SVK users: search.cpan.org/dist/SVN-Mirror/ | paste: sial.org/pbot/perl6 | pugs.blogs.com | pugscode.org | pugs.kwiki.org | www.treehugger.com/files/th_images/paradigm.jpg
Set by audreyt on 17 August 2006.
svnbot6 r12636 | lwall++ | Added .pm files, on assumption that duplicate parses are preferred to missing 00:04
r12636 | lwall++ | parses, and there might be examples that aren't named with the .pl extension.
svnbot6 r12637 | markstos++ | Document in Differences.pod that eval {} in Perl5 is try{} in Perl6. 00:37
marksto1 I don't see my syntax error. I keep getting "undeclared variable" for %args. Usually it seems declaring a data structure in the sub signature takes care of that. 01:19
multi method run_modes (%args) { return %!run_modes = %!run_modes, %args; }
markstos_ I tried a number of syntax variations, but I keep getting the error. 01:19
It strange, because I've done similiar things a number of times and haven't run into this.
markstos_ It looks like pugs treats subroutine signatures of (@a) and (%a) the same. 01:30
svnbot6 r12638 | fglock++ | v6 - s/$_V6_GRAMMAR/%_V6_GRAMMAR/ 02:05
r12638 | fglock++ | - implemented %(...)
r12638 | fglock++ | - fixed P6Term.pm to implement Term.pm API
r12639 | markstos++ | Attempt to enhance the usefulness of this page by cross-referencing the Apocalypse and Exegesis documents,
r12639 | markstos++ | and also try to explain what the heck the difference is, and which set of docs to look to first.
pasteling "markstos" at 12.176.101.89 pasted "selecting the right multi-method when choosing been arrays and hashes" (26 lines, 554B) at sial.org/pbot/19246 02:23
markstos_ So someone take a long at this failing test (it even is made with "use Test"!) ? I'm not sure if I've run into a pugs bug, or my own ignorance and confusion. 02:24
svnbot6 r12640 | markstos++ | remove "$self: " references here, which don't seem to add anything here. 02:59
r12640 | markstos++ | (Tests still pass after removing it. )
markstos_ I think what I don't understand is: How declare for a subroutine signature that the first arg is a hashref, as opposed to a hash? 03:04
PerlJam markstos_: What's a "hashref"? Perl6 doesn't have refs anymore 03:14
markstos_ \%h
What's \%h ? 03:15
?eval my %h = ( a => 'b'); say ref \%h; 03:16
evalbot_12640 OUTPUT[Hash ] Bool::True
markstos_ Hmmm.... smells like a working hashref.
I want to handle foo( a => 'b') and foo({ a => 'b' }); 03:17
PerlJam pugs understands refs, but it doesn't yet understand that they're going away 03:18
markstos_ Is \%h going away ? 03:19
Is foo({ a => 'b' }) going away?
PerlJam no, but they are just hashes to perl6
markstos_ So then pugs is buggy if it dispatches foo(\%h) and foo(%h) differently? 03:20
( Incidently, there don't seem to be tests for foo(\%h) and foo(\@a) in the test suite yet. 03:22
I could add them, but I want to be sure of the expected behavior1
s/1/!/
audreyt \%h dispatches to Capture
\@a too
%h and @a dispatch to Hash and Array respectively 03:23
{a=>'b'} dispatch to Array
er
{a=>'b'} dispatch to Hash
[1,2,3] to Array
(sorry, just woke up.)
markstos_ audreyt: Thanks. Is there an easy to express that I want \%h and %h handled as if they were the same? 03:24
audreyt the current runcore is Dec 2005's perl6, before the notion of Capture and before the semantics of \ got worked out
why would you want to do that?
markstos_ Ah.
PerlJam audreyt: thanks, I was trying to find a reference (heh!) that confirmed that \XXX always Captures XXX (all of the examples that I found use \(...) or \"constant")
markstos_ audreyt: I'm porting something from Perl5, and it allows you to do foo( a => b ) or foo({ a => b}) 03:25
Of course, I could drop support for one varation as part of the port...
audreyt multi foo (%x) { ... }
PerlJam markstos_: that's a Pair and a Hash
audreyt multi foo (*%x) { foo(%x) }
like that?
PerlJam: er no, that's named and a hash :)
markstos_ checks
audreyt Pair is when you do foo('a'=>'b') 03:26
note the lhs of => is not bare (which would be named)
PerlJam oh, that's right, I've obviously forgotten
you've just woken up and I'm still thinking in perl5 :) 03:27
audreyt markstos_: I think the most straightforward way is 03:28
sub foo (*@_, *%_) { %_ = (%_, @_); say %_.perl } 03:29
markstos_ checks that
audreyt er sorry 03:30
sub foo (%pos?, *%nam) { my %arg = (%nam, %pos); say %arg.perl }
the latter is what you've expressed in english, I think
unless you want to mutually exclude them
in which case you want 03:31
sub foo (%pos?, *%nam) { my %arg = %pos || %nam; say %arg.perl }
but hopefully the idea is clear :)
markstos_ audreyt: Thanks. I'll think about that. 03:32
audreyt :))
pasteling "markstos" at 12.176.101.89 pasted "to complicate matters, I want to support a hash, hashref, and arrayref calling styles..." (17 lines, 268B) at sial.org/pbot/19247 03:36
markstos_ audreyt: so to complicate things, the interface I'm porting also supports an arrayref. I'm having trouble modeling that as a multi method, in addition to the hash and hashref options. 03:37
PerlJam audreyt: Is someone working on teaching pugs about Captures?
audreyt PerlJam: sure, when you see "newland val" or "newval" or "vv" on channel mentioned by gaal luqui me etc 03:38
it's the new Capture-based runtime
markstos_ I wonder if I should just revert to the Perl5 style code, which takes in an array, and then inspects it to see whether a hashref or arrayayref was based in.
audreyt ?eval vv("newland string").method(x=>1)
evalbot_12640 "CCall \"method\" CaptMeth \{c_invocant = VPure (MkStr \"newland string\"), c_feeds = [MkFeed \{f_positionals = [], f_nameds = \{\"x\":=[VPure (IFinite 1)]}}]}"
audreyt see, a capture!
but of course we won't force everyone to write vv() :)
so once the method dispatch stabilizes, all oldland values will become newland values, and the vv() bridge won't be needed 03:39
Alias_ audreyt: Oh good, you're here
PerlJam audreyt: neato
Alias_ audreyt: We have to be really careful when using PERL5_CPANPLUS_IS_RUNNING
audreyt Alias_: mm?
Alias_ It gives false positives
Because the "RUNNING" isn't entirely accurate 03:40
audreyt right, kane mentions he's working on more fine grained reporting
Alias_ Yep, but the reason I hadn't written in support for that check already was the first time I tried, I got false positives and it was acting wrong
audreyt oy. I didn't know that 03:41
how do I reproduce the error? (I need to run for $job in a few mins, though, so better to mail me the recipe :))
Alias_ It's in ::With, but my tests were showing invalid results
I have a test dummy module up somewhere
audreyt markstos_: don't use multi
Alias_ hang on
markstos_ audreyt: Thanks for the tip. 03:42
audreyt sub foo (@a? %h?, *%nam) { my %arg = (%h, %nam, @a); say %arg.perl }
sub foo (@a?, %h?, *%nam) { my %arg = (%h, %nam, @a); say %arg.perl }
Alias_ audreyt: svn.phase-n.com/svn/cpan/trunk/PITA...akefile.PL
That's my "cpan dummy" module for testing ::With 03:43
You should be able to use that to test
audreyt k thx :)
Alias_ Once we get all those working (CPAN|CPANPLUS|manual) + (EU:MM|M:B) + (user|author) I'll start making everything else use it 03:44
audreyt that would be moosely 03:46
audreyt also needs to think about integrating Sage's work 03:47
Alias_ Which is that?
audreyt ./pugs some_perl5_script_here.pl 03:48
5->6 autotranslator
requires perl 5.9.4's MAD support
Alias_ Integrating it into M:I?
Or do you mean pugs
audreyt no, into pugs :)
Alias_ ah
markstos_ audreyt: following your suggestion, but it seems when passing a hash, it gets treated an array of pairs rather than a hash.
audreyt markstos_: nopaste code? 03:49
pasteling "markstos" at 12.176.101.89 pasted "more on passing hashes, hashrefs and arrayrefs" (13 lines, 196B) at sial.org/pbot/19248
markstos_ unrelated, my attempt at improving the Perl6 documentation has gone live: 03:51
feather.perl6.nl/~agentzh/syn/
lambdabot Title: Index of /~agentzh/syn
markstos_ I tried to turn that page into more of a central resource.
audreyt markstos_: indeed without mlultis it looks kinda hard 03:54
multi foo (Array $a) { foo(hash($a.values)); }
multi foo (Hash $h, *%nam) { my %arg = (%nam, $h.pairs); say %arg.perl;
}
would work
not pretty, though
will think about it some more :) but gotta run now
markstos_ tests audreyt's code.
Thanks audreyt++!
audreyt markstos_: on the portal page: I wonder if svn.openfoundry.org/pugs/docs/Perl6...rences.pod can be pod2html onthefly 03:55
aso singular of Exegeses is Exegesis 03:56
but looks very pretty :)
svnbot6 r12641 | markstos++ | typo fixes. audreyt++ 03:59
markstos_ I'm hoping agentz can help me Differences.pod into HTML. 04:00
I also added some "smart links" to Differences.pod that I hoped will start working. :)
pasteling "markstos" at 12.176.101.89 pasted "methods and subs dispatch an arrayref differently. Bug?" (29 lines, 507B) at sial.org/pbot/19249 04:13
markstos_ I need to go, but someone may be interested in reviewing my nopaste-- it appears to be bug where multi sub and multi method behave diferently. 04:14
audreyt MacVince: foo(@a) not foo(\@a) 04:24
er
s/macvince/markstos/
please not be using \ anymore :)
Khisanth the habits are going to be hard to kill 04:29
agentzh markstos++ 05:22
nice to see things had magically improved when i got up.
markstos: i'll autoHTML-ize Differences.pod on feather today. :) 05:24
agentzh now has a long TODO list for himself. 05:25
gaal moosees 05:39
svnbot6 r12642 | gaal++ | * t/README - English fix 06:07
agentzh feels like trying out mod_perl on feather. 06:11
svnbot6 r12643 | agentz++ | [docs/feather/~/index.html] 06:23
r12643 | agentz++ | - updated this file using gen.pl on feather.
svnbot6 r12644 | lwall++ | all_parse.t didn't interpolate filename correctly on failure. 07:07
svnbot6 r12645 | lwall++ | Handle aBOMinations like the one currently in t/subroutines/param_signature.t 07:16
r12645 | lwall++ | (Also handle Unicode whitespace in general, while we're at it...)
TimToady audreyt: that change uses oneOf. Could probably be faster by short circuiting characters known to be in the ascii range. 07:17
but my Haskell-fu isn't that good. 07:18
I can't tell whether the 10% slowdown is due to that or due to more of all_examples.t working... 07:21
(and other tests people have added recently)
masak "latest release" is out of synch on rt.openfoundry.org/Foundry/Project/?Queue=270 07:40
it says "6.2.3"
lambdabot Title: Pugs -- OSSF
masak is someone able to fix that?
integral masak: what should it be? the version that's on CPAN? 08:31
ah, I see the problem, openfoundry thinks 6.2.12 < 6.2.3 08:34
svnbot6 r12646 | agentz++ | [util/smartlinks.pl] 08:56
r12646 | agentz++ | - updated the comments
dolmans ?eval my @subs = (sub foo { $_ * 2 }, sub bar { $_ * 3 }); 10:03
evalbot_12646 Error: unexpected "f" expecting comment, subroutine parameters, trait or block
dolmans or the Synopsis outdated? though the sub name is meaningless there. 10:04
masak integral: ah, so that's why. still strange though, because we didn't skip from 6.2.3 to 6.2.12, did we? 10:21
integral no, but I guess no one bothered to create the versions on openfoundry
it's not really part of the CPAN release process at all
masak integral: it's a minor thing, but still it's good if the version number is correct
methinks
integral *nod* 10:22
masak is there some sort of "ultimate" version comparison module somewhere?
seems like there's always corner cases in version comparison
what with 6.2.12 > 6.2.3 10:23
and "alpha" and "beta" suffixes
and so on, etc
integral sadly not. even version.pm is too picky at times
agentzh masak: pugs' Test.pm can compare 6.2.12 and 6.2.3. :) 10:25
masak: but not with "alpha" or "beta" suffix. 10:26
audreyt first-class version object would be a win
dolmans: named sub exprs is not yet supported in pugs 10:28
dolmans audreyt: gotcha 10:29
audreyt goes looking at TimToady's haskell haxx0r 10:30
TimToady: I think isSpace is already uunicody 10:31
TimToady: but FEFF is not space 10:32
I'll revert and implement my Haskell-prime encoding detection proposal :)
# hackage.haskell.org/trac/haskell-pr...gDetection 10:33
lambdabot Title: SourceEncodingDetection - Haskell Prime - Trac
audreyt (which is, in part, based on perl5's encoding detection)
leo audreyt: from which publically readable dir on feather should I start svnbot.pl? (I don't want to have another pugs repo in ~lt) 10:38
audreyt ~audreyt/pugs 10:39
but I use URL anyway
(also it's not a svn checkout)
leo thanks 10:42
audreyt np :)
agentzh is there any commandline utility for Pod::Simple::HTML?
audreyt alias pod2html='perl -MPod::Simple::HTML -e Pod::Simple::HTML::go' 10:43
agentzh audreyt: yeah, i know the one-liner. but there're too few options. :) 10:47
for example, css file, index control, url prefix used by links, charset, title, and etc etc...
so i'll start util/podhtm.pl if there's no such script. :) 10:48
A lot of excellent features of Pod::Simple::HTML are just undocumented, which is really an unfortunate. :) 10:49
audreyt please do feedback to allison so she can include it in the next release 10:51
of Pod::Simple
agentzh audreyt: okay 10:53
svnbot6 r12647 | audreyt++ | * DrIFT.YAML: Make fromYamlBuf strict to save some more heap. 11:06
r12648 | audreyt++ | * Revert TimToady++'s Lexer patch as isSpace is already 11:09
r12648 | audreyt++ | Unicode-aware.
r12648 | audreyt++ | * To handle BOM on UTF8 files (and UTF16, UTF32, etc),
r12648 | audreyt++ | implement my source-code-encoding detection algorithm
r12648 | audreyt++ | proposal for the next Haskell standard:
r12648 | audreyt++ | hackage.haskell.org/trac/haskell-pr...gDetection
r12648 | audreyt++ | Currently only UTF8 is handled; UTF16 and UTF32 throws
r12648 | audreyt++ | an exception.
lambdabot Title: SourceEncodingDetection - Haskell Prime - Trac
agentzh markstos: with the help of util/podhtm.pl, smartlinks in Differences.pod now work perfectly. (haven't uploaded to feather) 11:29
svnbot6 r12649 | agentz++ | - added util/podhtm.pl which is a POD to HTML
r12649 | agentz++ | converter based on the excellent module
r12649 | agentz++ | Pod::Simple::HTML.
r12650 | audreyt++ | * Support for UTF32LE, UTF32BE, UCS2LE, UCS2BE.
r12650 | audreyt++ | Surrogate support coming soon...
r12651 | agentz++ | [util/podhtm.pl] 11:35
r12651 | agentz++ | - small tweaks
r12652 | agentz++ | [docs/feather/syn_index.html] 11:44
r12652 | agentz++ | - switched to the HTML version of the ``Differences''
r12652 | agentz++ | document
r12652 | agentz++ | - added notes on the resynchronization cycles from the
r12652 | agentz++ | Pugs repos. feather++
r12653 | audreyt++ | * Supprot for UTF16 surrogates: Now we can parse UTF16LE,
r12653 | audreyt++ | UTF16BE source code, with and without BOMs. That finishes
r12653 | audreyt++ | all canonical Unicode (8/16/32) encodings. :-)
r12654 | agentz++ | [docs/feather/syn_index.html] 11:46
r12654 | agentz++ | - small wording fixes
r12655 | audreyt++ | * add -fglasgow-exts to Pugs.Parser.Program to silence warnings
agentzh wow...
audreyt :) 11:47
audreyt is a happy unicoder
agentzh is a happy htmler 11:48
wolverian is a tired lama
llama I mean .. cough
audreyt O wolverian lama
wolverian :-) 11:49
agentzh markstos: Differences.html is now updated by feather for every 1 hour, and the smartlinks such as L<S04/...> in it are also working properly: 11:51
feather.perl6.nl/syn/index.html
lambdabot Title: Official Perl6 Documentation
agentzh and also: feather.perl6.nl/syn/Differences.html
lambdabot Title: Perl6::Perl5::Differences
agentzh thanks to podhtm.pl. :) 11:52
wolverian perlcabal.org/syn/Differences.html looks more official ;) 11:53
lambdabot Title: Perl6::Perl5::Differences
agentzh perlcabal?
gaal moose moose 11:54
agentzh oh, perlcabal has exactly the same IP as feather.perl6.nl. :)
gaal: got some cycles? 11:55
gaal yup! 11:56
agentzh i just can't understand why feather.perl6.nl/~agentzh/syn/ is forbidden while feather.perl6.nl/syn/ is not.
lambdabot Title: Index of /~agentzh/syn
agentzh they're both symlinks pointed to ~agentzh/syn. 11:57
wolverian agentzh, ~/public_html can't usually refer to outside the public_html
(I mean, up the path)
agentzh wolverian: oh...
is there any workaround? 11:58
wolverian I assume you can configure it, but you probably need to edit apache's own configuration to do it, i.e. you can't do it from .htaccess
(I'm assuming though.. and in any case, you can give .htaccess the right to modify the setting in apache's config)
agentzh no problem, i have root access.
wolverian :-) 11:59
agentzh but i don't know how to configure that...well... 12:00
wolverian, gaal: do you know how to do that?
agentzh btw, i hope people will like my podhtm.pl. ;-) 12:02
...and even contribute to it. 12:03
gaal agentzh: sorry, don't know off hand 12:04
agentzh gaal: no problem. :)
wolverian I don't remember either. 12:05
gaal this probably can help: httpd.apache.org/docs/2.0/howto/public_html.html
agentzh looking 12:07
svnbot6 r12656 | agentz++ | [docs/feather/index.html] 12:16
r12656 | agentz++ | - added a link to feather's Synopsis index to feather's
r12656 | agentz++ | homepage.
svnbot6 r12657 | agentz++ | [docs/feather/index.html] 12:22
r12657 | agentz++ | - adjusted the page layout.
agentzh oh, evalbot has been killed.
?eval 1+2
okay, i've found the answer in feather's httpd.conf: 12:25
Options SymLinksIfOwnerMatch
agentzh do i need to kill feather's apache and restart it so as to make my changes to httpd.conf take effect? 12:43
agentzh yay, it works! 12:49
the FollowSymLinks option and the command "apachectl -k graceful" have solved my problem completely. :)
gaal++ wolverian++ 12:50
wolverian yay :) 12:51
if you're using debian's apache, I think the preferred method is 'invoke-rc.d apache restart', or so 12:52
or reload
agentzh debian's apache? 12:53
etc/apache2?
wolverian the one installed with apt-get :) 12:54
i.e. not one that you've built locally
agentzh ah
wolverian the startup/restart/etc scripts are in /etc/init.d 12:55
agentzh has a lot of things to learn.
wolverian I've aliased 'service' to 'sudo invoke-rc.d' :)
agentzh okay 12:56
gaal audreyt: in the source encoding detection proposal, how does xs@[_] work? and xs@[0x00, _] etc.? 12:58
is this new haskell prime syntax? 13:01
Limbic_Region wonders if he was the only one suprised and a little dismayed by Ponie buying the farm 13:12
[particle] i certainly wasn't. 13:13
Limbic_Region perhaps I shouldn't have been 13:17
I normally would have thought that it would have been offered up as a community project prior to getting the death blow
but then again, the smart folks making decisions probably realized that too many such projects spreads the volunteer work force too thin 13:18
[particle] there was an offer some time ago for a new lead. nobody bit
Limbic_Region particle - that's not the story I got
Limbic_Region pulls up the email
audreyt gaal: no, it's just regular pattern matching 13:19
Alias_ This is open source, you can spread as thin as you like, if you are ok with things slowing down for a while 13:20
Look at Module::Install...
It's gone through a slow period while people are on other projects
Limbic_Region the only email I have is from Andy Lester saying he was trying to get the project fired up again but I also recall a conversation on IRC with someone on the inside that said they were sifting through candidates and an announcement would be made soon 13:21
*shrug*
audreyt hmm? new ponie news?
Limbic_Region it's dead 13:22
audreyt url?
Limbic_Region perl6.announce
audreyt (I know it's dead, just havn't seen the obituary)
Limbic_Region fetches a url
[particle] bbiab
Limbic_Region oh, it's on pugs blogs
that's where I first saw the headline
Limbic_Region planetsix.perl-foundation.org/? 13:23
audreyt ahh. www.nntp.perl.org/group/perl.perl6.announce/524
it's not on pugs blogs
lambdabot Title: nntp.perl.org - perl.perl6.announce (524)
Limbic_Region audreyt - I meant the "journals" link from the pugs homepage 13:24
they are synonomous in my browser the way I have the links set up
audreyt gotcha
Limbic_Region so no chance to work on the weird Win32 IO thingy yet?
audreyt obra++ # still preserved the "organ harvesting" meme
Limbic_Region has been conquering weird DNS problems the last 2 days for $work
audreyt Limbic_Region: which test is that? :) 13:25
Limbic_Region audreyt - t/examples/all_parse.t
the comment that says this is a horrible workaround that needs to be removed ASAP but is the only way Win32 currently works
audreyt if $*OS eq any(<MSWin32 mingw msys cygwin>) { $f.debug = 1;
}
this?
Limbic_Region $f.debug = 1; # should be completely unnecessary
yeah 13:26
Limbic_Region in a nutshell, in File::Find if there are not 4 say "something" strategically placed - it bombs on Win32 13:26
interestingly enough say "" does not work the same as say 1; # the latter working for some definition of work 13:27
audreyt yes, I remember that. booting into win32 now
Limbic_Region but running pugs -d on this particular example on Win32 produces no debug output (at least for me)
audreyt - fwiw, it was confirmed by another Win32 user 13:29
cjeris that is
Limbic_Region oh, you mentioned trying to upgrade to GHC 6.5, I might just do that as I have a few minutes 13:32
except www.haskell.org/ghc/download.html apparently doesn't list it - I'll check the IRC log as I thought you posted a link too 13:33
lambdabot Title: The Glasgow Haskell Compiler
audreyt oh no, my paralels image no longer boots 13:34
[particle] great. she drops the laptop running os/x, and windows breaks.
audreyt disk0s2: 0xe0030005 (UNDEFINED). 13:35
that sounds like a HD corruption.
[particle] sure does.
Limbic_Region well, perhaps the head scratched as the laptop plummeted to the floor
but knowing how important it was for you to continue your work
audreyt macbook supposedly has antiravity
I mean, gravity detection
Limbic_Region God saw fit only to damage the Win32 side
audreyt I wonder what's the standard procedure for this 13:38
gaal audreyt: so [1,2,3] in patmatch is the same as (1:2:3:[]) ?
audreyt gaal: yes
Limbic_Region gets brave and attempts to build 6.5 on Win32 instead of waiting for the binaries
agentzh audreyt: btw, evalbot is down. so the feather site is no longer up-to-date. :) 13:38
audreyt Limbic_Region: no, use the vinaries
www.haskell.org/ghc/dist/current/dist/
lambdabot Title: Index of /ghc/dist/current/dist
audreyt scroll to bottom
agentzh: evalbot is down but it will still autoupdate :) 13:39
agentzh oh? checking now...
audreyt: yes, you're right. :)
Limbic_Region audreyt - unless I am missing something, there is no way to tell where the binaries are as they all have the same naming convention 13:40
audreyt www.haskell.org/ghc/dist/current/di...w32.tar.gz
is the one you want
"mingw32" is your platform
Limbic_Region thinks that is what he was downloading
audreyt 20060823 is the latest snapshot
Limbic_Region over 60MB for a binary distribution 13:41
audreyt it's 45MB
Limbic_Region oh, must have grabbed the wrong date then
audreyt there's nothing there that's over 60mb, but some of them came close
Limbic_Region audreyt - ghc-6.5.20060101-i386-unknown-mingw32.tar.gz is listed as 61MB 13:42
just to nit
audreyt ahh you grabbed the first one
sorry, were scrolling backwards
indeed you are right :) 13:43
Limbic_Region yeah, wasn't realizing they were in ascending order but that makes sense
Limbic_Region believes his brain was damaged (hopefully not permanently) by the sustained high fever for the 2 weeks he was dealing with the acute symptoms of mononucleosis
audreyt oy 13:45
agentzh where should i put the timestamp in the Synopsis page? at the very begining, right after the `index', or in the VERSION section? 13:53
audreyt very beginning 13:54
I think
agentzh audreyt: okay
Limbic_Region audreyt - that particular snapshot does not include bin\ghc.exe ? 13:54
did the binary get renamed or something?
gaal Limbic_Region: read INSTALL 13:55
lambdabot Most guys don't understand me, but we really seem to connect.
Limbic_Region ahhh 13:56
lambdabot Whatever
Limbic_Region given that it was a binary installation I would have assumed installation wasn't necessary
Limbic_Region bonks himself in the head with a cement brick
gaal and you're blaming mono for brain damage....
Limbic_Region used the .msi last time 13:57
gaal - there is no configure so I will dig around somemore but if you happen to know of some Win32/MinGW install instructions - please let me know 13:59
gaal sorry, no haven't used that patform in a while 14:00
try #haskell? 14:02
Limbic_Region can't get there from here
Limbic_Region can only access channels/networks that feather allows 14:02
audreyt Juerd: add #haskell to cgi-irc please? :) 14:04
svnbot6 r12658 | audreyt++ | * "use" and "require" of modules that has different encoding
r12658 | audreyt++ | now works again.
Juerd audreyt: You have root :) 14:05
audreyt: I'm a bit busy getting my life back on tracks these days. I'm delegating as much as I can. Sorry about that.
audreyt sure. "you have root" is good enough response
gaal added 14:06
agentzh Juerd: i'm adding timestamp to smartlinks.pl. audrey said it had better appear at the very beginning of the HTML page.
[particle] #jifty and #svk would be nice, too
Limbic_Region ok - the fact that there is no configure anywhere to be found leads me to believe this no workey
Juerd agentzh: Somewhere at the top is a great place indeed.
gaal adding. (are they both both on freenode?)
[particle] yes
agentzh okay, soon you'll see the effect. :)
Juerd gaal: The "on" thing was my own hack. It no longer works since a system upgrade.
gaal: We're back to servers and channels.
Which it sees as two separate arrays, with no interrelation :| 14:07
Limbic_Region ok, heading over to #haskell now - thanks everyone
Juerd Limbic_Region: The thing lets you be in multiple channels, afaik...
"/join #haskell"
audreyt Limbic_Region: a sec. I have a hypothesis about ghc win32... 14:09
Limbic_Region Juerd - neato
audreyt - question already posed in #haskell but if you get to it first - great
it appears the channel is preoccupied with something ATM anyway 14:10
audreyt Limbic_Region: my hypothesis is that the 20M of reduced size 14:13
corresponds to a broken build that did not have ghc.exe and core binaries
svnbot6 r12659 | agentz++ | [util/smartlinks.pl]
r12659 | agentz++ | - added obvious timestamps to the resulting HTML pages
r12659 | agentz++ | according to the suggestion of Juerd++.
audreyt which means the last known-good version is www.haskell.org/ghc/dist/current/di...w32.tar.gz
so please use that instead. 14:14
Limbic_Region audreyt - my thought too except according to the build log - it worked
Limbic_Region needs to wander off but will be back
audreyt I checked the build log, and I suggest not trust it
gaal incomplete download?
insufficient disk space at unpacking site?
biteating worms?
audreyt apparently I need to boot with macosx dvd to fix the disk 14:17
agentzh Juerd: please comment on the brand new timestamps: feather.perl6.nl/syn/S02.html :) 14:18
lambdabot Title: S02
audreyt Limbic_Region: sorry that I don't have a win32 anymore to fix the win32 problem -- I'll see what the apple repair people say tomorrow and hopefully I can get my old laptop back sufficiently to continue working on this 14:20
audreyt decides to sleep instead of trying potentially increasingly destructive repair mechanisms 14:21
g'nite :) hopefully the biteating worms won't find their ways to my OSX data tonight...
agentzh audreyt: sleep well. ;) 14:22
gaal night audreyt 14:23
[particle] agentzh: i like the timestamp. is it possible to include the svn rev number? 14:24
agentzh particle: oh, that's easy. :)
adding now...
particle: but there's problem: where should i put the revision number? 14:25
*a 14:26
[particle] how about in parens, after time
agentzh but that way may be confusing
we have different rev number for pugs and syns
[particle] actually, i want both.
agentzh is shocked.
[particle] so it can be determined what the sources were. 14:27
agentzh well, i don't know where the pugs rev number stores.
is it in the pugs tree?
[particle] svn info
agentzh i know the synopsis rev number in at docs/Perl6/Spec/ 14:28
audreyt ./pugs -v |grep Version
[particle] i know it's in the parrot config file
ah, great.
agentzh ah, okay.
Juerd agentzh: I like the timestamp 14:29
agentzh Juerd: glad to hear that. :)
Juerd agentzh: May I place a request for the time format? :)
agentzh Juerd: oh, not really. :(
[particle] perhaps you could format it as C<< <br/>using pugs revision 999 and synopsis revision 999. >>
Juerd Okay then
agentzh particle: it's getting longer and longer, i'm afraid. :) 14:30
[particle] make the font smaller :)
Juerd pugs r999, syn r999 :)
[particle] sure, that works fine
agentzh good
particle: would like to add that to smartlinks.pl? 14:31
Juerd I'm always for concise info :)
agentzh i'm a bit tired.
Juerd As long as it's clear enough :)
[particle] agentzh: i'll try. but i haven't run smartlinks.pl yet
agentzh i've been fight with feather these days. :)
[particle] builds a recent pugs 14:32
Juerd agentzh: Has feather been naughty?
agentzh Juerd: no, feather is cute.
the ssh terminal here is terribly slow. :(
Juerd Aw
agentzh particle++ 14:33
particle: the synopsis rev info is in the file docs\Perl6\Spec\.spec-revision. 14:34
[particle] thanks 14:35
markstos agentz: Move the timestamps to the bottom of the page. Putting them at the top makes look they are most important thing to see.
The most prominent heading on the page should be "Synopsis 2: Bits and Pieces", because that's the key piece of information.
agentzh markstos: to be honest, adding timestamps as html footer is much easier, but more poeple here tend to see the time at the very begining. :) 14:36
*people
Juerd markstos: The information is very relevant at the beginning. Not because it's key info, but because it's very dynamic, and indicates that the system works. 14:37
markstos: Besides, POD starts after the index, so people scroll down anyway :)
markstos markstos: That's just my two cents as a professional website developer with some graphic design training... 14:38
markstos accidently replies to self. Weird.
Juerd markstos: In a marketing sense, you're absolutely right; but our goals differ... These documents are for interested people mainly... 14:39
afk
Limbic_Region attempts ghc 6.5 from a snapshot earlier in the month
[particle] C<nmake unoptimized> fails :(
agentzh particle: you don't need to build pugs at all. :)
[particle] oh? then how do i get the version info?
agentzh util/smartlinks.pl doesn't require a working pugs.
audreyt [particle]: fails how? 14:40
agentzh i think the version info must be stored in the source tree.
[particle] Writing new package config file... done.
Installation failed for filepath at util\build_pugs.pl line 164.
NMAKE : fatal error U1077: 'C:\WINDOWS\system32\cmd.exe' : return code '0xff'
Stop.
agentzh hasn't built pugs for days.
audreyt indeed the revinfo is src/Pugs/pugs_version.h
so you can read that instead
[particle] ah, great. i can read the header
audreyt [particle]: nopaste a more complete log? 14:41
pasteling "[particle]" at 144.81.84.139 pasted "win32 make failure" (85 lines, 3.9K) at sial.org/pbot/19251 14:42
audreyt fixed. 14:43
Limbic_Region audreyt - you are correct WRT snapshots being b0rk
Limbic_Region doesn't know who the right person to tell is so he will just throw it out in #haskell 14:44
audreyt [particle]: r12660 should fix nmake
markstos Juerd, check out how the darcs wiki works, which is moinmoin powered. They do included the changed timestamp on every page, but it's grayed out the bottom, so you have to look for it: wiki.darcs.net/ I think that's sufficient if anyone wants to check the freshness. 14:46
svnbot6 r12660 | audreyt++ | * modify build_pugs probing for older win32 Cabal
lambdabot Title: FrontPage - DarcsWiki
markstos Juerd, agentz: also, "S02.html" can removed from the timestamp, since it just repeats the URL. Something more useful would the source filename, like "Perl6/Spec/Operators.pod". 14:47
anyway, I'm grumpy and should get back to work.
audreyt markstos: kindly hack util/smartlinks.pl however you see fit :) 14:48
markstos Thanks to you both for your help with the documentation!
[particle] audreyt: fixed. make continues...
audreyt++
audreyt [particle]++
markstos audreyt: Agreed. I'll proceed in the wiki style once I return to FreeTime.
agentzh markstos++
audreyt speaking of FreeTime... I need to wake up early for $job tomorrow too 14:49
audreyt fades
cjeris $job-- 14:50
agentzh takes a break. 14:55
Limbic_Region particle - you on Win32 ATM? 14:58
[particle] yes 14:59
crap, another build error
Limbic_Region me too
thought it was related to my recent switch to ghc 6.5 15:00
[particle] util\build_pugs.pl :383 ?
Limbic_Region just a sec
no, mine looks different
let me nopaste
[particle] it seems there's a lot of assumptions in pugs now that Judy is installed 15:01
[particle] i don't have Judy 15:01
agentzh particle: Judy is in the pugs source tree. 15:02
[particle] is it required, or optional and broken if you don't have it?
Limbic_Region she's a real sweet lady, but nearly as sexy as Perl
[particle] hrmm
agentzh particle: your pugs has errors while compiling Judy?
pasteling "Limbic_Region" at 129.33.119.12 pasted "Current Win32 build error" (42 lines, 1.9K) at sial.org/pbot/19252
Limbic_Region particle - see third-party
Limbic_Region believes that's where Judy is
pasteling "[particle]" at 144.81.84.139 pasted "win32 make error (Judy ??)" (197 lines, 13.4K) at sial.org/pbot/19253
agentzh particle: if yes, then that's my fault. :)
looking now 15:03
[particle] lr: are you building unoptimized?
agentzh it seems to me the C part of Judy has been successfully built, no? 15:05
Limbic_Region particle - I never build unoptimized 15:06
my Judy error is different than yours anyway
agentzh LR: "nmake fast"
Limbic_Region blames agentzh
how is fast different from unoptimized?
agentzh LR: you may have a try...much faster! 15:07
at leaset on my win32 box.
LR, particle: you two are all having problems with the HS part of Judy, rather than the C part. 15:08
i wrote the code which builds the C Judy on Win32.
Limbic_Region agentzh - faster build but is the resulting executable faster? 15:08
agentzh LR: of course, the resulting executable is *a bit* slower. 15:09
Limbic_Region I don't mind pugs taking an hour to build - it's all about the runtime
agentzh fast == unoptimized
LR: but pugs is changing very fast.
Limbic_Region particle - what version of ghc are you running
agentzh so i have to rebuild pugs over and over again.
Limbic_Region agentzh - I seldom have to build from scratch so keeping up to date is not bad 15:10
agentzh i got the same error as particle.
Win32, GHC 6.4.2 here.
Limbic_Region I had no problems with Pugs with GHC 6.4.2 but I literally just switched to 6.5 within the last half hour
agentzh i think only audreyt or cmarcelo can fix this kind of things. :)
Limbic_Region is wondering if it is related to that 15:11
agentzh has no clue
Limbic_Region I did try removing stuff from third-party as that has been the solution in the past
so I am rebuilding now
agentzh realclean now... 15:12
Limbic_Region actually - I think removing stuff from 3rd party did fix it as it is beyond the point of failure last time
if it fails this time, it will be for a new reason 15:13
agentzh hehe
Limbic_Region sticks his tongue out at particle for being left behind in the dust
is Win32 your primary environment agentzh? 15:14
agentzh LR: true
[particle] 6.4.1
Limbic_Region so how much do you know about haskell - specifically as it relates to pugs internals 15:15
[particle] cleans the puppy
agentzh knows very little.
Limbic_Region particle - if you feel like throwing chicken bones - try this rmdir /s third-party; sv[kn] up; perl Makefile.PL; nmake fast
agentzh it seems that realclean helps. 15:16
Limbic_Region agentzh - not in this case
agentzh why?
Limbic_Region the realclean often misses files in 3rd party that are perpetually added after it is realized they should be included too
audreyt 6.4.1 cabal specific problem. 15:17
fixed.
Limbic_Region the entire 3rd party removal is overkill btw - I just didn't know which files in there needed to be removed
agentzh LR: i think i added many "rm" like commands to pugs' main Makefile.PL
Limbic_Region ahh
audreyt - since you are either working in your sleep or not in bed yet - did you see my /msg
agentzh yes, removing the whole 3rd party is indeed overkill.
Limbic_Region yes, but cpu cycles are cheap - especially when I am wasting time chatting here 15:18
agentzh yay, i've successfully built my pugs!
svnbot6 r12661 | audreyt++ | * in GHC 6.4.1, the third-party/installed/ layout is different,
r12661 | audreyt++ | so we need to probe to another place to put Judy .o files
r12661 | audreyt++ | into the HsJudy .a file.
audreyt Limbic_Region: no I did not.
[particle] hih! 'nmake clean' fails due to 'line too long'
audreyt did you perhaps forget to identify to nickserv?
agentzh particle: "realclean"?
audreyt [particle]: svn up, rm -rf third-party, svn up third-party, make 15:19
[particle]: that sequence should work. sorry for not considering GHC 6.4.1
Limbic_Region no , but here it is just in case
audreyt WRT the Win32 IO problem (t/examples/all_parse.t) - I noticed this in the build (using GHC 6.5) "Could not find module `System.Console.Readline':"
[particle] agentzh: switched from nmake 1.50 to 7.1... much better
agentzh particle: nmake 7.10.3077 here. :) 15:20
[particle] yes, me too
just forgot to set my environment
Alias_ nmake, ugh :/
agentzh heh
Alias!
Alias_ Have you tried with dmake?
We're already swapping from nmake to dmake for Perl 5 on Win32
agentzh really? 15:21
Alias_ (since nmake is reporting to not work on Vista any more)
[particle] fabulous.
Alias_ At least, according to me Micrsoft Corporate Support flatmate
[particle] dmake?
agentzh microsoft is working on xmake.
Alias_ my
agentzh a ms guy told me that.
Limbic_Region audreyt - would the missing System.Console.Readline conceivably have anything to do with the IO problem or without Win32 you have no way of making that guess?
Alias_ What'sit do?
agentzh xmake is the next generation of nmake. 15:22
and the vc team is working on that.
Alias_ So it's going to break everything again?
Limbic_Region IOW - they are going to start shipping a make that works out of the box
agentzh presumbly. ;)
Alias_ And we'll need to download Visual Studio to use it? :)
Limbic_Region instead of the crappy 1.5
audreyt Limbic_Region: not going to metter
Limbic_Region k
Alias_ Limbic_Region: You could just use gnu make
agentzh the win32 version of gmake has a lot of bugs. 15:23
Limbic_Region Alias_ - I have 2 versions of nmake.exe, msys make, MinGW nmake32-make, and dmake
Alias_ As of the very latest 3.81 from last month, it incorporates the MinGW stuff by default
Limbic_Region I think I got it covered
agentzh i just can't remember how many bug reports i've sent to their mailing list.
Alias_ So it all works now
The new Strawberry Perl alpha 2 will use it
Limbic_Region the hardest part with having all those makes is remembering to set your environment correctly so they don't step on each other
for instance - if you have both mingw and msys in your path then building parrot using mingw doesn't work 15:24
Alias_ msys?
Oh that's the "extra tools" or something right?
Limbic_Region yes, msys - the companion to MinGW
www.mingw.org/ # second paragraph 15:25
lambdabot Title: MinGW - Home
Alias_ We've managed to avoid it so far
Limbic_Region audreyt - GHC 6.5 on Win32 does not fix the heap problem fyi 15:26
Alias_ audreyt: Why aren't you alseep
hang on, why aren't _I_ asleep
Limbic_Region she is
it looks like pugs is going to successfully build using ghc 6.5 - as soon as it is finished I am off to find lunch 15:27
[particle] rebuilds pugs 15:28
Limbic_Region yep - pugs successfully rebuilt 15:35
Limbic_Region svn ups just to be sure the latest checkin(s) didn't mess things up 15:35
agentzh gaal: does `make smoke' use Test::TAP::Model to generate a .yml file? 15:37
gaal: i think smartlinks.pl itself shouldn't serve as a test harness anyhow. 15:38
gaal: it should read an existing .yml data file and represent that in the HTML pages.
of course, the current tests.yml is far from ideal. The serialization method of Test::TAP::Model should be used. :) 15:39
ah, Test::TAP::Model looks very nice. 15:40
Limbic_Region heads off to lunch
agentzh gaal: can you manage to obtain a typical .yml file from a real smoke which i can work on directly? 15:44
agentzh waves &
svnbot6 r12662 | agentz++ | [docs/feather/syn_index.html] 15:57
r12662 | agentz++ | - s/between Perl and Perl6/between Perl 5 and Perl 6/
gaal Test::TAP::Model doesn't have a native serialization format. But tests.yml that yaml_harness produces just happens to be contain a node (called 'meat') that TTM knows how to construct an instance from (see new_with_struct, util/testgraph.pl line 25) 16:13
i'll start a smoke to send you.
gaal audreyt: doesn't your autobuild pugs run a smoke? 16:14
TimToady I just did a smoke. 16:19
TimToady m19s28.vlinux.de/iblech/stuff/pugs-...9bf23.html 16:20
lambdabot Title: TAP Matrix - Thu Aug 24 15:58:13 2006 GMT
gaal TimToady: could you scp your tests.yml to feather or somewhere else agentzh can moose it? 16:21
TimToady It would be lovely if smokes were marked with deltas from the last smoke...
gaal yes! 16:22
gaal for a while someone put 'make test' outputs on a mediawiki 16:22
and nothingmuch was working on aggregating smoke results, but I don't know if that included history 16:23
[particle] you want smokestat, similar to diffstat
TimToady I put tests.yml into wall.org/~larry/tests.yml
gaal thanks, TimToady 16:24
the smokeserver currently works on client-rendered html matrices
we should update to use Best Practical's code, which moves the rendering to the server side 16:25
that way the serialized TTM data -- suitable for diffing/aggregating/otherwise moosing is all in one place
TimToady the rendering cranks the CPU for quite a while...
so you'd want to cache it.
gaal yeah, that should be fixed too :)
TimToady fixing is also good. :) 16:26
gaal one WTDI (resulting in smaller html, too!) would be to bite the bullet and move to javascript
TimToady now where did I put that JS backend for pugs... 16:27
gaal gotta run off, bye folks 16:28
TimToady ich auch. & 16:29
moi aussi. & 16:30
watashi mo &
kai ego &
gaal הגם אתה, לארי?
TimToady heh 16:31
but you spelled it backwards. :)
[particle] what interesting word usements you structure
TimToady anyway, going for a walk. & 16:32
dazjorz <b_jonas> and then there's this other bot that prints titles of urls on #perl6 16:59
lemme try that...
www.google.com/search?hl=en&q=r...gle+Search
lambdabot Title: random text string that shorten will shorten - Google Search
dazjorz Yes! :D
Limbic_Region wow - pugs on ghc 6.5 is MUCH faster 17:49
Limbic_Region in the past, what did a plus in front of a variable in a sub signature indicate? 17:56
svnbot6 r12663 | fglock++ | * v6 - working Pugs/Grammar/P6Term.pm; can be tested by uncommenting
r12663 | fglock++ | the 'use' line at P::G::Expression.pm;
r12663 | fglock++ | - included the precompiled P6Term.pmc, which is needed to bootstrap
Limbic_Region that it was optional?
s/signature/prototype/ 17:57
or not
no one remembers? 18:00
?eval my $foo = <HellO>; $foo.lc 18:07
evalbot_12663 "hello" 18:07
svnbot6 r12664 | Limbic_Region++ | Parse issue: lc - fixed 18:17
r12665 | Limbic_Region++ | Fixed example not to assume being ran from . 18:26
r12666 | Limbic_Region++ | Added -Iblib6/lib as many examples use ext/
r12667 | Limbic_Region++ | Added 'p5' to directories skipped as examples include p5 code too 18:32
Limbic_Region ?eval my ($true,$false,$string,$defined,$last); sub infix:{qn'\\'} ($lhs,$rhs){ $rhs // $lhs }; $last //= [\\] ($true,$false,$string,$defined,$last); $last; 18:36
evalbot_12666 Error: unexpected "[" expecting comment, "\\(" or term 18:36
Limbic_Region thinks that is a pugs parse bug - can anyone confirm/deny?
[particle] does [\\] operate on a list, or array? 18:38
Limbic_Region particle - the infix definition is in the ?eval
TimToady actually, you can't have a [\\] reduce operator.
TimToady now that I think about it... 18:39
Limbic_Region not my code btw - I am trying to clean up examples
TimToady - see examples/cookbook/01strings/01-02default-variables.pl
[particle] can {qn'\\'} be replaced with <'\\'> ??
Limbic_Region line 76
TimToady yes, I know, I saw it recently.
and it didn't even occur to me at the time that [\\] was illegal 18:40
[\...] is a metasyntax
Limbic_Region TimToady - also, what did it used to mean when there was a plus in front of a variable in a sub's signature?
TimToady that's the old named arg notation. would be :$foo now
Limbic_Region k - will fix 18:41
TimToady no, I'm wrong about [\\] 18:42
S03 says it chooses the infix operator in case of ambiguity
so that's a reduce of \\ if there is a \\ operator
and the triangular version would be [\\\] 18:43
the tricky thing is that the name of the \\ operator is infix:<\\\\>
since \ has to be backwhacked in <...>
Limbic_Region so why is that failing to parse correctly then I wonder 18:44
TimToady which is why I think I changed it to infix:{q:n'\\'}, if I recall
I think infix: and such aren't correctly modifying the parser currently.
It's not recognizing Unicode operators either. I noticed the set operators didn't recognize ∩ and such 18:45
also, it's possible q:n isn't implemented yet... 18:46
svnbot6 r12668 | Limbic_Region++ | Changed old named arg +$var to :$var - TimToady++
TimToady no, it is implemented, yay
Limbic_Region I still can't get any incantation to parse as far as pugs is concerned so I am moving on 18:47
TimToady so is qn//
because [\\] is not going to be recognized if \\ isn't also.
anyway, I tried a lot of permutataions too. I agree with moving on for now 18:48
one thing you could do is slurp the whole file in each time and examine it to see if it contains 'use v6', then eval (if that works). That would get rid of the failures with p5 not be recognized. 18:49
alternately, go through and rename all the p5 code with -p5 in the name
but that'll probably break some things.
Limbic_Region TimToady - I started to do s/script-p5.pl/script.p5/ 18:50
I wasn't sure of the proper way of removing the original file from the repo though so I left it alone
[particle] svn rm foo
Limbic_Region down to less than 30 examples not parsing
TimToady I think that .p5 was originally proposed, and for some reason was rejected. don't remember 18:51
which is why we ended up with -p5.pl
Limbic_Region if you want white space between the hash name %hash and the key <key> you need to using a leading . right as in %letter .{$hostname} = $letter; # ?? 18:52
TimToady but if eval can be made to test parsability, then I don't see why slurping and checking for "use v6" wouldn't work.
and then we could test any text file, not just ones ending in a particular extension.
Limbic_Region true
TimToady (unix programmers being loathe to put .pl on a file just to make it executable for Windows folks...) 18:53
PerlJam TimToady: Have you listened to the recent perlcast interview with pmichaud? There's some major kudos for you in there :)
Limbic_Region ?eval my %hash = 1..4; %hash .<3>;
evalbot_12668 Error: unexpected "." expecting comment, operator, postfix conditional, postfix loop, postfix iteration, ";" or end of input 18:53
Limbic_Region that's a pugs parse bug isn't it? 18:54
or has that whitespace thing changed too
Limbic_Region at least 4 or 5 of the examples are blowing up with the pugs error: pugs: Control.Concurrent.STM.atomically was nested 19:03
fglock mini-blog: use.perl.org/~fglock/journal/30742 19:06
lambdabot Title: Journal of fglock (5723)
TimToady PerlJam: uh, no, I haven't listened to it. My head's already way too big. What's the url? :) 19:10
Limbic_Region: that would have to be written %hash\ .<3> now 19:11
PerlJam For the audio it's www.perlcast.com/audio/Perlcast_Int...ew_032.mp3 19:11
TimToady postfix ops *never* *ever* have whitespace before them 19:11
(but "unspace" doesn't count as whitespace)
PerlJam TimToady: in the beginning pmichaud sings the praises of perl, and then the regex-refactor :) 19:12
Limbic_Region Thanks TimToady 19:13
svnbot6 r12669 | Limbic_Region++ | Changed my $test_file = shift || 'test' to @*ARGS[0]
Limbic_Region fixing a lot of these isn't difficult
unfortunately that particular test is now resulting in pugs: Control.Concurrent.STM.atomically was nested - still going to check it in 19:14
TimToady well, we're probably s/atomically/contend/ at some point anyway, so don't sweat that too much. Probably just running into an unimpl. 19:15
or is the test not actually saying "atomically"?
svnbot6 r12670 | Limbic_Region++ | changed %hash .<key> to %hash\ .<key> TimToad++\nUnfortunately it now bombs with pugs: Control.Concurrent.STM.atomically was nested 19:16
[particle] shame on patrick. not once did he mention that pge needs tests :( 19:21
svnbot6 r12671 | Limbic_Region++ | Added FindBin::Dir as not guaranteed to be run from .
Limbic_Region TimToady - WRT atomically, I didn't dig in to the guts just running pugs -c 19:23
is the correct syntax for a p5 require 'file.pl' still require 'file.pl'; # ?? 19:24
PerlJam [particle]: you can mention it in your perlcast interview :)
Limbic_Region anyone have any idea what the following LOC might be trying to do: my $life = 10 .?. 20; 19:27
[particle] was that ever perl 6 syntax?
TimToady there's a prior def of infix:<.?.> 19:28
Limbic_Region dunno - I find it hard to keep up being a casual observer but it is used in more than one place in this example
TimToady again, that appears not to be working
Limbic_Region TimToady - you have already looked at all these then?
TimToady I've at least glanced at most of them.
Limbic_Region ok, will leave it then as it *should* work 19:29
what about the require 'file.pl'; ?
Limbic_Region that's the only HOP example I haven't been able to fix (though you fixed most of them) 19:29
svnbot6 r12672 | Limbic_Region++ | Changed old named arg +$var to :$var 19:35
r12673 | Limbic_Region++ | Used FindBin::Dir as not always run from . but require 'file.pl' still b0rk 19:38
Limbic_Region YAY - down to only 22 examples that aren't parsing 19:42
Limbic_Region is going to count how many have that STM problem 19:47
TimToady it's odd--I've never seen the STM problem.
I wonder if it's Windows related...
Limbic_Region likely 19:48
[particle] what stm problem?
Limbic_Region Win32 is the redheaded stepchild in perl 6 development
particle - many of the examples are having an stm problem - wait a sec and I will have a list 19:49
[particle] using pugs, or v6, or what?
Limbic_Region pugs
particle - for instance, try pugs examples/nested_loops/coroutine.pl 19:51
[particle] wow, that barfed 19:53
Limbic_Region pugs: Control.Concurrent.STM.atomically was nested right? 19:54
[particle] no
Limbic_Region oh, what did you get?
[particle] no output to screen, just ms error reporter trying to send data to ms 19:55
exception code 0xc0000005
Limbic_Region wonder if that's a difference in ghc version
*shrug*
[particle] *brr* 19:56
Limbic_Region particle - let me give you another example and see if it has the same exception 19:57
pugs -c examples/vmethods/time.pl 19:58
[particle] same
Limbic_Region ok, so same problem - different output 19:59
pasteling "Limbic_Region" at 129.33.119.12 pasted "The following examples on Win32/MinGW/GHC 6.5 bomb with: "pugs: Control.Concurrent.STM.atomically was nested"" (6 lines, 204B) at sial.org/pbot/19259 20:03
Limbic_Region audreyt - apparently more Win32 only bugs
cmarcelo Limbic_Region: nested error here too (linux64 + ghc 6.5) on time.pl at least 20:04
Limbic_Region good to know though about all I am good for is identifying the bugs - not fixing them 20:06
Limbic_Region well, I am content with examples for the moment anyhow - between TimToady and myself, more than half the original parse-fails are now passing 20:10
svnbot6 r12674 | markstos++ | typo fix for 'minutes' and other minor phrasing cleanups on syn_index.html 22:15