratulations. 00:09
clkao putter: yay
audreyt putter: a good way to start is Class::MultiMethod::Dispatch 00:18
and extend it to support the notion of semicolons
00:18 Bit-Man joined
clkao what's this semicolon notion thing? 00:18
audreyt multi f (Int $x, Int $y; $z, $w) 00:19
clkao and this Class::MultiMethod::Dispatch thing isn't on cpan 00:20
audreyt sorry
Class::MultiMethod::Pure
thinko
clkao so after ; it is... ?
audreyt it is not used for tiebreaking purposes
putter: actually
(Int $x, Int $y; Int $z, Int $w) 00:21
is as good as
(Int $x, Int $y; Any, Any)
putter: we didn't cover that
i.e. any constraints after the final semicolon is not used for voting 00:22
clkao audreyt: might worth to document that prelude.p6 requires/expects for the internal backend to implement 00:23
is it christmas yet 00:24
svnbot6 r11166 | sanug++ | Renaming PHEME.html
r11167 | sanug++ | Renaming PHEME.jpg 00:26
audreyt putter: fixed 00:29
clkao: indeed, a rationalized "grep Pugs::Internals:::*" 00:30
00:31 tup joined
svnbot6 r11168 | audreyt++ | * mmd-draft.txt - add the exception that the parameters after the final 00:31
r11168 | audreyt++ | semicolon is ignored for the voting process.
rafl Why isn't the haskell module installed with 'make install' anymore? 00:34
audreyt rafl: it used to break with a per-user package.conf 00:36
putter audreyt: re CMMP, thanks. re after final semi...
svnbot6 r11169 | rafl++ | r168@ata: rafl | 2006-07-05 02:37:31 +0200
r11169 | rafl++ | * Fixed typos in Makefile.PL which prevented the detection of
r11169 | rafl++ | {installed,parrot,null}_config.o for installed parrots.
audreyt rafl: now maybe you can uncomment the Makefile rules
clkao audreyt: do you want to do some module::compile debugging?
audreyt actively want? no. to solve your problem? sure
putter the are still checked for compatibility, yes?
audreyt yes.
clkao audreyt: yes, solve my problem
putter is there a motivating example for this feature? 00:37
rafl audreyt: It's still likely to break as the generated installed-pkg-conf or whatever it is called contains paths to the build tree which fucks things when using the system-wide haskell module if the build-dir isn't present. 00:38
audreyt multi infix:<..> (Int $x, Int $y; $by)
multi infix:<..> (Int $x, Int $y; Int $by)
putter: well, "motivating" is maybe when you think about methods
rafl: yeah.
multi bark (Dog $x; Moving $target) { ... } 00:39
multi bark (Tree $x; Layer $n) { ... } 00:40
essentially the 2nd argument for each is "incidental"
putter so we want the infix:<..> example to be an ambiguity failure why?
(assuming handed an int)
audreyt sorry, that was merely a typo on my part. 00:41
hm.
otoh, I can't quite see why "always an implicit semicolon at end" is bad.
TimToady: help?
putter: the relevant spec: 00:43
Sometimes you want to have parameters that aren't counted as part of the
long name. For instance, if you want to allow an optional "step" parameter
to your range operator, but not consider it for multi dispatch, then put a
semicolon instead of a comma before it:
multi sub infix:<..>(Int $min, Int $max; Int $by = 1) {...}
The final semicolon, if any, determines the complete long name of a multi.
putter ok, pondering...
dduncan question ... does any .meta stuff work yet? 00:45
audreyt dduncan: only .meta.name
dduncan: the rest is pending The Arrival Of The Moose tomorrow
(=stevan)
dduncan so then I'll be able to say .meta.identifier.version tomorrow?
audreyt yes. 00:46
dduncan sounds good
audreyt in perl5 too.
I'll nudge stevan into releasing a new Moose with Moose::Meta::Module and Moose::Meta::Package
would be fun to express Rosetta with moose 00:47
dduncan fyi, I've decided to scale Relation back again to just a container, and not worry about subclassing at the moment ... I will also implement it out of things that work in Perl 6 and Perl 5 now, namely Hashes and Arrays ...
vsoni audrey: in PIR what is the difference b/w .lex .sym and .local
dduncan also, I expect to release a corresponding Perl 5 version on CPAN, using Moose
I will probably rename it to Set::Relation so it fits in better
how does that name sound to you?
audreyt vsoni: .sym is just .local 00:48
dduncan Rosetta will come after
audreyt dduncan: that sounds good
vsoni: .sym = .local; .lex lives on the pad, .sym does not
dduncan okay, I'm about to commit said name change now ...
putter so an f with (Int; Int) and (Int; Num) will always be ambiguous?
audreyt putter: yes, though you can always force disambig with (Int, Odd) 00:49
svnbot6 r11170 | fglock++ | v6 - simple option processing in s/// 00:50
putter is (Int, Num) and (Int; Int) on (3,4) ambiguous or variant first? 00:51
svnbot6 r11171 | Darren_Duncan++ | ext/Relation/ : first step of renaming Relation to Set::Relation and otherwise reworking it 00:53
r11172 | Darren_Duncan++ | renamed ext/Relation/ to ext/Set-Relation/, which should be less ambiguous and fit better into CPAN 00:57
clkao gr, i don't think fglock saw my changes 00:58
audreyt putter: variant first.
(Int;Int) is as good as (Int;Any) for voting.
putter but only for the final semi, yes? 01:00
audreyt parsefail 01:02
svnbot6 r11173 | sanug++ | 11159 perl6_compilation.graffle has now had each file exported to the perl6_compilation directory
r11173 | sanug++ | Added specific steps for Perl 6 pugs flow and Allison Randal's Parrot targets with specifics on future, present, and past languages.
putter in a signature with multiple semis, the final semi means something somewhat different than the others. 01:05
01:07 hikozaemon joined
putter a thought question might be "if we were using manhattan rather than second-to-none-everywhere to combine individual parameter distances, what would ; mean then?". 01:09
I'm am starting to be tempted to walk the design decision path just to get a better idea what the space looks like. 01:10
audreyt nod. 01:11
putter Or even better, look at a list of objectives, if there is one.
svnbot6 r11174 | rafl++ | r175@ata: rafl | 2006-07-05 03:11:42 +0200 01:12
r11174 | rafl++ | * Only use rpath if the build dir where the rpath points to really exists.
r11174 | rafl++ | r176@ata: rafl | 2006-07-05 03:12:51 +0200
r11174 | rafl++ | * Some debian/ updates towards 6.2.12 packages.
r11174 | rafl++ | r177@ata: rafl | 2006-07-05 03:14:25 +0200
r11174 | rafl++ | * Remove some obsolete files in debian/.
01:14 sanug joined
putter we have a story for individual parameter type distances, and one for how to combine them. we want to introduce inhomogeneities into the latter. 01:14
audreyt putter: have you committed v6-**? 01:16
putter oh, as an aside: might the second-to-none result in a "chase to the bottom"? authors narrowing their types to escape ambiguity, which creates ambiguity for others, who then... I haven't thought it through, but the possibility came to mind. 01:17
no. there was a note in t/README that the #! lines in test files help Test::Harness and prove. so I'm off doing a full regression, which takes an hour or two. 01:19
I could decouple the changes if you like? (hmm, probably should have done that the minute I hit the issue. ah well). 01:20
01:21 ludan joined
putter If yanking the #!'s from test files proves problematic, then the usefulness of use v6-** diminishes. 01:23
sigh. build goes boom. Cannot import `Data.ByteString': it was found in multiple packages: pugs-fps-0.7 fps-0.7 (imported from src/Pugs/AST/Internals/Instances.hs) 01:28
nothingmuch Perl::Generate dillema: 01:31
i have the subcall node
which is basically invoke as closure
it takes a code like thing (an expression) and optional args 01:32
if the expression is a bareword there is no ->
otherwise that stuff will be added
putter is unsure what "that stuff" is, but nods encouragingly... 01:34
nothingmuch anyway
i need a name for the type of the code thing
because Strs can coerce into barewords for that
Perl::Generate::AST::Node::Invokable? (ambig)
Node::Codeifyable? (sucks) 01:35
just Node::Bareword|Node::Expr ?
(which is redundant, because Node::Bareword does Node::Expr
putter so... the thing is a code or bareword? 01:38
svnbot6 r11175 | Darren_Duncan++ | ext/Set-Relation/ : rewrote part of the DESCRIPTION in Relation.pm to indicate the new focus and direction of this module
putter Node::CodeOrBareword?
almost end of day for me I think. 01:39
nothingmuch putter: hmm
putter: good night
btw, are you hackathoning with us tomorrow or later? 01:40
svnbot6 r11176 | fglock++ | v6 - rule/token/regex placeholders in parser 01:41
putter yes tomorrow, and also later, for some value of later. 01:43
nothingmuch were you here yesterday 01:44
i only met spinclad
putter I unfortunately do not have mobile computes, but I'm only a few minutes away, so will deal.
01:45 scook0 joined
putter yes. I was mostly in the other room, with audrey, gaal, and later spinclad. 01:45
nothingmuch okay
i thought i saw someone but was unsure 01:46
i mostly had my earphones on, and was being distracted
01:46 ivanfrey joined
putter I stepped in briefly to say hi, but didn't want to stomp on the non-hackathon work in progress, so it was low profile. I look forward to seeing you and knowing I've seen you. ;) 01:47
nothingmuch likewise =)
nothingmuch is not sure who is more insane
syd barret or ingy
putter any idea what time show starts? 01:48
nothingmuch fireworks?
ivanfrey Please forgive my ignorance but since I attended Damian's talk here in Toronto on Monday, I'm keenly interested in running pugs.
putter hopefully not, though that might be interesting. the hackathon tomorrow.
ivanfrey: welcome. let's see, do we have a page on download instr... oh, we just did a release, neat, one sec. 01:49
nothingmuch putter: probably ~10ish 01:50
putter ivanfrey: pugs.blogs.com/ has a download link.
nothingmuch maybe a little later
yesterday we started at around that time
but today wake up times varied from 8-somethingish to 13:00
putter nothingmuch: thanks :) 01:51
obra tomorrow, I need to be home from 8 - 12
y'all are welcome to head toward the office earlier
putter ivanfrey: after you download the search.cpan.org/dist/Perl6-Pugs-6.2.12/ release, there is a README and an INSTALL. 01:52
obra but non-bps hackathoners should make sure someone you know will be in the office before heading there.
lambdabot Title: "&#9786;&#21776;&#40179;&#9787; / Perl6-Pugs-6.2.12 - search.cpan.org"
ivanfrey Only problem is that when I tried to install the latest binary build of pugs on my PowerBook G4 running Mac OS 10.3.9 it was looking for a libperl.dylib that didn't exist on my system. I will check this link you've just posted.
obra surprising the office staff would be a not-good thing
nothingmuch what if we surprise them with party hats? 01:53
01:53 sanug left
clkao audreyt++ # fixing module::compile madness 01:55
putter: are you changing the tests? i can't wait to hook things up
ivanfrey putter: ok, I've already downloaded the source code. I was just checking to see if I could at all possibly avoid compiling, ;-) 01:56
putter ivanfrey: oh, curious. hmm, in about 14 hours there will be more people here, and thus greater potential for help. I'm not sure anyone mac-ish is here at the momemnt. 01:57
ivanfrey: right. :)
svnbot6 r11177 | audreyt++ | * revert fglock++'s last patch on Term because clkao claims that he had 01:59
r11177 | audreyt++ | it fixed already in a previous commit in a different way.
putter clkao: hit a snag, delayed until tomorrow. sorry. t/README says the test #!'s help Test::Harness and prove, so I was going to run a full regression test. But pugs HEAD doesnt build at the moment. 02:00
ivanfrey putter: sure, I'll wait. After all I haven't seen a more enthusiastic advocate for a language than Damian. 02:01
putter Would use the release to test, but out of time for today. Need a diff so you can play this evening?
nothingmuch clkao: ping 02:02
clkao nothingmuch: pong
svnbot6 r11178 | audreyt++ | * "use v6" now handles "use Test" without preloading the Perl5-land
r11178 | audreyt++ | "Test.pm", as we handle all block-rolling ourselves and refuse to
r11178 | audreyt++ | play with any Perl5-based Module::Compile precompilers "use"d within
r11178 | audreyt++ | this scope -- because our "use" doesn't mean the Perl5 "use" anyway.
r11178 | audreyt++ | * With this commit, this now works:
r11178 | audreyt++ | % cd pugs/perl5/Pugs-Compiler-Perl6
clkao nothingmuch: sighungry
svnbot6 r11178 | audreyt++ | % export PERL6LIB=../../ext/Test/lib
r11178 | audreyt++ | % perl -Ilib t/08-test.t
r11178 | audreyt++ | * We now complete 01-sanity/. All tests pass.
nothingmuch clkao: where's my v6 Test.pm with commented ugliness?
clkao nothingmuch: commented? not yet. 02:03
nothingmuch ;-)
it can be incremental
subetha?
oh wait, you don't use os
x
clkao audreyt: Use of uninitialized value in split at t/13-match.t line 5. 02:04
only in the first compile
putter good night all &
nothingmuch good night putter
audreyt clkao: I see that
vsoni audreyt: Saw your email response for PIR :) 02:05
audreyt vsoni: hi.
svnbot6 r11179 | audreyt++ | * don't warn when PERL6LIB is set to zero
audreyt vsoni: well, allison's point is well-taken; if PIR wants to be mixed-interleaved-compilation with full side effects -- which is to say, like Perl 5 and only Perl 5 -- I'm fine with that 02:06
I think it will make static analysis of .pir impossible, but allison also said she'll be okay with that.
so, I consider my job done :)
(and it'd be nice if you can proceed with submitting the remove-statics and hook-with-exceptions subtickets) 02:07
vsoni I agree......IMCC should be used to target other backends too.....
How bout running PERL on .net or JVM 02:08
A retargettable IMCC might come in handy
I will start working on the tickets......and keep you posted on progress 02:09
Audreyt: Let me know if you run into any other problems with IMCC in the mean time.
audreyt okay. 02:13
(I've posted (hopefully final) followup to allison)
<- learned to avoid any technical -- well, any kind of -- arguments with allison 02:14
clkao doh :$depends is not working yet. that means depends => $depends right?
or has that changed?
audreyt clkao: yes
clkao is there :@foo as well? 02:16
audreyt yes. 02:17
ivanfrey Anybody have pugs running on Mac OS 10.3.9? 02:19
audreyt not that I know of...
what error are you running into?
clkao audreyt: wait, it's depends => $depends or depends => 'depends' ? 02:20
audreyt depends => $depends
ivanfrey audreyt: All kinds of errors, with the latest binary builds from www.unobe.com/packages/, compiling pugs. I'm running into a real dog's breakfast. 02:23
lambdabot Title: "Index of /packages/"
audreyt ivanfrey: uhm, how about don't use binary build 02:24
and try getting a build of GHC?
haskell.org/ghc/download_ghc_641.html#macosx
lambdabot Title: "GHC: Download version 6.4.1"
clkao audreyt: Use of uninitialized value in string eq at lib/v6.pm line 41. 02:26
ivanfrey audreyt: yeah, i've installed GHC 6.4.1. But when I go to compile pugs, no runnable GHC found. Hmmm, I'm sure there's some option which will help me set the GHC folder. 02:27
audreyt: oops my bad there it is in INSTALL. 02:28
nothingmuch needs a pair programming buddy 02:29
motivational problems
any volunteers? 02:30
Perl::Generate semantics for operators
audreyt clkao: fixed 02:33
nothingmuch audreyt.find( :item<minion>, for => $?SELF ); 02:34
svnbot6 r11180 | audreyt++ | * more warning avoidance 02:35
02:46 kanru joined
ivanfrey Could not find module `Data.ByteString': 02:50
use -v to see a list of the files searched for
(imported from src/Pugs/AST/Internals/Instances.hs)
Build failed: 256 at util/build_pugs.pl line 120.
make: *** [pugs] Error 2
This is getting painful.
audreyt hm.
this is GHC 6.4.1, on the release tarball? 02:51
is it possible for you to try svn checkout svn.openfoundry.org/pugs/ and see if it works?
lambdabot Title: "Revision 11075: /"
audreyt I apologize for the pain. GHC641+OSX10.3 is as of yet an untested combination
ivanfrey audreyt: This might not have anything to do with GHC or does it? 02:53
audreyt ivanfrey: actually I'm pretty sure it does. GHC 6.4.1 ships with an older version of the Cabal library 02:54
which is the likely cause of this problem
rafl has checked in an alternative way of building this 02:55
in the pugs svn trunk
and so it maybe possible that it can be built easier that way
alternatively, if this is the release tarball
you can try
cd third-party/fps ; ../../Setup configure ; ../../Setup build ; sudo ../../Setup install
and then cd ../../ ; make 02:56
and see if it now works.
ivanfrey audreyt: ok, I'll try the alternative. 02:57
03:05 shachaf joined 03:08 shachaf joined 03:12 weinig|away is now known as weinig
ivanfrey audreyt: the alternative wasn't going to work because ghc wasn't in my PATH (/usr/local/bin/ghc). So I stuck it on my path. I now suspect that if I had done this from the beginning I wouldn't have gotten my last error. I'm still building, btw. 03:12
audreyt okay.
it might be very possible that is the problem. 03:13
svnbot6 r11181 | audreyt++ | * perl6_compilation.graffle - "Ponie 2.0" is not actively
r11181 | audreyt++ | developing -- it's planned for the future. Pirate
r11181 | audreyt++ | (Python on Parrot) is not dead; the list is active as
r11181 | audreyt++ | of June 25th; it seems just that they've branched the
r11181 | audreyt++ | Parrot tree to avoid the churn.
unobe ivanfrey: are you on panther or tiger?
cog tried to do install the latest images and it didn't work for him on panther
ivanfrey unobe: panther (10.3.9)
unobe hmmm..thought so
ivanfrey ubobe: You're the one who does the parrot and pug binary builds? 03:14
unobe apparently the linked libraries are different, so the images I made won't work (unless I can figure out what dependencies are needed)
nothingmuch hola unobe 03:15
unobe well, the ones on unobe.com/packages anyway
hey nothingmuch
you still in boston?
nothingmuch yep
nothingmuch is learning how to be an american couch potato
beer + sitting
no tv yet though
that's the advanced course
unobe hehe, welcome to the club :-)
how many days until you graduate to the advanced course? 03:16
nothingmuch i dunno
but clkao is a bad influence on my studies
we're going camping tomorrow
well, s/camping/hiking/ really
unobe ivanfrey: i still have panther, so I'll try to install it on my laptop, and build an image for panther later this weel 03:17
s/weel/week/
nothingmuch ooh! food
unobe ah sweet!
03:17 weinig is now known as weinig|sleeping
unobe which mountain range are you doing again? 03:17
ivanfrey: or you can tell me what libraries, and I can see if just copying them over to your comp will work for now. 03:18
from what i remember, they were just newer versions of libraries, so i think they can co-exist with the ones already present on panther 03:19
ivanfrey unobe: wait a sec, i'm on the phone. 03:24
unobe k 03:25
03:26 stevan joined
clkao sb end 03:29
unobe clkao: i just switched to irssi :-) 03:30
svnbot6 r11182 | clkao++ | Parse :$foo, which is a pair foo => $foo.
unobe i love /sb command, which, with my little knowledge of ircii, wasn't available in ircii 03:32
03:32 sanug joined
audreyt what does /sb do? 03:32
sanug: yeah :)
sanug Over here now ... newbie on irc - you were first reason for me to learn... 03:33
unobe audreyt: scrollback
so /sb home goes to the start of the session, /sb goto HH:MM:SS goes to a certain time in the past, and /sb end goes to the end 03:34
sanug What about the Main.p6 and CGI.p6 examples - those would seem unlikely candidates - yes? 03:35
03:35 nothingmuch_ joined
ivanfrey unobe: sorry 'bout that but the phone rang. I'm building pugs right now but I don't know what I'm going to get because I had lot of configure statements that said this was missing and that was missing and what not. Anyhow I tried to get the build I downloaded from your site running by searching for libperl.dylib and putting that where the build expected to find it. Bad idea because it just wasn't the right version 03:36
nothingmuch_ clkao++; # great food
nothingmuch: you need your modem fixed
audreyt sanug: unlikely candidates?
unobe oh wow! i just typed /help scrollback and the other stuff i didn't know about: clear, status, redraw 03:37
clkao nothingmuch_: glad you like it ;)
sanug Would CGI.pm build to javaScript?
unobe sweetness.
audreyt sanug: why would it not?
sanug: it involves no disk access
so persumably you can write a web server in javascript with it 03:38
and in fact, someone did that with javascript
sanug audreyt: the mind reels - I am still getting my head around this...
audreyt but well, Test.pm is a more neutral target 03:39
and you can s/CGI.pm/Test.pm/g with not much effort
clkao audreyt: help me get pc-p6 use lib ext/test/lib for make test ;0 03:40
sanug audreyt: Also the virtual machine on the left of the Perl6 to JavaScript - We would be using Parrot - or something else?
audreyt: s/left/right/
clkao is it almost christmas? 03:41
nothingmuch_ needs better minion locators
unobe ivanfrey: hmm...if i remember correctly, a new version of libncurses was needed
ivanfrey unobe: my build just failed. just hold on a sec while I'll post the error.
unobe you didn't have a problem with that?
k 03:42
ivanfrey Compiling Pugs ( src/Pugs.hs, dist/build/src/Pugs.o )
ar: creating archive dist/build/libHSPugs-6.2.12.a
*** Building: /usr/local/bin/ghc -hide-all-packages -package stm -package network -package mtl -package template-haskell -package base -package pugs-fps -package pugs-HsSyck -package unix -idist/build -Ldist/build -idist/build/src -Ldist/build/src -o pugs src/Main.hs -lHSPugs-6.2.12 -threaded
ld: archive: /Users/ivanfrey/Downloads/Perl6-Pugs-6.2.12/dist/lib/pugs-HsSyck-0.1/libHSpugs-HsSyck-0.1.a has no table of contents, add one with ranlib(1) (can't load from it)
ld: archive: /Users/ivanfrey/Downloads/Perl6-Pugs-6.2.12/dist/lib/pugs-fps-0.7/libHSpugs-fps-0.7.a has no table of contents, add one with ranlib(1) (can't load from it)
Build failed: 256 at util/build_pugs.pl line 217. 03:43
make: *** [pugs] Error 2
audreyt sudo ranlib /Users/ivanfrey/Downloads/Perl6-Pugs-6.2.12/dist/lib/*/*.a 03:44
and then "make"
svnbot6 r11183 | clkao++ | don't skip when test failed to run. This depends on audreyt++'s fix 03:45
r11183 | clkao++ | for make test to set PERL6LIB to ext/Test/lib to work.
nothingmuch_ is slowly discovering that types in Perl 6 are going to be much more useful than for just self documenting code and safety and speed 03:47
but in fact brevity too
audreyt ivanfrey: does that work? 03:48
clkao nothingmuch_: now you discover... 03:49
audreyt clkao: there you go.
nothingmuch_ clkao: i never thought i'd use coercions as much as I do now
and how much they actually help ergonomics
clkao nothingmuch_: you need more beer to understand many things 03:50
svnbot6 r11184 | audreyt++ | * Pugs-Compiler-Perl6 - Makefile.PL: Add PERL6LIB as requested by clkao
r11184 | audreyt++ | so that the 08-test.t always passes.
clkao audreyt: yay
so requestor doesn't get ++ on the comment log ;) 03:51
nothingmuch_ clkao: actually the beer is at the stage where it's slowing me down
clkao orz
system($^X, '-p', '-i.old', '-e', 's,PERL_DL_NONLAZY=1,PERL6LIB=../../ext/Test/lib,g', 'Makefile');
ivanfrey audreyt, unobe: well my build just failed, but for some other reason. But before I proceed further I have Perl installed via Fink in /sw/bin/perl and the build scripts are finding this version of Perl presumably because it's in my path. Is this bad? 03:52
unobe ivanfrey: export PATH:/usr/bin:$PATH 03:53
the perl that comes with 10.3.9 should be fine
03:54 justatheory joined
unobe oops. s/export PATH:/export PATH=/ 03:54
03:57 MenTaLguY joined
ivanfrey Could not find module `Data.Yaml.Syck': 03:59
use -v to see a list of the files searched for
(imported from src/Pugs/AST/Internals/Instances.hs)
Build failed: 256 at util/build_pugs.pl line 120.
make: *** [pugs] Error 2
I fixed PATH, but I'm getting this error now.
clkao audreyt: btw 10moose isnt compiling moose1.pm at all
weird
sanug audreyt: re: perl6_compile: Page on Perl 6 on Perl 6 - is it still a pretty printer as stated in the plan or now more advanced? 04:01
unobe ivanfrey: sometimes i've gotten that error. try running make again
audreyt ivanfrey: try the alternative path: 04:02
cd third-party/HsSyck/ ; ../../Setup configure ; ../../Setup build ; sudo ../../Setup instal
unobe or even ./configure again
audreyt cd third-party/HsSyck/ ; ../../Setup configure ; ../../Setup build ; sudo ../../Setup install
and then go back to "make"
on the toplevel
clkao, fglock: t/10-moose1.t fails here - known? or do I have an old moose? 04:05
clkao audreyt: no, it's not compiling moose1.pm at all
i suspect something module-compile-y
nothingmuch_: we should book the car
nothingmuch_ clkao: could you do that? 04:06
ivanfrey audreyt: in the ../../Setup build step there is a
nothingmuch_ i'm really not sure which you prefer
clkao nothingmuch_: i don't care
nothingmuch_ i'd say we should pick it up at 19:00 and return it at 23:00 04:07
clkao as long as we don't have to go all the way to the airport
nothingmuch_ i thought you had various milage deals
clkao don't seem to worth it
nothingmuch_ what's jesse's address? 04:08
Ibbetson?
23
nevermind
ivanfrey Argggg pressed return by mistake ... is a "ar: creating archive dist/build/libHSpugs-HsSyck-0.1.a" statement. Should I run ranlib on this file?
audreyt yes.
with sudo if needed
I think GHC 6.4.1 doesn't know about OSX needing to run ranlib 04:09
ivanfrey audreyt: GHC 6.4.2 is coming soon for Mac OS X. Do you think it's worth reporting this bug to them? Is it a bug? 04:10
vsoni audreyt: Any good sites to learn Haskell 04:11
audreyt ivanfrey: it is a bug, and I believe it's fixed; if you want to check, you can finish building pugs first 04:12
and upgrade your Cabal
nothingmuch_ clkao: herz at $125
audreyt and see if it's indeed fixed
nothingmuch_ pick up at 17:00
drop off at 11pm friday
.1 miles from jesse's house 04:13
audreyt Cabal 1.1.4 is available from haskell.org/cabal/download.html
nothingmuch_ hyundai accent
lambdabot Title: "The Haskell Cabal"
audreyt vsoni: check out the READTHEM file in pugs.
nothingmuch_ we can upgrade it for $5 to a midsize chevy 04:14
err, $7
or for $24 to a ford taurus
i don't think that's necessary, right?
ugh 04:16
it doesn't support .il for phone numbers
clkao: please do the reg yourself
i have the page open
clkao coming
04:18 sanug joined
audreyt ivanfrey: once you've got this working, please run util/package-osx.pl 04:19
unobe: do we have a build for 6.2.12-release? 04:20
for 10.4 powerpc
ivanfrey: so other people can get a binary installation without having to download or dance with ghc
04:20 guaumiau joined
clkao audreyt: oh the use lib is translated bad. fixing 04:22
ivanfrey audreyt: when I saw Damian's code examples, he was running pugs on Tiger I believe. Anyhow I've successfully built it. Just need to install it I guess, but wait there is a make test step. 04:23
audreyt ivanfrey: G4 CPU? 04:24
unobe audreyt: not yet. i remember some of the smoke tests were failing, and had uploaded the smoke report: xrl.us/oear
audreyt on single CPU G4, depending on memory, the test can take up to 3 hours.
lambdabot Title: "TAP Matrix - Tue Jun 27 22:41:20 2006 GMT"
audreyt unobe: right, but those are embparrot tests, no? 04:25
unobe ahhh...right. sorry bout that. i'm rebuilding as we speak.
audreyt ivanfrey: you can ensure that Pugs is basically sane by running
ivanfrey audreyt: mine's a G4, I don't know about Damian's.
unobe err...type
audreyt env PERL_HARNESS=./pugs prove t/01-sanity/*.t
if that passes, "sudo make install" and run the package script in utils should work 04:26
ivanfrey audreyt: Oh, you were talking about mine.
audreyt yeah
if you are going to run "make test", maybe download Bundle::Pugs::SmokeKit from CPAN first
and run a "make smoke upload-smoke" instead
so that other people can see the smoke test results 04:27
04:27 MenTaLguY left
unobe ivanfrey, audreyt: PREFIX, as far as I knew, wasn't working, so I had changed the Makefile.PL to install to /usr/local 04:28
audreyt ivanfrey: ivan dot frey at utronto dot ca?
unobe: yeah, rafl subsequently fixed that on trunk
unobe ah sweet
audreyt so maybe try trunk again?
rafl however havn't uncommented the "copy" lines from Makefile.PL so Pugs still wouldn't install as GHC libraries 04:29
ivanfrey audreyt: utoronto
audreyt ivanfrey: right. a commit bit is on its way to you
you can checkin svn.openfoundry.org/pugs/ using that usernamd/passwd and with subversion
lambdabot Title: "Revision 11075: /"
04:29 merlyn joined
merlyn in the current "v6.pm" in the cpan - Can't locate inc/Module/Install/Admin.pm in @INC 04:30
merlyn sighs
nothingmuch_ oi, bad make dist
audreyt merlyn: fixing
merlyn: fixed
reuploading
04:30 frederico joined
ivanfrey audreyt: please excuse my ignorance, but what's a commit bit? 04:31
merlyn 12.5 cents, same as in town. :)
sorry, old joke, bad joke
nothingmuch_ ivanfrey: the ability to check in code 04:32
a user in the SVN server
ivanfrey: if you have an email we can give you one
audreyt nothingmuch_: I already gave him one. 04:33
nothingmuch_ oh
audreyt ivanfrey: it is the login to write access to the pugs source repository with subversion
clkao audreyt: 10moose now passes. but if you see moose1.pmc, it's doing the wrong thing having the use lib for PERL6LIB - because it thinks itself being main, before it actually turned into "class main' 04:34
svnbot6 r11185 | clkao++ | Make 10-moose pass - emitting import list for use lib.
clkao audreyt: same thing for test.pmc 04:35
audreyt k
fglock: I'll rename v6-pugs to v6 in dist
fglock: also I wonder if it's okay to rename perl5/Pugs-Compiler-Perl6/ to perl5/v6/ 04:36
to save typing and all
clkao mmm 04:38
svnbot6 r11186 | audreyt++ | * "use v6-pugs" is now "use v6-**" for neutrality. 04:40
ivanfrey audreyt: do I have to run make install and /util/whatever before I run make smoke upload-smoke? 04:41
04:41 mdiep joined
audreyt no. 04:41
TimToady: "use v6-**" triggers -w 04:43
rather unfortunate
clkao :(
audreyt Argument "*main::*" isn't numeric in negation (-) at -e line 1. 04:44
clkao audreyt: 08-test.t no longer works 04:45
audreyt due to ** ?
ivanfrey audreyt: ok, the prerequisites of Bundle::Pugs::SmokeKit are Perl 5.6.1 or 5.8.3 whereas the Perl that comes with 10.3.9 is 5.8.1. This might be a problem. 04:46
clkao audreyt: ya :/ i don't know why
audreyt ivanfrey: oh? 04:47
hm.
where is such a prepreq decled?
ivanfrey audreyt: in the pod for this module. 04:48
audreyt aha.
nothingmuch_: your doing?
ivanfrey audreyt: oh sorry, I've misread the pod. 04:49
svnbot6 r11187 | sanug++ | Moving pugs/docs/perl6_compilation.pdf to perl6_compilation with the HTML and JPEG files
ivanfrey audreyt: it actually refers to Perl6::Pugs 04:50
audreyt clkao: fixed
nothingmuch_ audreyt: what did I do? 04:51
audreyt nothingmuch_: nothing much
sorry, false alarm
nothingmuch_ WHAT HAPPEN?
clkao whhot
nothingmuch_ SOMEBODY SET UP US THE BOMB
audreyt ok. now let's think about this ** business.
svnbot6 r11188 | audreyt++ | * Pugs::Grammar::Term - parse "use v6-**" as valid.
audreyt it seems that -_ and -() and -0 and -any and -Any and -pugs are valid 04:52
since -bareword and -digit works 04:53
hm, what to do what to do. 04:54
oh well, it can always wait another day. releasing as v6-pugs. 04:55
clkao: give us a Data::Bind without underscore please? 04:57
svnbot6 r11189 | sanug++ | Updated Perl 6 on JavaScript page to correct title
audreyt (can't release this on CPAN with an underscore dep)
clkao :/
should we do v6 release?
audreyt: also 08 might want to skip if there's no ../../ext/Test 04:58
audreyt clkao: also you actually require 0.09 of Devel::Caller
and also test::exception
svnbot6 r11190 | audreyt++ | * revert v6-** back to v6-pugs because, well, ** triggers perl -w,
r11190 | audreyt++ | and "make test" uses -w.
audreyt so register them and release D::B?
I'll get v6 packed up with a t/Test.pm in it
clkao sounds good 04:59
uploaded 05:00
svnbot6 r11191 | audreyt++ | * update dependencies. this be v6.pm 0.004.
r11192 | audreyt++ | * This be v6, 0.0004.
05:00 vsoni left 05:02 sanug left
svnbot6 r11193 | clkao++ | Data-Bind 0.25 05:02
05:03 nothingmuch joined
clkao audreyt: bah, you didn't make the test.pm 05:03
audreyt clkao: huh?
clkao 05:58 < audreyt> I'll get v6 packed up with a t/Test.pm in it 05:04
audreyt yes?
clkao but you release 004 without it ? 05:05
audreyt 005 :) 05:08
svnbot6 r11194 | audreyt++ | * v6 0.0005 to fix ChangeLog and mispackaging
clkao butstill no t/Test.pm ? 05:09
audreyt no way
I didn't check it in doesn't mean it's not in the pkg 05:10
:)
clkao is it in the manifest? ;)
audreyt merlyn: give the new Data::Bind and v6 a try?
why, it is
clkao intersting
audreyt merlyn: check pause incoming 05:11
clkao: I'm getting 05:14
t/07-tailcall......Use of uninitialized value in concatenation (.) or string at /Users/audreyt/tmp/v6-pugs-0.005/blib/lib/Pugs/Emitter/Perl6/Perl5.pm line 515.
Use of uninitialized value in concatenation (.) or string at /Users/audreyt/tmp/v6-pugs-0.005/blib/lib/Pugs/Emitter/Perl6/Perl5.pm line 515.
clkao ya. dunno what changed 05:16
will look after shower
but probably not important
audreyt $name is under perhaps?
s/under/undef
yeah.
clkao name shold be there
fixed 05:17
need multisub
gaal sup yo 05:18
05:18 ivanfrey joined
svnbot6 r11195 | clkao++ | Fix a warning emitting empty signature. 05:20
ivanfrey audreyt: How do I install Bundle::Pugs::SmokeKit? 05:22
audreyt ivanfrey: cpan Bundle::Pugs::SmokeKit
sudo as neede
d
gaal irc.pugscode.org/ is down? 05:23
(but colabti.de/irclogger/irclogger_log/perl6 is up) 05:24
clkao . o O (Bundle::v6)
ivanfrey GOT /Users/ivanfrey/.cpan/sources/modules/03modlist.data.gz 05:26
Going to read /Users/ivanfrey/.cpan/sources/modules/03modlist.data.gz
Going to write /Users/ivanfrey/.cpan/Metadata
Warning: Cannot install Bundle::Pugs::SmokeKit, don't know what it is.
Try the command
i /Bundle::Pugs::SmokeKit/
to find objects with matching identifiers.
gaal urp? do other modules install OK? 05:27
ivanfrey: try for example 'cpan YAML::Syck' 05:28
ivanfrey This was the tail end of the result of perl -MCPAN -e 'install Bundle::Pugs::SmokeKit'
gaal does this work? perl -MCPAN -e 'install "YAML::Syck"' 05:29
(have you installed cpan modules on this machine before?)
audreyt gaal: you going to join the 'thoners tomorrow? 05:30
afternoon that is 05:31
gaal thoning?
sure
audreyt @ obra's office
ivanfrey gaal: no I haven't installed cpan modules at all.
gaal i was just meeting two local friends today
audreyt cool. I'd like to get MM mapped
to P6AST
gaal good
audreyt very happy that Test.pm and 01-sanity/* runs from v6.pm land 05:32
ivanfrey audreyt: I'm getting this error:
gaal I can use some time near the eye of the insanitorm
audreyt which means we can deploy+test on both hs and p5 rapidly without worrying too much about surface scaffolding
ivanfrey Trying with "/bin/wget -O -" to get
ftp://ftp.perl.org/pub/CPAN/authors/01mailrc.txt.gz
--01:26:35-- ftp://ftp.perl.org/pub/CPAN/authors/01mailrc.txt.gz
=> `-'
Resolving ftp.perl.org... 192.75.95.80, 142.132.65.108
gaal audreyt: yes that is quite spectacular! CBS doesn't know it, but the fireworks were for that?
ivanfrey Connecting to ftp.perl.org[192.75.95.80]:21... connected.
Logging in as anonymous ... Logged in!
==> SYST ... done. ==> PWD ... done. 05:33
==> TYPE I ... done. ==> CWD /pub/CPAN/authors ... done.
==> PORT ... done. ==> RETR 01mailrc.txt.gz ...
audreyt gaal: might be
ivanfrey Error in server response, closing control connection.
Retrying.
mdiep so if Perl6 has n-backends, something like DBI will require n implementations? (even if they share most of the underlying code)
audreyt ivanfrey: that's... usual.
mdiep: yes, very good analogy
gaal does that mean your net setup is unhappy with FTP active mode?
ivanfrey: try an HTTP mirror? 05:34
audreyt ivanfrey: export FTP_PASSIVE=1
and try again
s/export/setenv/ if you are on tcsh
for my life I never understood why that is not the default.
gaal 2nd that :(
audreyt actually I do understand 05:35
gaal audreyt: actually I think there are versions of Net::whatever it is that ignore that setting
audreyt some early broken servers did not impleent PASV
but all servers (allegedly) support active
gaal ISTR I had to hack a hardcoded value in the installed module :(
audreyt so the one is default
gaal: LWP.
gaal yeah but of course not all firewalls do 05:36
audreyt LWP::Protocol::MyFTP
I had to
my $self = $class->SUPER::new(@_, Passive => 1) || return undef;
gaal do new perls ship with something that speaks http?
audreyt gaal: not afaic
gaal perl--
that makes new installations fiteen times more difficult than they need to be
audreyt yeah. 05:37
though, Perl 5.10 would ship with CPANPLUS 05:38
which does force passive.
ivanfrey audreyt: same error doesn't know what Bundle::Pugs::SmokeKit is.
gaal ivanfrey: do you maybe have wget installed? 05:39
audreyt ivanfrey: hm. type
sudo cpan
and then
force install YAML URI MkDoc::XML Petal Test::TAP::Model Test::TAP::HTMLMatrix Best
and see if that works?
gaal audreyt: I think he doesn't have the index 05:40
so first order of business is fixing transport for him.
audreyt oh right
gaal ivanfrey: if you do have wget, please try:
audreyt have to force reload.
gaal cpan
> o conf urllist unshift www.perl.org/CPAN/ 05:41
lambdabot Parse error
Title: "CPAN"
gaal > o conf commit
lambdabot Not in scope: `commit'
gaal > reload index
lambdabot Not in scope: `reload'
gaal shush lambdabot
and then try the d/l again
if you do NOT have wget installed, you need to go to search.cpan.org and manually d/l a library that does http 05:43
ivanfrey gaal: No, I have wget. 05:44
gaal whew (geting lwp w/o something that fetches dependencies for you is painful)
gaal has done it several times on broken windows boxes
ivanfrey gaal: there's a newer version of cpan I'm installing now. 05:46
clkao audreyt: does devel::caller test pass for you? 05:49
audreyt: it's failing bizzarely for me. got and expected are the same
audreyt not on win32 05:50
clkao darwin?
audreyt testing
same failure mode. hmm
clkao you meant you didn't test when you installed it?
ok, i am pinging richardc
audreyt my cpanp is set to perpetual skiptest.
clkao ... 05:51
orz
gaal sure speeds it up!
clkao can you do that in bundle file?
audreyt gaal: exactly!
clkao: no
clkao nothingmuch: don't fall asleep hugging your laptop
gaal clkao: too late
it would be cool to intercept the packets from last.fm and send him a wake up call 05:52
zzZ& 05:57
ivanfrey audreyt: the Bundle::Pugs::SmokeKit install is asking whether I want the EUMM installer or the MB installer. 06:02
audreyt you want EUMM, I think.
shower, bbiab & 06:04
gaal audreyt: maybe nothing at all is not as good as a mock shebang 06:13
audreyt #!mock
?
gaal # this is a perl6 program
not recognized by perl or unix
audreyt #!perl6 ?
gaal no, not a shebang
audreyt $ echo '#!moose' > moose ; perl moose 06:14
Can't exec moose at moose line 1.
gaal because #!perl6 would be bad too
audreyt indeed
gaal sorry, "mock" was the wrong word
audreyt that's another thing for bleadperl to address
gaal non-#! line but still a comment
audreyt well
use v6-pugs
would be that comment
I think.
use v6-perl; # unless it's this to be more explicit 06:15
but we can sort out the -suffix mess tomorrow; it was -pugs, let's stay -pugs for tonight :)
gaal if only because some day we'll want (maybe) to perl -pi -e to another real shebang 06:16
and having a neutral marker will make that easier
w/o falling into files that aren't tests under t/
audreyt well, we'll do a 06:17
gaal "# this is a perl6 program" works
ivanfrey audreyt: I'm giving up for tonight. Gotta get some sleep.
audreyt perl -pi -e 's/^use v6-pugs;/#!realshebang\n$&/'
ivanfrey: that is a wise choice
ivanfrey: at least we made progress...
ivanfrey: sorry for the pain :)
gaal wonders why B:P:SK has an installer at all, being a bundle, but it being 2:18 precludes his further inquiry 06:18
audreyt suggests gaal retool it into Task::Pugs 06:19
or Task::Smoke
gaal not toniht though
audreyt definitely tonight
which is to say, ~18 hours from now
gaal that sounds like a good schedule to me
audreyt :)
gaal moose moose
audreyt moose moose 06:20
shower for real &
gaal i still have a niggling feeling a dummy line wold be helpful but don't know why
night for real &
ivanfrey audreyt: Yes we did make some progress. Take Care.
audreyt you too :)
clkao: it's taking forever to commit 06:21
06:22 lisppaste3 joined
clkao audreyt: :( 06:29
svnbot6 r11196 | audreyt++ | * Sweeping change over all t/*.t t/*.pm files to make them executable with perl5. 06:36
r11196 | audreyt++ | * All the #! lines are gone.
r11196 | audreyt++ | * All the "use v6;" lines are replaced with "use v6-pugs;".
audreyt clkao: there you go
06:36 mjk joined
dduncan audreyt, didn't you use update S11 to be implementation neutral, such as "use v6-**"? ... that sounds like a step backwards then 06:42
audreyt dduncan: yeah indeed.
dduncan or is it okay because they're bundled with pugs?
audreyt $ perl -we 'use v6-**'
Argument "*main::*" isn't numeric in negation (-) at -e line 1.
it is not okay 06:43
it's just ** fails -w
"make test" defaults to -w
so we either find someway to subtly circumvent -w
(which is unlikely)
or switch back to
use v6-()
or use a bareword
use v6-0
use v6-any
use v6-perl
that is the leading candidates
dduncan maybe put a 'no warnings' block around that? ... or that could be worse
audreyt I think that's worse
as the idea is to let people effortlessly write perl6 code 06:44
that can be run in both perl5 and pugs
so far, I like
use v6-perl;
best
but I can certainly live with v6-() or v6-0 or any of the other alternatives.
dduncan if you want to be open-ended, the use v6-any seems closest in spirit to v6-** 06:45
in my opinion
audreyt yeah.
I like v6-perl because it's more self-descriptive
especially now we don't have #! anymore
it makes people see that "it's the version 6 of perl"
but otherwise I like -any 06:47
has the spirit of "anydbm"
in any case, this thing can wait till tomorrow with putter timtoady etc are all awake 06:48
dduncan now, what about ext/ modules etc? some or all of them still use shebang lines; should those be changed to be like the tests?
audreyt dduncan: I think so, but not before we 1)make the tests work with v6.pm 06:49
and 2)decide what -suffix to use
dduncan good idea
07:04 kane_ joined
clkao kane! 07:06
kane_ hey cl
clkao how are you
kane_ tired :) but only 3 more days till a month of vacation :)))
clkao we've just released a v6.pm that compiles perl6 Test.pm
whoot
kane_ gratz :) 07:07
its' going nicely i see
clkao so you are going honeymoon before wedding?
kane_ nope, somewhere in the fall, when weather is crap again :)
so, what's your visa status saying? 07:08
07:08 SM_ax joined
clkao i think i can go back to uk 07:09
pugs still takes forever to compile on my laptop 07:10
kane_ so does perl5 though :) 07:11
so the plan is to reside in the uk for now then? at least that's a place where we can come look you up :)
clkao for a while 07:12
i don't know yet. attempting to plan makes me sad
kane_ hah
is it now the P*** word?
07:15 cognominal joined 07:18 typester joined
clkao audreyt: pugs -BPerl5 needs to be smart about perl5lib 07:30
07:31 Aankhen`` joined
clkao audreyt: and perl6lib 07:31
audreyt clkao: how, exactly? 07:36
07:44 renormalist joined, drrho joined
clkao audreyt: should the perl5/* things go under blib rather than blib6/pugs/perl5 ? -M really hates the latter, because there is ../../../../blib 07:46
orz
07:54 bsb_ left
audreyt clkao: sorry, need to sleep now. let's continue tomorrow 08:10
08:37 cognominal joined 08:45 Aankh|Clone joined 08:59 kanru joined 09:00 Aankh|Clone is now known as Aankhen`` 09:09 elmex joined 09:13 kanru joined 09:16 visq joined 09:36 joples joined 09:44 SM_ax joined 09:49 Aankhen`` joined 09:51 elmex joined 09:57 froh-doh joined 09:59 drbean joined 10:03 dduncan left 10:09 buetow joined 10:41 mauke joined 11:04 nothingmuch joined 11:08 chris2 joined, frederico joined 11:10 lisppaste3 joined 11:15 bpphillips joined 11:32 EarleMartin left 11:34 ludan joined 11:52 Nouk joined 12:00 bpphillips left 12:02 Nouk left 12:23 weinig|sleeping is now known as weinig|away 12:40 Limbic_Region joined 12:47 Debolaz2 joined 12:53 vel joined 12:57 ruz joined 13:00 lollan joined 13:02 ivanfrey joined
ivanfrey audreyt: Maybe I copied this command wrong: 13:04
13:04 christopher joined
ivanfrey env PERL_HARNESS=./pugs prove t/01-sanity/*.t 13:05
env: prove: No such file or directory
pasteling "broquaint" at 217.206.131.214 pasted "An idea to simplify installing Pugs::Compiler::Regex rules" (30 lines, 535B) at sial.org/pbot/18224
broquaint You need Test::Harness installed, ivanfrey. 13:07
search.cpan.org/~petdance/Test-Harness-2.62/
lambdabot Title: "Andy Lester / Test-Harness-2.62 - search.cpan.org"
broquaint Or not, ignore me. 13:10
meppl guten morgen 13:17
ivanfrey broquaint: thanks, that did the trick. 13:18
13:20 kanru joined 13:21 fglock joined
ivanfrey audreyt: when I'm back on later I'll post my errors from "env PERL_HARNESS=./pugs prove t/01-sanity/*.t" 13:25
13:28 bpphillips joined
svnbot6 r11197 | fglock++ | v6 - a bit more of 'rule' parsing 13:28
13:29 cdpruden joined
nothingmuch morning fair lambdmoose 13:49
fglock nothingmuch: good morning 13:54
broquaint: thanks! # backlogging 13:58
broquaint np, fglock. 13:59
That code could suck less too, but it should work in theory. 14:00
fglock broquaint: do you have a commit bit? 14:01
broquaint I believe so.
Shall I make it so then?
(with tests)
fglock sure - many thanks! 14:02
source = perl5/Pugs-Compiler-Rule/
broquaint Right oh, I'm on it :) 14:03
14:04 daxim joined
clkao audreyt: devel::caller 01.0 should pass tests 14:04
fglock clkao: hi! 14:05
14:06 rzoz joined
nothingmuch clkao: morning 14:07
i hope I didn't wake you
clkao nothingmuch: no you didnt 14:08
fglock: hey
fglock clkao: I'm working on compiling rules (p6-on-p6) 14:09
clkao fglock: the s// rules are reverted, as it conflicts with my changes. i changed how the token substitution is parsed 14:10
14:10 buubot joined
fglock clkao: ok 14:11
14:12 buubot joined
fglock clkao: I don't know what to write in the ChangeLog - we changed a bit of everything :) 14:13
clkao fglock: "Changed everything to make Test.pm compile?"
14:13 justatheory joined
nothingmuch wonders what jesse meant by having to be up at 8 14:14
svnbot6 r11198 | fglock++ | v6 - ChangeLog on previous versions 14:17
clkao nothingmuch: but it
's for furniture devliery. os if they didn't ring it's probably alright
nothingmuch i was just kidding 14:18
i've been downstairs for a while and there has been no suspicious activity
so it's prolly OK =)
fglock clkao: ' sub { 42 } ' generates 'Data::Bind->sub_signature( \&, );' - looks wrong 14:20
clkao that's wrong
create a test case? 14:21
i don't have much time for perl6 hacking today
merlyn but if you don't work on perl6, the terrorists have already won! 14:22
clkao devel::caller has a macro in caller.xs. 14:23
define #WORK_DAMN_YOU
merlyn: i think my country care less about terrorists ;) 14:24
clkao wakes up
nothingmuch: we need to go shopping. you still need shoes?
nothingmuch ywa, od xoueaw
yes, of course
let's make breakfast, then go to herz and make sure the car is OK
then go downtownish as per obra reccomendations 14:25
clkao nothingmuch: you want to make breakfast? ;)
nothingmuch sure
clkao i know the gear shop, let me google map it
nothingmuch i got some reccomendations for jwang (ex bostonian), but they were fairly "anything works"ish 14:26
in Israel you have to be quite careful
because return policies are hard
so, breakfast arrangements have been delegated 14:27
rgs return policies?
nothingmuch as in, if these hiking boots are not good for me at all (which they may be, i have a pretty unique foot shape), in israel i could probably never return them
14:28 xdg joined
nothingmuch in the US i can 14:28
svnbot6 r11199 | fglock++ | v6 - 14-anon-sub.t - testing anon-sub compilation 14:29
nothingmuch morning xdg 14:30
xdg morning
nothingmuch, belated happy birthday, by the way 14:31
nothingmuch clkao: btw, the next time i fall asleep on the couch you are entitled to slapping rights
thanks, xdg =)
nothingmuch was trying to think and then somewhat nodded off
14:32 rzoz joined
svnbot6 r11200 | fglock++ | v6 - emitter: warn Dumper( $ast ) if $ENV{V6DUMPAST}; 14:35
r11201 | fglock++ | v6 - s/Dumper/.perl/ 14:38
14:38 hexmode joined
ajs FYI: > sudo cpan v6 14:42
Can't locate inc/Module/Install/Admin.pm in @INC
> perl -v 14:43
lambdabot Not in scope: `v'
ajs This is perl, v5.8.6
just FYI
fglock ajs: try 'reload index' - there is a new version already 14:44
ajs Just did. Still same msg. When I "cpan Module::Install::Admin" it says it's up to date as well 14:45
I tried force installing Module::Install::Admin, and still no change
14:47 shachaf joined
fglock oops - it is still broken in the last version: search.cpan.org/src/AUDREYT/v6-pugs...e/Install/ 14:48
lambdabot Title: "Index of /src/AUDREYT/v6-pugs-0.005/inc/Module/Install"
fglock ajs: I'll upload it again - thanks
it's missing in the MANIFEST file 14:49
14:53 weinig|away is now known as weinig
fglock audreyt: pugs/inc/Module/Install doesn't have Admin.pm 14:55
clkao fglock: really? it seemed to work for me. i instaslled 00 from pause incoming last night
ajs clkao: Did you have a previously installed, working copy? 14:56
clkao i was trying a fresh install
ajs clkao: huh. dunno
14:57 scook0_ joined, scook0_ left
fglock Module::Install in cpan doesn't have Admin.pm either 14:58
clkao admin is some special case in pugs/ only iirc
fglock ajs: try 'install Module::Install' - Admin.pm is in lib/ 14:59
ajs Q: is sub defined(Scalar $thing, ::Role?) {...} the correct way to describe defined? 15:00
S06 is a bit fuzzy on how type passing works vs. generics. 15:01
fglock ajs: I think this is describing 'does()' 15:02
ajs fglock: Well svn.openfoundry.org/pugs/docs/Perl6...ctions.pod says: $x.defined(SomeRole);
fglock ajs: sorry, I don't know 15:03
ajs Which is the method variant, but the same would have to apply to both the sub and method
fglock: NP. thanks anyway. I'll assume it's right for now.
15:05 chris2 joined
svnbot6 r11202 | broquaint++ | * Added the C<install()> method + tests + updated docs. 15:08
fglock broquaint: thanks! 15:09
broquaint Happy to help :)
gaal meows 15:13
audreyt ajs: ping 15:15
ajs: the M::I::Admin bug -- does it occur on v6.pm 0.005 on CPAN 15:17
and if yes, can you give me a shell access so I can duplicate and fix it?
szbalint audreyt: morning?
audreyt I've seen people reporting it but I can't quite reproduce
szbalint: greretings
szbalint I was wondering whether there are any plans to fix that try/exception handling bug, as it is interfering with my DBI project. I managed to work around it but that's kind of ugly. 15:19
svnbot6 r11203 | fglock++ | Pugs-Compiler-Rule - ChangeLog install()
audreyt szbalint: how did you work around it?
15:21 elmex joined
nothingmuch win14 15:21
broquaint Thanks, fglock, forgot about that :) 15:22
fglock np :)
szbalint audreyt: using perl5 code to do the eval. 15:24
clkao audreyt: can you see d::caller work for osx? it failed on threaded perl for me on bsd
ajs audreyt: pong 15:26
F/FG/FGLOCK/v6-pugs-0.003.tar.gz 15:27
Should I be getting a newer version? I've updated my CPAN indexes...
nothingmuch clkao: let's have lunch at diesel 15:28
audreyt ajs: please download 0.005 from search.cpan.org/dist/v6/
nothingmuch and then go on the subway
clkao nothingmuch: that's wrong planning. we should go porter sq and harvard sq 15:29
nothingmuch clkao: what part is wrong?
clkao nothingmuch: diesel is the other way, davis sq 15:31
15:32 rzoz_ joined
svnbot6 r11204 | fglock++ | v6 - parses very basic rule declarations; full parsing with 15:33
r11204 | fglock++ | Pugs::Grammar::Rule.rule is not possible yet
ajs audreyt: doing
fglock bbiab &
ajs audreyt: Works, but wants Devel::Caller which fails tests on install. Other than that installation went cleanly 15:35
15:36 JimKraj joined
audreyt ajs: there's a new Devel::Caller too. 15:40
pause.cpan.org/incoming/Devel-Caller-0.10.tar.gz
15:41 rzoz_ is now known as rzoz
ajs Thanks 15:41
15:43 prefiks joined
audreyt szbalint: wait a sec 15:46
szbalint: you had
use perl5:Foo
but there is no Foo
is that the error you want to trap?
szbalint no
but the test is path sensitive I'm afraid
15:47 bpphillips joined
audreyt oh. better to use core module 15:47
I'll fix
szbalint ok
15:48 putter joined
putter hi folks. 15:49
so is there a 'thon plan for lunch?
audreyt szbalint: it appears method-not-found does get captured
szbalint hm 15:50
putter fglock: congrats on recent milestones.
audreyt and explicit "die" is not
investigating
szbalint okay.
15:50 stevan joined
svnbot6 r11205 | audreyt++ | * exception_handling.t: switch to core module "Carp" for testing 15:53
15:53 buubot joined
putter hi stevan 15:55
gaal stevan: moose? 15:56
15:57 Daveman joined
putter Hmm. The pugs HEAD build is still broken. Same error as last night. Cannot import `Data.ByteString': it was found in multiple packages: pugs-fps-0.7 fps-0.7. Is there a workaround? 15:57
gaal putter: try ghc-pkg unrefister fps 15:59
*unregister
stevan gaal: moose
putter: hey
gaal stevan! welcome?
audreyt szbalint: I have a workaround
szbalint: namely demoting all uncaught "die" in perl5 land into warnings
stevan gaal: I am waiting downstairs for my hello hug 16:00
clkao stevan: while you are gone, v6.pm became crazy
gaal one hello hug, coming down!
audreyt szbalint: would that be acceptible?
szbalint audreyt: The reason I need try is to use it like the example in Inline::Java for exception handling (to interface with JDBC), so I need to trap the error/warning anyway somehow. 16:01
16:02 buubot joined
audreyt ok. so you do want $! in p6 land to contain the error in p5 land 16:02
szbalint yeah.
audreyt thinking. 16:03
gaal c 16:04
svnbot6 r11206 | szbalint++ | Removing unused directory. 16:06
16:06 buubot joined
stevan putter: are you joining our happy hackathon today? 16:07
in meatspace that is
Daveman Greetings, gaal :) 16:08
gaal hey Daveman
audreyt szbalint: I think I've got it fixed. testing 16:10
16:10 thepler joined
szbalint audreyt: oh, nice! Thanks for the fast help with it. 16:12
audreyt np :) thanks for prompting
the trick is essentially forcing a
return ($@, ...)
from p5 land
putter stevan: yes. though noone was at the office when I stopped by earlier. don't know what the plan is. 16:13
rgs hi all. just to check... (my $foo) = 42 is forbidden in P6 ? 16:14
audreyt ?eval (my $foo) = 42
16:14 evalbot_11162 is now known as evalbot_11168
evalbot_11168 (\42,) 16:14
audreyt that works just fine, thanks :)
rgs damn
I need a new pugs
audreyt yeah. that's fixed in 6.2.12 16:15
16:15 buetow joined 16:17 buubot joined
svnbot6 r11207 | audreyt++ | * Pugs::Compile::Rule - Fix Makefile.PL as noted by DanKogai++ 16:19
16:20 rzoz_ joined, guaumiau joined
fglock putter: thanks! 16:23
putter Ok, I removed ~/.ghc (didn't realize it was there - that sure made my past release testing a whole lot less useful). ~/.ghc had the only package.conf with fps. starting from scratch, make fails with: Could not find module `Data.Yaml.Syck'. 16:24
fglock: :)
16:24 thepler left
audreyt putter: can't be "scratch"; perl Makefile.PL will put a Data.Yaml.Syck there 16:24
16:25 shachaf joined 16:26 thepler joined
gaal putter: I think this has happened to me too but I wasn't paying attention when I hackarounded it. What 'make' invocation are you using? 16:26
try just 'make' 16:27
stevan putter: I am not sure of the details of the plan myself, but I know we will be migrating to the office somtime this afternoon
pasteling "spinclad" at 209.94.133.172 pasted "build failure at r11196" (28 lines, 1.1K) at sial.org/pbot/18237
gaal not unoptimized or pugs or anything
audreyt spinclad: r11207 fixed that
gaal putter, spinclad: yes, we should be moving to the office in an hour or so
putter audreyt: re scratch - was, didnt, retrying.
16:28 notThis joined
putter perl Makefile.PL >log_1 2>&1; make >log_2make 2>&1 16:28
16:31 amnesiac_ joined
putter Unless I'm being very careless... attempted build from scratch. problem unchanged. Hmm... 16:31
16:31 amnesiac_ is now known as amnesiac
putter ... nope, none of my package.conf's mention Yaml, so I'm not confusing things. 16:33
audreyt ok. 16:34
is this your laptop?
pasteling "putter" at 66.30.119.55 pasted "perl Makefile.PL >log_1 2>&1" (94 lines, 3.4K) at sial.org/pbot/18238 16:35
audreyt putter: ok. hm. I think "make" builds HsSyck
but this may be somehow prevented
try "make >log_1 2>&1" ? 16:36
pasteling "putter" at 66.30.119.55 pasted "make >log_2make 2>&1" (409 lines, 27.5K) at sial.org/pbot/18239 16:37
audreyt ok, so it registered pugs-fpc 16:40
-fps
but ghc-pkg can't find that
that is curious.
do you have write permission to ~/.ghc/ ?
spinclad audreyt: thanks, i finally caught up with that in bl
16:41 rzoz_ is now known as rzoz
putter I disabled ~/.ghc per above. I'll try putting it back. 16:42
Can users be depended upon to have a .ghc ?
"disabled" ~~ renamed 16:43
spinclad gaal: thanks
audreyt putter: yes ghc adds that automatically 16:44
gaal spinclad: we'll ping you when we leave for the office, about 15 before getting there
audreyt szbalint: please test.
szbalint: I _think_ I got it right 16:45
svnbot6 r11208 | audreyt++ | * szbalint++ requested that we propagate Perl 5 "die" errors into
r11208 | audreyt++ | Perl 6 "die" errors so that it can be caught and handled at "try" site.
r11208 | audreyt++ | We implement it with prepending $@ to each return list.
audreyt but I'm not sure of DBI6's test case
so please test :)
16:45 buubot joined
putter ok. thanks. will add renormalizing it to my test scripts. "yaks are cute creachers. bit black wet noses. friendly. sigh" 16:46
szbalint audreyt: okay.
spinclad gaal: good
szbalint tests
audreyt putter: working with Cabal is like... working with Module::Build
it clearly solves real problems, it's even made into core module, but the user-side incompatibility among versions makes it hard to work around 16:47
just yesterday I had M::B failing on me in a similar way.
PerlJam So ... what's this I hear about ponie 2.0? Anything interesting? Or should I just wait for the official announcement? 16:48
audreyt it needs ExtUtils::ParseXS to support XS modules, but it can only use it if it's installed _before_ it
as it's not a dependency, only a recommends
16:48 iblechbot joined
audreyt PerlJam: it's ideaware at this moment; the idea seems to be rebuilding Perl 5 without using Perl 5 core, using TGE/PGE alone 16:48
putter What can we do so that perl6 has less of this hair? 16:49
audreyt I hope it's started after Punie passes more than 3 or 50 test files...
s/it's/it'll be/
gaal audreyt: which reminds me, does MB/MI have the concept of "require one of these n modules?" thinking of Task::PugsSmoke which surely recommends YAML::Syck but requires at least YAML.
putter: builting PAR and many pure-perl bootstrapping modules 16:50
audreyt gaal:
if ( can_cc() ) { requires('YAML::Syck' => '0.27') unless can_use('YAML' => 0.35);
} else { requires('YAML' => '0.35') unless can_use('YAML::Syck' => 0.27);
}
but no it's not abstracted away yet
stevan use Best :)
gaal excellent
stevan: we do in fact :) but this is at install time
PerlJam audreyt: so, they're going to forego bug-for-bug compatibility and start from scratch?
(assuming the idea gets off the ground)
audreyt PerlJam: it's all ideaware at this moment so I have no idea 16:51
wait for announcement later, maybe
putter re fps unregister, one needs something like ghc-pkg -f ~/.ghc/x86_64-linux-6.4.1/package.conf unregister fps because the default ghc install has the main package.conf owned by root, pugs builds modify the ~/.ghc package.conf, and ghc-pkg goes boom when it finds it cant write the main .conf, even though it doesnt need to. :/
audreyt ghc-pkg --user unregister fps
is the incantation
spinclad I noticed in my makelog last night: after sudo ../../Setup install of both pugs-fps and pugs-HsSyck, the make went through but appeared to be installing in ~/.ghc where it would not be found. (before rafl's changes they were found there.) 16:52
audreyt oh!
hm.
trying to duplicate here 16:53
spinclad s:last/it/they/
(it found the global installed ones so went through, but 'make' is no longer keeping them up to date...) 16:54
gaal eep, /me remembers sudo chmod a+w his ghc pakcage registry after some losing battles with lossage of this sort
audreyt spinclad: I'm seeing something else here 16:55
gaal no doubt the Wrong Thing
audreyt spinclad: here it looks into --user and found it
ingy I want to be recognized as the one who coined TMTOWTJFDI
TimToadJifty 16:56
gaal recognizes ingy on sight
spinclad tmtowtfiu
gaal cogs ingy again and again
16:57 ruoso joined
spinclad Cabal-1.1.3, btw, debian latest 16:58
if we can get it looking in --user again reliably, we shouldn't have to sudo Setup install 17:01
putter (btw, my .ghc is back, its package.conf is now [], make error apparently unchanged, package.conf remains unchanged despite fps registering. now exploring the fps registering process.) 17:04
putter wonders whether you could shampoo yaks, and thus not mind them standing around all the time. hay smells nice. might need to housebreak them. 17:06
svnbot6 r11209 | fglock++ | Pugs-Compiler-Rule/compile_p6grammar.pl - compiles Rule.pm
17:20 Debolaz2 is now known as Debolaz
putter yipes. 33 thousand spam emails in just a couple of weeks. 17:21
gaal spinclad, putter: we're moving over to the office now
rafl spinclad: What do you mean with "would not be found"?
gaal see you soon &
rafl spinclad: also 1.1.4 is the latest Cabal in Debian.
putter gaal: thanks.
spinclad rafl: 1.1.3 is sid latest (that's experimental, right? is there later?) 17:24
szbalint hm putter, the from-scratch build I've tried bombed out too
"Could not find module `Data.Yaml.Syck'"
rafl spinclad: There's ghc6 6.4.2-2 which comes with Cabal-1.1.4
spinclad rafl: there's my clue, thanks 17:25
putter szbalint: good to know. it looks like the chain is Data.Yaml.Syck not found because HsSyck build failed because error reading ./.setup-config because cannot satisfy dependency pugs-fps-any because... something silently went wrong with fps registration or install. 17:28
spinclad rafl: "would not be found": i saw cases where the ghc-pkg register/unregister succeeded (presumably to ~/.ghc) but pugs-fps or pugs-HsSyck were not found when needed. 17:29
rafl spinclad: It shouldn't use ~/.ghc anymore. I build wrappers around ghc and ghc-pkg which install/register things inside the build tree. 17:30
spinclad: third-party/installed/packages.conf or somesuch. 17:31
putter (the registration at least is not happening. either to .ghc, or a chmod a+w main package.conf)
rafl: does sial.org/pbot/18239 look like what you would expect?
lambdabot Title: "Paste #18239 from "putter" at 66.30.119.55"
rafl To see if it's registered in third-party/installed you could use ./util/ghc-pkg-wrapper -l 17:32
putter is there a .conf file?
spinclad ok, it is registered there 17:33
rafl putter: That's not what I expected. There's a conf file in third-party/installed which is created on the first run of either the ghc or ghc-pkg wrapper.
spinclad the dependency search may not be looking there
17:34 notThis joined
pasteling "putter" at 66.30.119.55 pasted "third-party/installed/packages.conf" (1 line, 1.1K) at sial.org/pbot/18241 17:34
rafl putter: fps seems to be installed properly. Could you check if your ./util/ghc-pkg-wrapper -l can see it?
putter checking... 17:35
spinclad i'm going to unregister them from global and ~/.ghc and try again 17:36
putter yes, pugs-fps-0.7
rafl putter: You could also write a small haskell file which uses Data.ByteString and compile it with util/runcompiler -hide-all-packages -package pugs-fps
putter writing... 17:37
17:39 Aankhen`` joined 17:40 bernhard joined
rafl putter: You'll also need -package base, I guess. 17:41
putter yes...
17:44 frederico joined
TimToady Cabal: A sufficiently advanced technology is indistinguishable from 90% crap... 17:44
spinclad rafl: yes, that works (making ./pugs) (aside from needlessly rebuilding both packages, not a new problem) 17:50
Limbic_Region um, did the GHC version requirement change or is Win32 b0rk a known problem?
and WTF is a Cabal installation? 17:51
rafl spinclad: I'll try to fix the latter right now.
pasteling "Limbic_Region" at 129.33.119.12 pasted "Win32 perl Makefile.PL failure(s)" (13 lines, 534B) at sial.org/pbot/18243
rafl putter: If that didn't work, try removing ~/.ghc. I'll be offline for about an hour now. 17:52
spinclad not a problem with the wrapper that i can see
Limbic_Region fwiw - this is after an nmake realclean and a svn up 17:53
17:54 jkanter joined
svnbot6 r11210 | fglock++ | Pugs-Compiler-Rule - fixed quoting in emitted :ratchet code 17:54
spinclad & # food, shower, check back here, head over to davis^2 17:55
audreyt Limbic_Region: try now? 17:56
svnbot6 r11211 | audreyt++ | * Makefile.PL - fix win32 path in ghc-pkg wrapper
r11210 | fglock++ | Pugs-Compiler-Rule - fixed quoting in emitted :ratchet code
17:57 stevan joined
audreyt Limbic_Region: it's rafl and his unix-centric code again :) 17:57
but it should be fixed nnow
putter: come to the office and let's debug it f2f? :)
(with your laptop)
stevan yes putter please come by :) 17:58
fglock does <@terms> capture to $/<terms> ? 17:59
17:59 Aankh|Clone joined
putter rafl: sorry, a call. A simple ./util/runcompiler -hide-all-packages -package pugs-fps -package base Main.hs works, where Main imports but does not use ByteString. I'll try using next... 18:02
audreyt: no laptop. :( awaits having a job. 18:03
Limbic_Region tries again 18:05
stevan putter: come anyway 18:06
Limbic_Region so far so good audreyt 18:07
audreyt
szbalint audreyt: I seem to have the same problem as putter, if you want I can offer shell access to my development pc.
Limbic_Region grrr - my ability to use plus symbol is gone again
18:08 particle_ joined
kolibrie stevan: I submitted a talk proposal for pghpw.org 18:08
lambdabot Title: "Pittsburgh Perl Workshop - 2006-09-23 - About"
particle_ howdy all
fglock Limbic_Region: no plus symbol here too
TimToady fglock: no
stevan kolibrie: nice
fglock TimToady: thanks 18:09
TimToady only <alphabetic> captures by default.
particle_ do regexes allow routine modifiers (eg. multi rule foo { ... } )?
TimToady which is why <?ws> disables capturing ws.
putter steven: right, will do
particle_ ?eval multi rule foo { ... } 18:10
evalbot_11168 Error: unexpected "f" expecting comment, "handles", bare trait, subroutine parameters, trait or block
TimToady yes, but a multi rule isn't going to do you much good without a different signature.
putter anyone: could someone give me a oneliner use of ByteString. my haskell-fu is failing me...
particle_ understood, just making parser mods
kolibrie maybe tomorrow I can finally get around to looking at grammars again, after the nice fixes audreyt++ and pmichaud++ put into pugs and parrot for me
TimToady rules are specifically allowed to declare sigs though. 18:11
putter oh, neat (rules allow sigs)
kolibrie notes that in his brain for future reference
TimToady we already use it for things like <operprecparser(@terminators)> 18:12
though that's presumably just an ordinary sub there. 18:13
particle_ yep, pge does that
s/pge/pgc/
TimToady or rules of the form <abc stuff>
anyway, you can parameterize rules.
particle_ hrmm, so a space seperates the rule name from it's args 18:14
TimToady rather important in recursive descent parsing to know what the expected terminators are.
particle_ iirc perl6 subs allow spaces in names
TimToady <abc stuff> parses stuff as regex. <abc(stuff)> parses stuff as arglist.
ingy hola
18:14 Aankh|Clone is now known as Aankhen``
particle_ howdy ingy 18:14
TimToady currently <abc: stuff> is equivalent to <abc('stuff')> or some such. 18:15
putter re ByteString, got it, exploring
ingy ang it 18:16
d
I made coffee an hour ago
and forgot about it
a sure sign I haven't had my coffee yet
particle_ should 'my Int sub foo ( --> Int )' be a parse error? 18:18
?eval my Int sub foo ( --> Int ) { ... }
evalbot_11168 Error: unexpected "-" expecting comment, formal parameter or ")"
18:19 larsen joined
TimToady ?eval my Int Int sub foo () { ... } 18:20
evalbot_11168 Error: unexpected "I" expecting comment, "\\", ":", "*" or parameter name
TimToady ?eval sub foo (Int Int $x) {...}
evalbot_11168 Error: unexpected "I" expecting comment, "\\", ":", "*", parameter name or ")"
TimToady Looks like it doesn't handle "anded" types yet. 18:21
In any event, shouldn't be a parse error.
particle_ great, was hoping that
TimToady could possibly be a semantic error if you say my FloorTopping sub shinola (--> DesertWax), but it should always at least parse. 18:22
*DessertWax
and even if the two types can't be combined, it'd probably just be a warning about unreachable code. 18:23
particle_ and the --> must follow all params, yes? 18:24
TimToady yes.
particle_ okay, next. 18:25
rule signature { $<invocants>:=(<parameter>? [, <parameter>]*)
$<paramlist>:=(\: <parameter>? [, <parameter>]*)? }
TimToady and we've reserved the ( --> Foo; ) syntax (note semi) for possible mmd tiebreaking on return type.
particle_ ooh, good to know
that's an optional semi denoting mmd type tiebreak... will do 18:26
the signature rule above populates $<invocants> in absense of a \:
this seems incorrect
i believe $<paramlist> should be populated if there's no '\:'. agreed? 18:27
TimToady nowadays invocants (plural) is always incorrect. we no longer call mmd longname parameters 'invocants'
that seems right. 18:28
the invocants vs longnames is partly why we differentiated : into : or ;
particle_ i see.
does anything every follow the ; ? 18:29
TimToady so there can only ever be one :
and if it's missing, the invocant is implied
putter rafl: Ok, re runcompiler, my only note is it errors if Main is not defined. Other than that, it worked in different directories, and ByteString was used ok (didnt try fps itself). Happily dropped a.out's. 18:30
TimToady previously, if we saw multi ($self:...) we had to guess...
sure, you can replace every comma with ; if you like, then you get common lisp mmd semantics
if you leave out all ; then there's one implied currently after the last param but before any --> 18:31
putter TimToady: so what is a "oneliner" definition of ; ?
TimToady well, the real definition is being worked out, but the oversimplified meaning is that anything to the right of a semicolon is treated as 18:33
Any unless the information to the left is insufficient to determine a unique function.
particle_ does the following look (more) correct for the signature rule body? 18:34
$<invocants>:=(<parameter>? [, <parameter>]* <':'>)?
$<paramlist>:=(<parameter>? [, <parameter>]*)
rafl putter: So it's probably breaking because of ~/.ghc. 18:35
TimToady you can't have multiple parameters before the colon. 18:36
particle_ ah, in that case: $<invocants>:=(<parameter> <':'>)?
putter TimToady: if the information to the left is insufficent, but a few parameters later we are down to a unique, are the remaining params Any?
TimToady see ./misc/pX/Common/mmd-draft.txt 18:37
merlyn should Devel::Caller fail badly on OSX? 18:38
putter mmd-draft.txt doesn't seem to have changed since I last saw it... 18:39
TimToady $<invocant>:=(<parameter>? <':'>)?
merlyn I'm trying to install the new v6.
putter sigh. I keep running the same mind path. perhaps needs to be a note in mmd-draft. to wit, 18:42
18:42 arguile joined
Limbic_Region fglock - Juerd had fixed it once, but apparently some upgrade has reverted the change 18:45
Juerd ping
fglock merlin: audreyt tested in OSX, afaik 18:46
Limbic_Region audreyt - the Makefile.PL fix did not fix the Win32 build 18:47
Limbic_Region tries again just in case there was a fix checked-in that he didn't see
svnbot6 r11212 | fglock++ | Pugs-Compiler-Rule - implemented '<@rules>' in :ratchet mode 18:49
18:49 weinig is now known as weinig|away
Limbic_Region interesting nmake realclean is b0rk now 18:50
particle_ if you updated your working copy and files/dirs were shuffled, realclean may break 18:51
unless you regenerate the makefile
Limbic_Region particle_ - well, first perl Makefile.PL was b0rk (audreyt fixed)
so then I did a realclean, svn'd up, and did a nmake 18:52
particle_ which compiler on win32?
Limbic_Region that bombed not too far into the make so I did a nmake realclean again (to see if there was another update since I was AFK)
but that failed
which compiler?
GHC 6.4.1
particle_ oh 18:53
Limbic_Region er 6.4.2
and AS perl 5.8.8 though it really isn't doing any of the compiling
Limbic_Region is attempting another nmake to see if the previous problem is still present and will nopaste accordingly 18:54
particle_ ouch... Devel::Caller tests just broke perl 18:55
(segfault)
Limbic_Region well, they don't call it Devel for nothing ;-) 18:56
putter given (Int;Int;Int,Int) (Int,Num,Int,Int) (Int,Int,Num,Int) (Int,Int,Int,Num) and (10,11,12,13), with the sigs taken pairwise, 1&2 -> 1, 1&3 -> 3, 1&4 -> 1. no ambiguities. are all those intended?
merlyn heh
putter rafl: any tests you want me to run before heading out? 18:57
19:00 Aankh|Clone joined
particle_ merlyn: Devel::Caller breaking badly on win32, too 19:00
Limbic_Region pugs is currently b0rk on Win32 19:01
merlyn it's broken all over, then, not just me. good. as long as others are sharing my pain. :)
Limbic_Region both nmake
and nmake realclean
pasteling "Limbic_Region" at 129.33.119.12 pasted "Win32 nmake failure" (94 lines, 5.2K) at sial.org/pbot/18245 19:02
19:03 frederico joined
putter no, my examples are broken, let's see... 19:03
particle_ notices a new version of Devel::Caller on cpan today
perhaps the older Devel::Caller works better
stevan putter: come to the offices,.. I am bored,.. they are babbling about ASTs 19:06
pasteling "Limbic_Region" at 129.33.119.12 pasted "In addition to the nmake failure nopaste, here is the nmake realclean failure for Win32" (58 lines, 5.1K) at sial.org/pbot/18246
spinclad rafl: same question as putter 19:07
spinclad heads out now too & 19:09
(about 45 min trip) 19:10
putter stops composing mmd cases
ok, off to 'thon. (about a 5 min trip;)
&
19:13 korihor joined 19:15 korihor left
svnbot6 r11214 | audreyt++ | * SVG diagrams for old, and new, runtimes 19:17
r11215 | audreyt++ | * p6ast.hs - sync with gaal on examples of new runcore 19:21
19:22 amv joined 19:25 Aankhen`` joined 19:40 dduncan joined 19:46 xdg joined
svnbot6 r11216 | gaal++ | * Bundle::Pugs::SmokeKit has been obsoleted in favor of Task::Smoke 19:51
Limbic_Region ok - so do any of the thon'ers know why Win32 is b0rk? 19:52
audreyt Limbic_Region: I think I fixed it 19:53
Limbic_Region k 19:54
Limbic_Region tests now 19:55
oh audreyt - my wife said the following about you when she saw your pic "she is pretty and she must be smart"
19:55 cognominal joined
theorbtwo She can tell that from a photo? 19:56
Limbic_Region well the pretty part is mostly subjective so - yes
the smart part - I have no idea why she said that 19:57
perhaps it is because she doesn't often see me oogling pictures of hot female hackers and assumed for audreyt to be there - she must be smart
Limbic_Region wasn't capable of conveying the extent to how smart but it made me laugh when she said it 19:58
svnbot6 r11217 | fglock++ | v6 - added lib/Pugs/Grammar/P6Rule.pm - Rule grammar translated to P5 20:00
clkao oh bloody scary 20:01
Limbic_Region audreyt - still b0rk
will nopaste the results in case they are different
audreyt ok
pasteling "Limbic_Region" at 129.33.119.12 pasted "lastest Win32 nmake failure" (136 lines, 7.5K) at sial.org/pbot/18247 20:02
Limbic_Region testing nmake realclean to see if *that* is also still b0rk
nope, nmake realclean still b0rk too 20:04
audreyt Setup.exe: ..\..\gcc-lib\ld.exe: runProcess: does not exist (No such file or directory) 20:05
hm.
where is gcc-lib\ld.exe originally?
is it in your c:\ghc\
Limbic_Region doesn't have a c:ghc\ 20:06
audreyt ok, where is your ghc?
c:\Gat\ghc
Limbic_Region everything is under C:\gat\<root>
yep
audreyt is there a ld.exe under c:\Gat\ghc ?
Limbic_Region so let me find ld.exe
C:\Gat\GHC\gcc-lib
it's the only one there
audreyt ok 20:07
thinking
is ghc.exe in c:\gat\ghc\ghc.exe ?
Limbic_Region I do have mingw on this box btw - it is in the PATH later though
c:\gat\ghc\bin\ghc.exe 20:08
fwiw - this is a bin install with no tweakage on my part
audreyt understood
ajs Revisions to S29 away... I'm finally very, very slightly ahead of the curve ;) 20:10
rafl spinclad: I thought it's working for you now.. 20:11
audreyt clkao: fixed 20:12
Limbic_Region: if you copy c:\Gat\ghc\gcc-lib into the pugs tree 20:14
does it suddenly work?
svnbot6 r11218 | audreyt++ | * Handle ./pugs -BPerl5 -e 'use Test' by recognizing lack of package
r11218 | audreyt++ | as the main progrem in v6.pm
r11219 | fglock++ | v6 - parser does rules! ' rule xxx { a.b$ } '
audreyt (it's not the correct fix, just diagnosis)
kolibrie fglock++ 20:16
ajs fglock: that's very cool.
audreyt ooh, bootstrap.
today seems to be the day of bootstraps; stevan just got Class::MOP's immutable (fast) object layout bootstrap itself, so all metaobjects are suddenly fast 20:17
20:17 notThis joined
kolibrie oooh, stevan++ 20:18
20:19 spinclad_ joined
Limbic_Region audreyt - checking now, was AFK 20:20
20:20 fglock joined
audreyt Limbic_Region: ok. is ghc in PATH? 20:21
Limbic_Region you want all the files copied - or the dir itself
audreyt the dir itself with all files in it
Limbic_Region c:\gat\ghc\bin is
kolibrie fglock: we all noticed r11219
audreyt fglock: regex and tokens too?
svnbot6 r11220 | audreyt++ | * runcompiler.hs - use rawSystem to avoid whitespace trap 20:22
r11220 | audreyt++ | on dirs with whitespace.
fglock kolibrie: cool - I used compile_p6grammar to translate the grammar
audreyt: yes
Limbic_Region audreyt - svn up'ing now
kolibrie fglock: cool, I'll have to look and see what improvements it has grown since I last touched it 20:23
fglock it's not hooked in emitter yet
20:23 xdg joined, bpphillips left
kolibrie one step at a time 20:25
20:26 notThis left
Limbic_Region audreyt - if this does make the problem magically go away - what is the *right way* to fix it? 20:27
20:27 vel joined
Limbic_Region cause it appears to be compiling now 20:28
20:28 nothingmuch joined
audreyt Limbic_Region: the right way is to fix findpath 20:31
Limbic_Region findpath is in build_util.pl ? 20:32
nothingmuch clkao_: how do we plan on picking up the car?
walk over?
Limbic_Region though I am not sure I see how to fix it 20:33
kolibrie nothingmuch: six people
Limbic_Region gcc-lib is not part of my PATH
so walking the path is out
nothingmuch kolibrie: moose pie
what do six people have to do with it?
20:33 TimToady joined
kolibrie six strong people 20:34
for lifting
nothingmuch oh
i doubt six are enough 20:35
1500 k / 6 is quite a bit
i doubt i can lift 200+ kilos
kolibrie me neither
Limbic_Region audreyt - must go now, will touch base with you tomorrow if it isn't fixed by then ;-)
szbalint audreyt: Is there a way I can work around that Data.Yaml.Syck b0rk? I can't build Pugs without that to test your fix.
clkao nothingmuch: sure 20:37
nothingmuch clkao: so we should leave roughly now 20:38
audreyt szbalint: give me shell?
clkao i think we can be late a bit
szbalint audreyt: ok
nothingmuch clkao: when i looked up 18:00 it said "there are no open rental agencies at the hour you selected" 20:39
maybe we can use jesse's car and two drivers?
clkao nothingmuch: uhm, the opening hours saz till 1800 20:40
nothingmuch clkao: oh
okay
i didn't read the sign
szbalint audreyt: see pm for details. 20:46
spinclad_ rafl: it is... was wondering if any tests i could do would help (like reregister by hand in ~/.ghc, watch wrapper want to register in global)
but now i'm away. could try things on feather (where sudo is not an option anyway) 20:47
fglock clkao: how about making Data::Bind optional for now - default to the old behaviour if it is not installed 20:56
20:56 weinig|away is now known as weinig
clkao fglock: why? then things will break 20:57
no named args
fglock clkao: I can't run in windows :(
I'm only testing simple statements 20:58
clkao i thought audrey made the windows
fglock we can make it proceed with warnings 21:00
21:01 DaGo joined 21:03 dolmen joined
svnbot6 r11221 | fglock++ | v6 - added "invalid rule syntax" message 21:07
21:11 sahadev joined 21:14 Sage joined
rafl spinclad_: Nothing is registering globally anyway. I could only think of problems when things are already registered in ~/.ghc.. 21:16
unobe audreyt: uploaded latest smoke test with PUGS_EMBED=perl5, and created new image file @ unobe.com/packages/pugs-6.2.12-OSX-10.4.dmg
audreyt woot 21:24
21:25 ludan joined 21:26 Daveman joined 21:29 bpphillips joined
spinclad_ rafl: i was seeing attempts to register globally before i got rid of ~/.ghc entries... audreyt was working with Limbic_Region last night with a manual fix step of C<sudo ../../Setup install> to succeed in registering globally 21:32
can we detect old ~/.ghc entries and remove them?
21:35 tup joined
svnbot6 r11222 | fglock++ | v6 - placeholder for rules in emitter; 21:40
r11222 | fglock++ | - "Useless use of a variable in void context at (eval 200) line 71."
clkao audreyt: what's the win32 data::bind situation? 21:47
21:55 Sage left 22:01 cognominal joined 22:06 vsoni joined
svnbot6 r11223 | fglock++ | Pugs-Compiler-Rule - fixed a problem with each() initialization in the 22:06
r11223 | fglock++ | emitters
vsoni seen audreyt
jabbot vsoni: audreyt was seen 42 minutes 5 seconds ago
22:06 nothingmuch joined 22:07 weinig joined, bpphillip1 joined
svnbot6 r11224 | fglock++ | v6 - emits code for 'token { <a> }'; untested! - needs more work on parameters 22:12
22:13 bpphillip1 left 22:14 frederico joined 22:16 Eimi joined
fglock it's now possible to implement s/// with p6 rule syntax inside 22:17
& 22:19
22:19 fglock left
gaal TimToady /audreyt: L<06/"Immutable types"> is missing a Tuple type. 22:22
TimToady s/Tuple/Seq/ 22:23
TreyHarris TimToady: what should I do with typos in Synopses? send you a diff?
gaal ah.
TimToady TreyHarris: that works
gaal TimToady: but that implies only positional lookup, doesn't it? 22:24
rgs TimToady: audreyt: I'd love to have your advice on www.xray.mpe.mpg.de/mailing-lists/p...00148.html
TimToady we wanted to avoid conflict with the relational meaning of tuple.
lambdabot Title: "state variables progress"
gaal TimToady: what about "simple" immutable records?
are those simply locked mappings? 22:25
I don't know how important this is, but haskell gives records you can query either by position or by name. 22:26
TimToady rgs: 1 is fine. assignment to (state $x) is basically just erroneous, and they should be glad if they get a warning. 22:27
rgs TimToady: OK. this one was my favourite actually. :)
TimToady gaal: I suppose. I'm not entirely sure how much we distinguish these things by type, and how much we can just recognize immutability as an optional "feature" of mutable types. I suspect Audrey has a clearer idea about this... 22:28
perhaps captures can be made to cough up their named associations sequentially somehow... 22:29
even a hash can give you the elements in an order, if you don't mind random order. :) 22:31
gaal :) yes, I noticed the ordedness of Mapping isn't specced...
TimToady but Perl doesn't really care to confuse the Array and Hash roles...
TreyHarris oh, c'mon. it worked so well for PHP *choke8 22:32
TimToady on the other hand, it certainly doesn't care if you wnat to define an object with both roles.
after all, captures do just that...
but not with the same data 22:33
22:33 bpphillips joined
TreyHarris TimToady: that's what i thought, but why doesn't this work then: 22:34
?eval my $z = \($x, $y, :a, :!b); say $z<a>
evalbot_11168 Error: Undeclared variable: "$x" 22:35
TreyHarris oops. well, it doesn't work even if you declare them :-)
22:35 stevan joined
TimToady I don't think captures are completely implemented yet. 22:35
TreyHarris ok. so that *should* work, and say "1"? 22:36
TimToady I seem to recall it was one of the things that was waiting on .12
yes, it should, I think.
or rather, probably "Bool::True". 22:37
gaal the forky view could be useful in XML<->Capture mappings, or maybe not. 22:44
I mean, does $node<foo> mean attribute foo or a child node <foo/>
22:46 JimKraj left, bpphillips left
ingy hola 22:55
TimToady gaal: I would take it to be attribute foo. There's no guarantee of uniqueness for child node <foo/>. 23:01
gaal sure. XML-Simplisti^H^H^H^He nonwithstanding. 23:10
svnbot6 r11225 | audreyt++ | * change the ref. implementation for multi-dispatch to
r11225 | audreyt++ | cover the new change proposed by spinclad++:
r11225 | audreyt++ | each variant eliminates itself at semicolon points
r11225 | audreyt++ | if it's lost at that level.
23:13 joples joined 23:23 marksto1 joined 23:31 rashakil joined
kolibrie audreyt: can send me your .vimrc (or at least your minibufexpl.vim settings) 23:42
gaal kolibrie: just put that file in ~/.vim/plugins/ 23:44
(nad hi :)
*and
kolibrie gaal: moose!
gaal: but I want my filenames down at the bottom 23:45
gaal heh, interestingly I don't see an obvious setting for that 23:49
kolibrie it's a very long file to browse through looking for settings 23:50
kolibrie thought audreyt might have other nifty things set up, too 23:52
shower & 23:55

Powered by Cro and the Raku® Programming Language.

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

Thank you!