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.
00:02 lollan joined 00:20 neoesque joined 00:23 hikozaemon joined 00:29 nekokak joined 00:41 christo joined, christo left
svnbot6 r12532 | fglock++ | * PCR - uses plain insideout; Test.pm compiles with less than 80M ram; about 5% faster 01:06
xerox @arr 01:09
lambdabot I'll keel haul ya fer that!
xerox @arr
lambdabot Drink up, me 'earties
xerox @arr
lambdabot Avast!
01:11 justatheory joined
Patterner Avaunt! 01:11
01:24 SamB joined 01:46 zgh joined 01:54 dolmans joined 02:48 mjk joined 02:49 agentzh joined
agentzh is back now. 02:49
gaal: oh, coloring the tests using smoke results (just like catalog_tests.pl) is attractive. 02:50
but i think we should make it optional, since smoking pugs usually takes a long time, and the smoke results are implementation-specific, or even backend-specific. 02:51
today i'm going to make the HTML/CSS source generated by util/smartlinks.pl a bit nicer. :) 02:53
it seems to me fglock has been doing something amazing on v6.pm's optimization recently. :D 02:54
fglock++ fglock++ fglock++
agentzh goes back to hack. 02:55
gaal: i've seen your svn log. :) 03:01
i didn't get cycles last night to add snippet boundary. too much time was sepent on debugging JavaScript code yesterday. ;-) 03:02
i'll do it now.
svnbot6 r12533 | agentz++ | [util/smartlinks.pl] 03:07
r12533 | agentz++ | - updated the comments
r12534 | Sage++ | Lots and lots of code cleanups. Just about every function gets some better documentation and some whitespace syntactic sugar. 03:10
r12534 | Sage++ | Just doing some cleaning in anticipation of code review.
r12535 | agentz++ | [util/smartlinks.pl] 03:19
r12535 | agentz++ | - suppressed error messages if it's not running with the --check option (but with a hint).
r12535 | agentz++ | - tweaked the errro message format a bit.
TreyHarris agentzh: gaal and i were talking last night about pugs-specific tests, did you see that? 03:20
agentzh oh? looking now...
TreyHarris no, i can repeat 03:21
basically, apparently xx-uncategorized used to be called "pugsbugs" and was used for pugs-specific issues
its new name, and your comment in TASKS, suggest that all those tests should be on language issues--which leaves the question of where pugs-specific tests should go
agentzh so what's your opinion? 03:22
TreyHarris make a pugsbugs subdirectory, or create a separate t/ elsewhere
agentzh i think the motivation of isolating tests into "pugsbugs" was probably to get audreyt++ easier to fix pugs bug. 03:23
*bugs
but you're talking about testing pugs-specifc features, right? 03:24
TreyHarris no... well, it's not like such bugs shouldn't be runnable by other implementations, just that they might seem odd and/or redundant in the context of being a language-specification test suite 03:25
like the one i checked in last night, that just checks if ".say for %hash" works, since it was causing pugs to segfault.
agentzh TreyHarris: oh, well. my gut feeling is that pugs' bugs could probably become bugs of implementation X. :) 03:26
svnbot6 r12536 | Sage++ | More code cleanups. Documentation improved, formatting better, and all is right with my little corner of Pugs.
agentzh TreyHarris: that test can go into t/syntax/... or t/statements/for.t, i think. 03:27
TreyHarris gaal said that such tests that cause segfaults should be isolated and placed alone in a file 03:28
and suggested pugsbugs as the appropriate place, not knowing it no longer exists
agentzh TreyHarris: yes, that makes sense.
TreyHarris: the dir name xx-uncategorized or "segfault" is okay. 03:29
TreyHarris and also those kinds of tests should be handled a little differently. you shouldn't todo them in any circumstance, though audrey might choose to todo them just prior to a release if she wants to call out a known and as-yet-unfixed bug
agentzh *nod* 03:30
TreyHarris: go ahead and see how well it works in practice. :) 03:32
TreyHarris how about "critfails", or "critical" or something like that? let all impls run them, but just put them together so that a) segregating into lots of little files won't be a pain to the organization of the rest of the tests, and b) developers can pay special attention to hard failures in that space?
agentzh will look later. sorry, lunch time now & 03:33
03:33 agentzh left 03:39 SamB joined
cmarcelo audreyt et al: cmarcelo.blogspot.com/2006/08/pugs-...-judy.html && cmarcelo.blogspot.com/2006/08/sold-...ps.html... if you folks have any suggestion, new data, or something, please tell me... =) 03:40
lambdabot Title: cmarcelo: Pugs meets Judy
audreyt cmarcelo: would you be so kind and post the same article on pugs.blogs.com? :) 03:56
(the meets judy one)
o/~ hey judy, don't make it bad / take a slow pugs / and make it faster o/~ 03:57
obra heh
wolverian hrm. I think I grog typeclasses now. thinking of them as java interfaces is really disparaging. 03:58
evan_tech++ # great journal on haskell at evan_tech.livejournal.com/ 03:59
lambdabot Title: evan_tech -- Recent Entries
audreyt indeeed 04:04
cmarcelo audreyt: sure. how do I do it? 04:08
audreyt cmarcelo: check your email for a blog bit :) 04:10
wolverian: the oleghack he links to -- www.haskell.org/pipermail/haskell/2...14939.html -- is dazzling 04:11
lambdabot Title: [Haskell] Typeful symbolic differentiation of compiled functions
TreyHarris audreyt: do you have any opinion about the question of where pugsbugs should go?
audreyt TreyHarris: they should go squashed 04:13
but meanwhile, try to find relevant t/ to put them, and failing that, t/xx-uncateg.
TreyHarris ah, gaal said put them in separate files to themselves since they disrupt smoke reports otherwise 04:14
audreyt well, as long as it's not crashing.
if it is crashing, t/xx-uncateg is perhaps better
TreyHarris no, i'm talking specifically about crashing
ok 04:15
can evalbot survive a segfault? 04:16
i don't have a linux installation, so i've been tempted to try these as they come up against evalbot, but if i make it go away that would be bad
audreyt try it now? :) 04:19
TreyHarris ?eval my $i; $i++ for %?CONFIG; say $i
04:19 evalbot_12525 is now known as evalbot_12536
evalbot_12536 Error: Can't modify constant item: VUndef 04:19
audreyt you are not supposed to change %? hashes... 04:20
oh wait, it's $i not $_
nvm
TreyHarris that segfaults on my machine (also running r12536)
?eval %?CONFIG.elems 04:21
evalbot_12536 Error: No compatible subroutine found: "&elems"
TreyHarris oh, i thought hashes supported elems
TimToady ?eval %?CONFIG.pairs.elems
evalbot_12536 0
TimToady ?eval [%?CONFIG].elems
evalbot_12536 1 04:22
TimToady ?eval %?CONFIG.say
evalbot_12536 OUTPUT[ ] Bool::True
TreyHarris those all work for me. but %?CONFIG by itself in list context segfaults 04:23
audreyt investigating
fixed 04:25
TreyHarris ?eval my @a = @(%?CONFIG)
evalbot_12536 [undef,] 04:26
TreyHarris (i realize that's incorrect. but it segfaults for me)
svnbot6 r12537 | audreyt++ | * TreyHarris++ pointed out that listifying Hash was
r12537 | audreyt++ | crashing when it shouldn't, apparently a regression
r12537 | audreyt++ | from Judyizing the interface.
TreyHarris making
audreyt++
cmarcelo haskellers: do you know hot to "export" instances? in CollectionsM I instantiate MapM (Data.Map.Map) and I thought that when I imported both Data.Map and CollectionsM the instance would be implicitly working... but doesn't looks like =P (actually I'm pretty sure that yesterday it works but doesn't today ...) 04:27
audreyt: (i got the blog bit, i'm reviewing a bit and will commit into blog) 04:28
audreyt cmarcelo: you don't export instances, they are assembled automagically 04:30
almost makes me wonder if we can allow separate "does" declarations
which seems like a good way to solve the Expression Problem (tm)
TimToady sort of a require... 04:31
audreyt (not runtime infix does, compile-time separate decl does) 04:32
TimToady I wonder how much "does" implies "use"... 04:34
wolverian audreyt, that is insanely cool. now, to port it to perl6.. 04:35
(oleg's diff hack)
cmarcelo audreyt: expression problem => how to add new "cases" for a type and add new functions "on" that type without "trouble" (recompiling, changing lots of code, doing lots of boilerplate)? or is a deeper thing?
audreyt cmarcelo: exactly that 04:36
04:41 zgh left
cmarcelo audreyt: mind checking if you get "no instance error" changing "_BufToID" type to "Map ByteString ID" instead of "H.StrMap ByteString ID"...? 04:43
04:46 Khisanth joined
audreyt sure. sec 04:48
surely you mean IORef?
IORef (Map ByteString ID)
ioref works here 04:49
cmarcelo ops. my mistake
svnbot6 r12538 | audreyt++ | * Pugs.Internals: Even more encode/decode `seq`ing 04:50
wolverian oh, what does the | do in: class Term t a | t -> a where...? I didn't find it in the specification 04:59
audreyt it's fundep 05:02
it's so much fun and addictive, you'll grow dependent on it
wolverian yes, but what does it mean? :-)
oh, should I look at the extensions? 05:03
found it. thanks :)
...and it links to a paper. :D 05:04
audreyt see ghc userguide 7.4.3. Functional dependencies
roughly: multiparam typeclasses encode relationships 05:05
wolverian ah. thanks
audreyt some relationships can be fully determined only by some part of it
wolverian I went straight to 7.4.7 .. reading the paper anyway
audreyt in haskell-prime, the committee is considering fast-tracking associated types to replace fundep, though 05:06
since fundep force you to write prolog in haskell :)
wolverian hm.. the core doesn't use fundeps I gather? 05:10
(of ghc)
05:10 agentzh joined
TreyHarris audreyt: how can I rebuild Prelude.pm.yml? i don't need to know anymore since i'm making pugs again, but for next time i wondered 05:10
agentzh TreyHarris: i was thinking about the "pugsbug" issue more seriously during my lunch and walk.
TreyHarris ok? 05:11
agentzh TreyHarris: perl util/gen_prelude.pl -v -i src/perl6/Prelude.pm -p .\pugs.exe --output blib6/lib/Prelude.pm.yml
audreyt TreyHarris: rm blib6/lib/*yml ; make
TreyHarris audreyt++ # r12537 fixes issue 05:12
agentzh hmm, i think i should add the yaml refresh commands into prove6.
TreyHarris ahhh... i left out the -p... part
agentzh because i'm also using this command constantly.
TreyHarris audreyt: that didn't work :-(
agentzh TreyHarris: now i totally second your opinion. we should have a pugsbug directory or something like that. 05:13
this directory holds separate files for special failing tests. 05:14
the tests should be as simple as possible. so that audreyt++ can fix the bugs more easily. 05:15
and more effectively.
TreyHarris yes. the only question i have is does it just accumulate tests forever? or do we create a subdir for tests that now succeed
agentzh but this directory shouldn't be under t/ since these tests are not supposed to run by ordinary test harness or smoke engine.
it shouldn't accumulate tests. 05:16
TreyHarris agentzh: eh?
agentzh most of the tests in this directory should also be put into corresponding test files in the main test suite.
TreyHarris old bugs come back to haunt you, you shouldn't delete them once fixed.... 05:17
agentzh if they're segfaulting, then comment out the copies in the main test suite.
TreyHarris ah, i see.
we might want to have a hardfail directory then
agentzh TreyHarris: no problem, once the old bugs come back, we can detect them from the main test suite. 05:18
TreyHarris with stuff that *isn't* in the language, that shouldn't work, but shouldn't cause the implementation to blow chunks
agentzh and this directory shouldn't be included in the pugs release, i think.
so no TODO trick is needed there.
i'll be very willing to add new tests to such a directory because it's more likely to raise audreyt et al's attention. 05:19
there shouldn't be too many noises in this directory and the test files should be kept minimal, or we'll lose the point. 05:21
05:21 kanru joined
agentzh TreyHarris: how about creating a "bug" directory at the same level of t/ ? 05:23
audreyt er no
xx-uncateg was known as t/bug/
agentzh audreyt: but i think the tests in t/bug should be skipped by the test harness. 05:24
audreyt but why?
agentzh segfaulting is smoke's killer.
(assume there's no timeout mechanism) 05:25
also pugs' bugs are not necessarily to be other implementations' bugs.
anyway these are solely my personal opinions. criticism welcome :) 05:27
agentzh goes back to util/smartlinks.pl. 05:28
svnbot6 r12539 | trey++ | With segfault fixed by audreyt++, 05:32
r12539 | trey++ | moving hash tests to t/data_types/hash.t
05:37 justatheory joined
svnbot6 r12540 | agentz++ | [util/prove6] 05:45
r12540 | agentz++ | - added the --update-yml option. so now you can use the
r12540 | agentz++ | following command to refresh Test.pm.yml and
r12540 | agentz++ | Prelude.pm.yml:
r12540 | agentz++ | util/prove6 --update-yml
agentzh loves JSAN's css and decides to steal that for smartlinks.pl. 05:48
cmarcelo english speakers: is it "Pugs' trunk" or "Pugs's trunk"? or both? 05:52
gaal agentzh++ but please consider factoring out the actual generation to the Makefiile
for Test.pm there's already a target 05:53
agentzh audreyt: what's the simplest way to upload HTML files to feather? via SVN? (i always use SOCKS4 proxy to connect to feather)
gaal for Prelude.pm.yml you'll need to add an option to build_pugs and run that
agentzh gaal: (?)
gaal note that build_pugs possibly constructs a larger Prelude than just src/perl6/Prelude.pm
agentzh gaal: would you please fix util/prove6 for me? 05:54
gaal agentzh: scp file feather.perl6.nl:
agentzh gaal: what is scp?
gaal agentzh: I will when I get cycles but that won't be for 10 hr 05:55
cmarcelo: I think either works.
agentzh: it's the ssh version of rcp
oh you're on windows
agentzh ah, thanks!
gaal @google pscp
lambdabot www.chiark.greenend.org.uk/~sgtatha...nload.html
Title: PuTTY Download Page
agentzh pscp is already in my perl installation. yay! 05:57
gaal you probably need to set the username:
pscp file [email@hidden.address] 05:58
agentzh thanks!
gaal also look in to "Paegant" on the same page, which allows you to do transfers with no need for manual password typing
TreyHarris cmarcelo: the usual rule (though it isn't 100%) is that if when sounded out, it ends in a zzz sound, you do s', but if it ends in a sss sound, you do s's. Pugs' sounds right to me
agentzh k 05:59
cmarcelo i'll stick with "Pugs' " then =)
05:59 Khisanth joined
agentzh gaal: when runing the command "pscp -load Feather [email@hidden.address] S02.html", i got the following error: 06:04
06:04 ludan joined
agentzh ssh_init: Host does not exist 06:04
ludan hi 06:05
agentzh but i *can* login to feather via putty.
ludan: hi
ludan i'm going to make a coffee, see you later
agentzh looks into psftp 06:07
svnbot6 r12541 | Sage++ | A first attempt at translating open(...). Not nearly complete, but at least some cases are handled. 06:10
TreyHarris agentzh: that syntax doesn't look right to me, though it's been awhile since i've used pscp. i don't think you need the "-load Feather" bit
agentzh TreyHarris: i'm behind a firewall. i need the proxy settings.
cmarcelo audreyt: pugs.blogs.com/pugs/2006/08/pugs_meets_judy.html 06:11
SageLT Whew, I think I'm through for the night. Guess that's it for my Summer of Code. 06:12
TreyHarris agentzh: ah. did you do the set PATH=C:\putty\directory;%PATH% bit?
agentzh TreyHarris: putty, pscp, and psftp are all in my perl's bin directory. :) 06:13
obviously the -load option doesn't work for me. :( 06:16
06:19 Gruber is now known as Grrrr
cmarcelo audreyt: I asked english-enabled friends to review my original post, but fix it if something is wrong.. 06:19
sleep &
06:19 cmarcelo left
agentzh gives up pscp and psftp, returning to svn-based file transfer. 06:22
i've checked in the HTML files generated by util/smartlinks.pl to my svn repos on berlios: 06:37
svn.berlios.de/svnroot/repos/unisimu/Perl/Syn/
lambdabot Title: Revision 663: /Perl/Syn 06:38
agentzh however, i think it's not appropriate to check them into the pugs repos. :) 06:39
06:39 traecer left 06:40 Aankhen`` joined
agentzh wow, has any kind person silently fixed "dev.perl.org/css/perl.css" for me? 06:46
the code snippets in the HTML files dramatically get bounded. :D 06:47
wolverian bounded? 06:48
agentzh wolverian: yes. 06:52
wolverian: look at svn.berlios.de/svnroot/repos/unisim...n/S02.html 06:54
lambdabot Title: S02
svnbot6 r12542 | agentz++ | [util/smartlinks.pl]
r12542 | agentz++ | - fixed command line options with arguments.
r12542 | agentz++ | - formated the code snippets by stripping extra
r12542 | agentz++ | indentation.
agentzh oh, the HTML looks much nicer now. yay! 06:55
wolverian oh. nice. I was looking at dev.perl.org
agentzh now we have code snippets from the pugs test suite embedded directly into the synopses! hooray! 07:00
gaal agentzh: try w/o a saved session? btw are you pushing to feather or d/ling from it?
agentzh gaal: that didn't work either. 07:01
gaal because that command line is for pulling from
agentzh gaal: it was the connection timeout error. 07:02
gaal: how about pusing files onto feather?
*push
gaal agentzh: pscp myfile [email@hidden.address]
if you want it on the web area, then pscp myfile [email@hidden.address] 07:03
agentzh gaal: does this command take my proxy settings into account?
gaal dunno, let's look at the docs...
hold on, which proxy? this is over ssh 07:04
agentzh SOCKS4
in fact, i wouldn't be able to connect to feather if there's no proxy. ;-) 07:05
gaal: you fixed the css on perl.org? 07:06
gaal how does your putty connection work?
agentzh: not I
agentzh it's lovely. :)
gaal: putty has a GUI.
gaal agentzh: I'm aware of that :)
agentzh i set the proxy in its GUI.
JIT fixing++ 07:07
gaal that's a proxy it creates, not uses, no?
agentzh i'm using a proxy offered by a company.
gaal agentzh: I'm looking at the berlios html, is that the one you liked? it looks like yersterday's html to me... 07:08
agentzh really?
agentzh is surprised. 07:09
gaal: there's no border around the code snippets?
i can see the border here in both IE and Firefox. 07:10
gaal yes, there is a border
apart from that?
agentzh so that's it.
gaal ah
okay :)
I thought I added the border in smartlinks.pl...
agentzh gaal++ 07:11
gaal but that's not my change. it says 07:12
style="border: 1px solid ; display: block;"
I said
agentzh oh, yes, i see that.
gaal style="display:block; border: 1px solid" :-)
agentzh i didn't notice the change. :)
gaal so someone else must have dunnit
(someone else)++, heh 07:13
agentzh looking into the log now...
gaal anymoose I should go now
have a nice day :)
agentzh thanks!
07:25 kane-xs joined
agentzh believes the Perl 6 test suite should serve as a good extension to the Perl 6 Spec. 07:31
07:33 masak joined 07:36 iblechbot joined
audreyt agentzh: want to write about this a bit to perl6-language and/or pugs.blogs.com? (just sent you a blog bit) 07:37
TreyHarris agentzh: I can't figure out why the snippet of syntax/force_context.t appears immmediately above "return 0 but True" in S02, can you explain it to me? 07:38
agentzh audreyt: already sent a mail to p6l and p6c.
TreyHarris the smartlink to me doesn't specify that place
agentzh TreyHarris: looking
audreyt: thanks for the blog bit. :) 07:39
TreyHarris: yes, that's a bug in smartlinks.pl. 07:40
thanks for the catch. ;-)
there's also a bug related to "empty" snippet.
i'll try to fix them in the next 2 hours or so. :) 07:41
TreyHarris heh. right, i was about to say that. "empty" meaning consisting of just the test, right?
07:43 ilogger2 joined, gantrixx joined, Bit-Man joined, rafl joined 07:44 takanori joined, rashakil joined, miyagawa_ joined 07:46 lollan joined, nperez joined, dakkar joined, lumi joined 07:47 pmurias joined, broquaint joined, JarJarBinks joined, hcchien joined, xern joined, rgs joined, kalven joined, revdiablo joined, dvorak joined, cookys joined, notsri joined, LCamel joined, Frances joined, obra joined, Yappo joined, c6rbon joined, clkao joined 07:48 pmurias_ joined, cognominal joined, f0rth joined 07:50 marcus_ joined 07:52 Maddingue joined 07:53 tcliou joined, Grrrr joined, Southen joined 07:55 Psyche^ joined, Psyche^ is now known as Patterner 07:59 Juerd joined 08:02 masak joined, Eidolos joined, shachaf joined 08:03 ajs_home joined, glasser joined, awwaiid joined 08:04 buu joined, xerox joined 08:10 ruoso joined
pasteling "agentzh" at 210.22.200.67 pasted "I wonder why i can't get into #perl6 :(" (1 line, 8B) at sial.org/pbot/19176 08:11
08:12 agentzh joined
agentzh oh, finally back... 08:12
audreyt was netsplit
agentzh TreyHarris: no, being "empty" means the smartlink is pointing to the same code snippet of the next smartlink. 08:13
08:13 baest joined
agentzh in other words, adjacent smartlinks share the same code snippet right below them. :) 08:14
i think i should have added this to t/README... 08:15
adding now...
Juerd: can i keep a small perl script running on feather so that it can awake every 6 hours and update the HTML version of synopses automatically using smartlinks.pl? 08:16
(the sentence above is a bit long...) 08:17
audreyt yes you can . :)
agentzh thanks!
08:19 evalbot_12542 joined 08:21 nekokak joined
agentzh audreyt: are you using the pugs working copy actively on feather? can i point the program to read the test suite under your working copy? 08:25
audreyt sure
~audreyt/pugs/ is constantly updated
(by evalbot)
agentzh oh, it's very nice!
audreyt :)) 08:26
agentzh oh, svnbot is no longer here. :( 08:36
gaal: (someone else)++ === (gaal)++ 08:38
lol
08:40 svnbot6 joined
audreyt the bot should be made netsplit-safe 08:40
agentzh encounters the concept of netsplit for the first time. 08:49
svnbot6 r12544 | audreyt++ | * Remove third-party/TextRegexLazy now we have scw++'s alternate
r12544 | audreyt++ | plan based on parsec combinators.
agentzh after fixing this bug in smartlinks.pl, luqui++'s point that we can only copy closures at the same source code position is now much clearer to me. 08:52
svnbot6 r12545 | agentz++ | [util/smartlinks.pl]
r12545 | agentz++ | - fixed the "empty" snippet bug. multiple adjacent smartlinks should share the same snippet of tests right below them.
clkao fglock: v6 is generating something stupid like " return ( unless ( [ \( $Test::testing_started, ) ], {} ) );" 08:58
svnbot6 r12546 | agentz++ | [util/smartlinks.pl]
r12546 | agentz++ | - warn a message to stderr if empty snippets are detected at runtime (so as to make sure the previous bug is truly fixed)
agentzh supper & 09:03
09:03 agentzh left 09:07 elmex joined 09:08 drrho joined 09:15 pmurias_ left 09:17 tup joined 09:35 chris2 joined 09:39 buetow joined 09:56 pmurias_ joined
pmurias_ how do i flush stdin in p6? 09:57
for =<> -> { say "?" } is not lined buffered on feather 09:58
audreyt hmm. might be accidentally disabled by the decode strict fix 10:06
a sec
yeah. fixing 10:12
fixed (I think.) testing 10:15
try again? 10:16
svnbot6 r12547 | audreyt++ | * pmurias++ pointed out that =<> and =<file> and =$fh
r12547 | audreyt++ | was accidentally made strict.
10:31 rindolf joined 10:35 borisz joined 10:36 borisz left
pmurias_ audreyt: i use the feather build so i can't try it yet 10:43
audreyt there's the evalbot build 10:46
~audreyt/pugs/pugs
that worksforme
pmurias_: what is this feather-build you refer to? 10:47
pmurias_ /usr/bin/pugs - isn't it the stable realase... :) 10:50
Juerd That lags far behind. rafl doesn't maintain this installation anymore, I think.
audreyt er no
that is 6.2.10
it's built about 10 months ago 10:51
I wouldn't mind ln'ing it to ~audreyt/pugs/pugs
Juerd audreyt: Great. I'll apt-get remove pugs then. Will you do the rest?
audreyt sure! 10:52
pmurias_ undecleared $module
Juerd Super
Done
That is - doing.
Done.
audreyt done. 10:54
pmurias_: try again?
pmurias_ works
audreyt yay 10:55
pmurias_ audreyt: does your version use PGE? 11:01
audreyt try again 11:03
I populated /usr/bin/parrot too
pmurias_ it uses PGE 11:06
audreyt good 11:07
pmurias_ could you look at ~/pmurias/ed.pl 11:09
it prints "print" on output not containing numbers,why?
audreyt drop the -> 11:10
$_ is not yet lexical
so it's seeing bad things
for @x -> {...} # this means to ignore values in @x 11:11
for @x {...} # this means to put values in @x to $_
Juerd What's hindering lexicalization of $_?
audreyt previously, it's all the $CALLER::_ defaults that was hindering it
now, there's nothing hindering it, so it should be done as part of Code refactoring that luqui/gaal/audreyt/etc are pushing 11:12
audreyt is glad that $CALLER::_ defaults are gone.
Juerd Sounds positive :)
audreyt sure :)
Juerd has to go to $office, but doesn't want to.
I've led a group of 15 kids (ages 13..16) for a week, and don't feel like going back to work :) 11:13
audreyt maybe you found your calling :)
Juerd I found it years ago, but now for the first time had the opportunity to actually do it :) 11:14
This is with an organization where I've been one of the kids for 10 years :)
They organize summer computer camps.
In three age groups, 9..13; 13..16; 16..20
audreyt pmurias_: also for some reason it needs to be written 11:15
if $_ ~~ m/\d/
investigating
pmurias_ it doesnt work 11:20
for me
audreyt you mean even with dropping -> and saying (if $_ ~~ m/\d/) ? 11:21
for =<> { if $_ ~~ /^\d/ { say "print" } } 11:22
this works for me
pmurias_ i tire if m/
*tried 11:23
audreyt that is a bug. fixed
rindolf What is the "~~" operator ?
audreyt search.cpan.org/dist/perl/pod/perls...ments_____ 11:24
lambdabot Title: NAME syntax - search.cpan.org
pmurias_ the generalisation of =~
rindolf pmurias_: I see.
pmurias_: is =~ still available in p6?
lumi Yes, it's spelt ~~ 11:25
pmurias_ is there an equivalent of Perl::Tidy for p6?
audreyt no, not yet. want to write one? :) you can probably hook a Emit::Perl6 to v6.pm 11:26
pmurias_ i abbandoned my plans for an yhc backend and i'm working on an ed(1) in p6 11:27
and indenting code in ed doesnt work great...
audreyt nice!
svnbot6 r12548 | audreyt++ | * In boolean context, m// and // should match against $_.
r12548 | audreyt++ | Reported by pmurias++
audreyt pmurias_: there you go :) 11:28
(/usr/bin/pugs on feather will be refreshed every 10 min or so)
Juerd [fast compiling pugs]++ 11:30
svnbot6 r12549 | agentz++ | [t/README]
r12549 | agentz++ | - added discussion of multiple adjacent smartlinks.
r12549 | agentz++ | - added the command to generate HTML using smartlinks.pl.
r12548 | audreyt++ | * In boolean context, m// and // should match against $_.
r12548 | audreyt++ | Reported by pmurias++
audreyt ?eval $_ = 10; if /10/ { say 'hi' } 11:32
11:32 evalbot_12542 is now known as evalbot_12548
evalbot_12548 OUTPUT[hi ] Bool::True 11:32
audreyt ?eval $_ = 10; if /11/ { say 'hi' }
evalbot_12548 undef
audreyt cool
xerox audreyt: watch this: 11:38
[1 of 1] Compiling ByteStringParser ( ByteStringParser.hs, interpreted )
Ok, modules loaded: ByteStringParser.
*ByteStringParser>
audreyt url? 11:41
xerox I can mail it to you, it doesn't do much now anyway 11:42
nothingmuch wonders whether Verby should be broken up into many dists 11:43
11:48 cookys joined
audreyt xerox: is this based on parsec? 11:53
xerox No
It's a rewrite from scratch
*ByteStringParser> parseTest (pMany (pChar '1')) (Data.ByteString.Char8.pack "1111")
"1111"
audreyt oh. I did that once...
with PArrows
xerox That is stepcut's work
audreyt nod 11:54
xerox Interesting
11:54 bpphillips joined
audreyt src/Text/Parser/PArrow.hs is my work, based on musasabi's 11:54
mail me a copy anyway?
xerox Okay
audreyt have you talked to stepcut about licensing?
xerox Nope 11:55
It says:
-- Module : ByteStringParser
-- Copyright : (c) Jeremy Shaw 2006
-- Copyright : (c) Daan Leijen 1999-2001
-- License : BSD-style (see the file libraries/parsec/LICENSE)
audreyt good enough
xerox Where do you want it mailed? 11:58
svnbot6 r12550 | agentz++ | [util/smartlinks.pl] 12:04
r12550 | agentz++ | - fixed the duplicate "syntax/force_context.t"
r12550 | agentz++ | snippets bug reported by TreyHarris++.
r12550 | agentz++ | (it was matching against code snippets as well as
r12550 | agentz++ | true POD paragraphs.)
audreyt xerox: audreyt at audreyt dot org 12:11
rindolf Hi audreyt
xerox Sent
audreyt danke 12:12
12:14 weinig joined
audreyt cool minimal design 12:18
xerox Yes!
audreyt but I like Parsecs error reporting 12:19
Unexpect / Expect
audreyt ponders hacking that in 12:20
nah. MOP first :)
12:20 Limbic_Region joined
xerox Okay :) 12:21
audreyt but thanks :) my $job can use this :)
lumi Can it be that make doesn't drift but make fast does? 12:22
xerox Uhu
audreyt lumi: possibly
Limbic_Region audreyt - did you happen to work on the Win32 opendir/readdir/etc over the weekend?
audreyt er wait
no
Limbic_Region: sorry, I didn't realize it was broken 12:23
so no
was it broken?
(I didn't add error reporting)
xerox audreyt: send back patches, if you happen to have sendable ones :D
Limbic_Region yes - you must have been really tired when you went to sleep at 4am on Friday
Limbic_Region will check this morning as I am building fresh
audreyt that I was...
so, remind me: how is it broken? which test should I look at? 12:24
Limbic_Region it was actually my File::Find code for testing examples better
Limbic_Region will re-nopaste in a sec
want to tell you the only revision it actually worked in
audreyt oh that 12:25
no I didn't look into that
but I wonder if the bug was in opendir or not
Limbic_Region 12424
your thinking on Friday was that it was getting an undef dir or something
audreyt if you svn up now 12:26
Limbic_Region and I was going to sprinkle debug prints throughout File::Find
audreyt (no rebuild needed)
does t/builtins/io/dir.t
pass?
Limbic_Region audreyt - already building fresh from most recent checkout
audreyt okie. I just commented out the win32 line on dir.t
Limbic_Region does so after weekends regardless because Win32 seems to require realcleans far more than any other OS
svnbot6 r12551 | audreyt++ | * opendir tests - uncomment in win32
Limbic_Region it just saves me time 12:27
audreyt I understand
as a consolation, win32's GHC in optimized mode is much faster than on unix.
Limbic_Region cool beaners - moldy chicken
audreyt did you say you were currently building?
pasteling "Limbic_Region" at 129.33.119.12 pasted "File::Find code that was (and possibly still is) bombing on Win32" (14 lines, 295B) at sial.org/pbot/19179 12:28
Limbic_Region yes
but it is early enough to restart if there is something you wanted to try 12:29
audreyt ok. give me 2min
Limbic_Region so build 12424 was the only build that actually worked once you started using ghc's code for the dir io stuff
after that - it just kept silently exiting
but then again, I didn't touch it all weekend myself 12:30
audreyt I didn't touch that codeline after r12424 either 12:34
svnbot6 r12552 | audreyt++ | * build_pugs.pl: "make profiled" will also rebuild third-party
r12552 | audreyt++ | as profiled if needed.
Limbic_Region over 100 commits over the weekend - yippee 12:35
audreyt please build r12254 12:36
svnbot6 r12553 | audreyt++ | * Remove the lazy UTF8.lhs and fold encode/decodeUTF8
r12553 | audreyt++ | back into Pugs.Internals as strict inlinable functions,
r12553 | audreyt++ | which netted us lots of space win.
r12554 | audreyt++ | * Pugs.Parser.Operator: Let-float the currently-static levels
r12554 | audreyt++ | of the operator table for both speed win (10%) and clarity.
audreyt pugs should be noticeably faster over the weekend
Limbic_Region well, it was already noticeably faster on Friday afternoon 12:37
audreyt smoke on my machine drops to <8min
on Friday it was still 12min or 12min
yeah, that is was
(bbiab) 12:38
12:42 explorer joined 12:47 weinig joined
xerox Why isn't the Haskell community hacking ghc at large as perl6 is on his compiler? :( 12:48
audreyt the ghc hackathon will change that :) 12:50
(we talked about this in ICFP'05, where I foolishly jumped in and hacked GHC without knowing anything about it)
xerox Yarrr
Another reason to be there...
Sigh.
audreyt and identified (darcs, trac, hackathon) as the neccessary ingredients
12:52 explorer joined, jferrero joined 12:53 Qiang joined
Limbic_Region audreyt - 12552-12554 aren't necessary for me to test the Win32 dir io stuff right? 12:55
. o O ( he says being very very paranoid ) 12:56
audreyt no
just uncomment the win32 line in dir.t
t/builtins/io/dir.t
and run it, and see if it works
Limbic_Region ok, just made it to the heap exhaustion point - the second nmake doesn't take nearly as long to finish 12:57
audreyt actually if you are feeling adventurous 12:58
I wonder if heap exhaustion is gone in GHC 6.5-win32
which can be installed from www.haskell.org/ghc/dist/current/di...w32.tar.gz
but test that only if you have spare cycles :) 12:59
13:07 lollan joined
audreyt xerox: looks like the upcoming GHC hackathon may have 30+ people 13:07
that's comparable to the largest perl6 hackathon yet
xerox Really? Wow. 13:08
audreyt bottom of hackage.haskell.org/trac/ghc/wiki/Hackathon
lambdabot Title: Hackathon - GHC - Trac
xerox On another front, I was thinking about mailing all the soc people and find out what's going on now that is finished
Maybe have them write two lines on the wiki
svnbot6 r12555 | fglock++ | v6 - optimize Parse::Yapp output; gets another 10% speed 13:10
Limbic_Region audreyt - all 18 passed
yippee
now to test my code
ok, my code still no workie - and no error messages in output 13:11
so sprinkle lots of debugging print statements throughout File::Find?
13:12 mauke_ joined
audreyt yeah 13:13
I think the bug is in the rel2abs call
13:13 cjeris joined
Limbic_Region Ok - this is odd 13:14
I disagree 13:15
13:15 chaoslawful joined, fglock joined
Limbic_Region putting a say "Entering sub" as the first line of the find method doesn't even print 13:15
13:15 lisppaste3 joined
Limbic_Region is going to wipe out the entire body of the sub and see what happens 13:15
audreyt you can also use pugs -d 13:16
to see what's really happening
pugs -d yourprog > log
Limbic_Region will do - but first
changing the method sub to the following body { say "entering sub"; return 1 }
results in no output either
13:17 mauke_ is now known as mauke
Limbic_Region oh wait, I think I know what is going on here 13:17
duh 13:18
can't modify the File::Find in ext/ and expect it to affect the one in blib6/lib 13:19
it must be Monday
fglock can I get an older revision using plain svn?
audreyt svn up -r 12380 13:20
fglock thanks 13:21
audreyt np :)
Limbic_Region audreyt - found the first anomoly - @start = @dirs || @.dirs; is empty though in my script $f.dirs = ('examples'); 13:22
audreyt yes that's a bug 13:23
try cgiving it an array
$f.dirs = ['examples']
Limbic_Region ok
same result - nada 13:24
is there a way to dump a code ref to see the actual code? 13:25
audreyt ?eval class C { has @.dirs; method x { say @.dirs } }; my C $c .= new; $c.dirs = ['x']; $c.x 13:26
13:26 evalbot_12548 is now known as evalbot_12555
evalbot_12555 pugs: Enum.toEnum{Word32}: tag (-1229482828) is outside of bounds (0,4294967295) 13:26
audreyt ?eval class C { has @.dirs; method x { say @.dirs } }; my C $c .= new; $c.dirs = ['x']; $c.x
evalbot_12555 pugs: Enum.toEnum{Word32}: tag (-1227385676) is outside of bounds (0,4294967295)
audreyt wt...
that works locally.
actually the other way works too
can you try
@start = (@dirs ?? @dirs !! @.dirs) ? 13:27
lambdabot Maybe you meant: part state
13:27 baest joined
Limbic_Region sure 13:27
nada 13:29
IOW - I am fairly certain @dirs and @.dirs is empty
audreyt my $f = File::Find.new( wanted_file => ..., dirs => ['examples'] ); 13:30
does that work?
audreyt ponders what's wrong...
Limbic_Region tries 13:31
took much longer to run, but still no output 13:32
svnbot6 r12556 | audreyt++ | * Allow pretty-printing subroutine body so that {...}.perl 13:33
r12556 | audreyt++ | would print something. It's not perl, though -- takers
r12556 | audreyt++ | welcome to implement a real Pretty instance for Exp that
r12556 | audreyt++ | dumps back to perl6.
Limbic_Region remember - this worked in 12424 which is really odd to me
audreyt - ok, just discovered that things are working - sorta 13:36
pmurias_ how do i specify the http proxy in svk? 13:37
audreyt I got it. 13:38
pmurias_: in ~/.subversion/servers
pmurias_ thanks
Limbic_Region audreyt - I put a say statement in the wanted_file code ref - it is executing 13:39
repeatedly I might add
audreyt the bug was in that named list params 13:42
when there's no default
assumes a default =undef
so @dirs is actually true
and always shadows @.dir
?eval sub { 1234 }.perl 13:43
13:43 evalbot_12555 is now known as evalbot_12556
evalbot_12556 pugs: Enum.toEnum{Word32}: tag (-1227381580) is outside of bounds (0,4294967295) 13:43
audreyt ?eval 1+1
evalbot_12556 pugs: Enum.toEnum{Word32}: tag (-1229478732) is outside of bounds (0,4294967295)
Limbic_Region evalbot is fubar
so this is really strange 13:44
since @start is empty - it is starting at pwd 13:45
it is hitting every single file in the pugs build dir
and it isn't pushing any of them onto $.results because
well, I am not sure why nothing is getting pushed onto $.results 13:46
return 1 if $file ~~ m:P5/^.*\.pl$/; # should return 1 for any file ending in .pl correct? 13:48
ayrnieu return 1 if $file ~~ m:p5/\.pl$/; 13:49
audreyt the defaulting prolem is fixed 13:52
Limbic_Region ayrnieu - it's not my regex, it is in the synopsis for File::Find - my question was that it should work
and in fact it does work, and $.results.push($abs) is happening - but for some reason the code is not finishing to produce the results 13:53
13:54 thepler joined
Limbic_Region audreyt - do you suspect this will sove the problem? 13:54
audreyt let me commit the :@dir first 13:55
no, it's orthogonal, but at least it will let you get 'examples'
Limbic_Region I don't for the following reason - in method find, I hardcoded @start to examples and it still is not working
ok
so this is just so we can find the next whacky thing going on
audreyt fwiw, it worked for me. 13:56
i.e. your code verbatim now runs here
svnbot6 r12557 | audreyt++ | * Fix defaulting of named :@x so it gets bound to (), not (undef),
r12557 | audreyt++ | when there's no incoming argument 'x'.
Limbic_Region well, thats a good thing
audreyt and the content of @file is correct
Limbic_Region is $f.dirs = qw/examples/; ok or should it still be ['examples']? 13:59
audreyt cmarcelo: r12701 made feather segfault :-/
cmarcelo: offending instance:
instance Refeable Int where toRef i = return $ toEnum i
i.e. negative ints will be rejected 14:01
that's kinda wrong :/
we want to map negative ints to large unsigned word instead
pmurias_ can rules take params like: 14:02
rule greet($who) {Hello $who,?} 14:03
i skimed through the rule part of the synopsis but i can't find anything 14:04
svnbot6 r12558 | audreyt++ | * Provide hand-crafted instance to make toEnum happier 14:05
r12558 | audreyt++ | so we can continue using IntMap.
SamB re-feable?
what does that mean?
audreyt referable, persumable 14:06
referable, persumably
pmurias_: yes, and pmichaud uses this calling syntax: /<greet: foo>/ 14:07
but I think he and fglock eventually settled to /<greet('foo')> 14:08
and both are canonicalized to S05 14:09
line 642 I think.
pmurias_ i was vagly aware of the calling syntax, 14:11
i wasn't sure how you define them
pmichaud codes some of his rules in pir 14:12
14:13 hexmode joined
audreyt S05:642 explains it succintly 14:16
svnbot6 r12559 | audreyt++ | * build_pugs.pl: oops, typo.
r12560 | audreyt++ | * Switch back to HashTable in Pugs.Parser.Operator for now
r12560 | audreyt++ | as the ReversibleHashIO instance's use of toEnum seems
r12560 | audreyt++ | still to cause panic on feather -- no time to debug tonight.
pmurias_ audreyt: i use printed out synopses,i'll check the online copy 14:17
y
audreyt ok. grep for =head1 Extensible metasyntax (C<< <...> >>) 14:18
at about 20% into the text
Limbic_Region audreyt - I am almost finished building 12558 14:20
if this no workie then I will try and debug on my own to give you something specific to look at
audreyt ok
locally it works
so if fails for you, might be a win32 thing
I'm using unmodified Find.mp
14:21 neoesque joined
Limbic_Region ok, step 1 is good @start contains examples 14:22
14:22 evalbot_12560 joined
Limbic_Region and so the program is dying somewhere in the while $dir = $.queue.shift loop since a say at the end of that loop is never printed out and thus $.results is never returned 14:23
so I try and determine what dir and/or file it is bombing on
svnbot6 r12561 | masak++ | collected another reference. I'm done backlogging August now... July, here I come 14:25
audreyt ?eval 1+1 14:26
evalbot_12560 2
Limbic_Region yeah, evalbot is fixed
14:27 vel joined
Limbic_Region s/yeah/yay/ 14:27
audreyt sorry I need to fade now :) $job tomorrow morning 14:32
*wave*
Limbic_Region audreyt - good new
news even
telling it to ignore .svn directories caused it to work
audreyt yay
Limbic_Region moves on to finishing the rest
audreyt Limbic_Region++ 14:33
I don't have a .svn file here, that might be the diff :)
Limbic_Region .svn should be ignored anyway
audreyt yup
*wave* &
Limbic_Region thanks and sleep well
14:53 prefiks joined
svnbot6 r12562 | fglock++ | v6 - some fixes, works with current Test.pm 14:56
14:58 [particle] joined
fglock new v6.pm times: 1.1s for one-liners, 5.5s for Test.pm 14:59
14:59 alinbsp joined 15:00 pbuetow joined
svnbot6 r12563 | fglock++ | v6 - fix warnings running Test.pm 15:14
15:15 fglock left
Limbic_Region anyone running Win32 able to test something for me? 15:22
cjeris Limbic_Region: ? 15:25
15:26 dbrock joined
Limbic_Region cjeris - are you up to the latest rev? 15:26
cjeris just built r12560 15:28
Limbic_Region close enough
pasteling "Limbic_Region" at 129.33.119.12 pasted "put this in foo.pl in the pugs build root dir and run it as pugs.exe -Iblib6/lib foo.pl" (21 lines, 378B) at sial.org/pbot/19183
Limbic_Region then nopaste the results to me 15:29
pasteling "cjeris" at 140.247.124.140 pasted "result of Limbic_Region's run" (2 lines, 48B) at sial.org/pbot/19184 15:30
cjeris sorry, it's r12561 not 12560 15:31
Limbic_Region cjeris - now change the $f.debug = 0; line to $f.debug = 1; tell me if the last line of output is "bye\n";
15:31 flips_and_rails joined
Limbic_Region no worries - it is recent enough to have included all the lastest audreyt commits 15:31
15:31 amnesiac joined
cjeris watches the script enumerate the examples directory, slowly :D 15:31
Limbic_Region this is truly weird 15:32
svnbot6 r12564 | audreyt++ | * Pugs.Internals: Revive the ability for Var/ID to be CString
r12564 | audreyt++ | (critical for embedded nulls in face of Judy.StrMap), by
r12564 | audreyt++ | encoding \0 internally as \xC0\x80.
r12564 | audreyt++ | -- In the \0 case, we diverge from the Unicode standard to remove any trace
r12564 | audreyt++ | -- of embedded nulls in our bytestrings, to allow the use of Judy.StrMap
r12564 | audreyt++ | -- and to make passing CString around easier. See Java for the same treatment:
r12564 | audreyt++ | -- java.sun.com/j2se/1.5.0/docs/api/ja...fied-utf-8
r12564 | audreyt++ | Note that we should not do that across I/O boundaries...
r12564 | audreyt++ | So a more elaborate treatment in ByteString is needed anyway.
Limbic_Region you are getting the same results as me
cjeris ok, so there is a long list of +d/-d, etc lines, then a BIG line which I guess is the out of say ~@files; then two lines 'bye' '1..0' . 15:33
Limbic_Region cjeris - I am going to nopaste something and hopefully you can make sense of it 15:34
pasteling "Limbic_Region" at 129.33.119.12 pasted "See blib6\lib\File\Find.pm and comments below" (61 lines, 1.7K) at sial.org/pbot/19185 15:37
Limbic_Region basically, the inclusion of say "something"; gets it to work and produce the correct result if placed in 4 locations but the absense of which exits abnormally with no error
IMO - that's weird
cjeris so, when you put the 'say's in only some of the places, does it give you only some of the results? 15:38
or does it die the first time it hits a branch with missing 'say'?
Limbic_Region cjeris - it is hard to say since adding debug lines affects the behavior 15:40
though I think I have found a problem (not necessarily the problem)
cjeris Limbic_Region: this is well beyond the point at which I can have bright ideas :) 15:41
Limbic_Region ok - thanks 15:42
Limbic_Region will continue to try and isolate but this is likely above his head too (meaning internals)
cjeris np 15:43
15:44 justatheory joined
Limbic_Region ok - guess I will take this up with audreyt in the morning but it is almost as if - it requires IO in order to continue looping 15:45
audreyt all four say is needed 15:46
or any one will do?
(it works without say here, so I'm relying on your info)
Limbic_Region audreyt - all 4 are needed if you want it to finish 15:47
audreyt cjeris: are you on win32 too?
Limbic_Region audreyt - yes, cjeris is on Win32
basically - any loop that doesn't result in IO exits prematurely with no error output
so with 3 out of the 4 it goes further than with 1 out of the 4 15:48
so earlier when I thought excluding the .svn directories fixed the problem it really was because I had debugging turned on - but you should be sleeping
audreyt change say() to Thread.yield() 15:49
and see if that still works
Limbic_Region ok
I found a parse bug if you are interested 15:50
audreyt it looks like my Pugs.Compat code using internal API for readdir might be subtly broken after all
sure, what's it?
Limbic_Region else { say "foo" } # works without ; as it should
else { #say "foo" } # does not work regardless of ; or not 15:51
actually - I am not sure why I even mentioned the ; other than multi-statements was part of my testing
though if the open and close brace are on diff lines - works as expected
audreyt uhm.
Limbic_Region the Thread.yield does not fix the prob
audreyt you mean you comment out the rest of the line
including braces
and expect the brace to reassert itself? 15:52
[particle] thinks l_r has an in-head parser problem
Limbic_Region reminds himself it is Monday
ok - so ignoring my ignorance for a second, you should be sleeping though if you want to stay up and debug why adding say statements makes this thing work I am happy to oblige 15:53
audreyt Limbic_Region: ok. try "close $*ERR"
in place of say
Limbic_Region ok - in all 4 places?
audreyt yup
Limbic_Region and btw - was the Thread.yield() supposed to be empty arguments?
audreyt yeah 15:54
Limbic_Region k - thought so
audreyt it's noop that helps coop threading on systems without true preemptive threading
Limbic_Region oh, not sure if this does anything for you but say ""; and say 1; do not produce the same results 15:55
say 1 in 4 places works, say "" does not
and close $*ERR in all 4 places didn't fix the prob either
15:55 mako132 joined
audreyt ok. please do a -d dump 15:56
one for say1 and one for say""
and put them up somewhere (nopaste, upload to feather, etc)
pugs -d yourprog.pl > out.txt
Limbic_Region ok - I am verifying what I just said in case I missed one of the 4 places in part of my experimenting
audreyt reading -d dump will almost certainly show where it went wrong 15:57
Limbic_Region yep - I wasn't having another Monday moment - will nopaste for you momentarily
audreyt a poor subst of true interactive debugger, but hey :)
audreyt remembers she woke up for food, and goes to fetch some. bbiab 15:58
Limbic_Region last few times I have tried -d it hasn't produced any output that the prog wasn't already btw, but will run on this particular prog to make sure
audreyt really? I'm surprised
what does
pugs -de1 15:59
print on your system?
Limbic_Region was assuming it was a Win32 thing
audreyt not likely
Limbic_Region ooh, that's printing lots of stuff
audreyt good
including each subexpression, context, nesting depth, etc etc 16:00
Limbic_Region not good
on the real prog - no debugging output (perhaps because it exited abnormally?)
audreyt it's huge, which is why I don't think nopaste = good idea
ugh
what about with "say 1"?
16:00 kanru joined
pasteling "Limbic_Region" at 129.33.119.12 pasted "this is with 4 say "" and -d" (17 lines, 39B) at sial.org/pbot/19187 16:01
Limbic_Region running now
ok, even more bizzare results 16:02
not only did running with -d not produce any debug info
it also caused the previously working config to not work
16:03 mdiep joined
audreyt :/ 16:03
ok, clearly my internal posix-emu hack doesn't like itself 16:04
this is GHC 6.4.2, Win32, right?
I'll boot to win32 and debug it tomorrow
Limbic_Region yep
16:04 tup joined
audreyt and for now -- arrange to say "\n" or something so you don't get blocked 16:04
Limbic_Region for the purposes of getting the test examples thing done - I will just turn debugging on to write it
audreyt k. sorry for all this trouble, and thanks for persisting :) 16:05
Limbic_Region no worries
btw - not redirecting to a file causes the say 1 to finish and produce the correct results
so to recap - using -d on a prog.name doesn't appear to do anything though -e from the command line works as expected 16:06
redirecting to the file (possibly in conjunction with -d) causes the working config to not work
say "" does not behave the same as say 1
and finally - there must be non newline IO in those 4 places for it to work at all 16:07
but apparently the -d on prog.name is only this prog specifically as a simple use v6-alpha; say "foo"; in foo.pl produce the correct debug output 16:10
Limbic_Region wanders off to finish writing the test examples stuff
16:13 fglock joined 16:22 Revision17 joined
Limbic_Region someone refresh my memory - if I do my $out = `some command`; and I want to know if `` ran correctly - what var do I check - I assumed $! 16:23
16:24 Psyche^ joined, Psyche^ is now known as Patterner
fglock I compared the v6.pm in svn with the latest CPAN version: 2.2x faster, 3x less ram :) 16:26
Limbic_Region fglock plus plus 16:27
16:27 cmarcelo joined
fglock Limbic_Region: :) 16:28
Limbic_Region oh wait, this should be wrapped in an eval anyway so I guess I should be using system and checking $@ 16:30
TreyHarris Limbic_Region: $?
you are writing Perl 5 code, right :-)
Limbic_Region no, Perl6 16:31
err Perl 6
TreyHarris oh.
audreyt Limbic_Region: $! is correct. 16:32
svnbot6 r12565 | fglock++ | * Pugs-Compiler-Rule 0.16 16:33
Limbic_Region there must be something else that does this that I can steal - my brain is refusing to work today
audreyt there's no $@ and there is no $? anymore
Limbic_Region audreyt - ok, don't yell at me if I tell you that it is empty then
pass or fail
audreyt that's because you don't have r12566 :D 16:34
svnbot6 r12566 | audreyt++ | * `` and qx{} now sets $! as they should.
Limbic_Region oh
Limbic_Region svn ups
audreyt (j/k, thanks for prompting the fix.)
Limbic_Region well, I am almost finished with the test examples script 16:35
was going to make a compliment one for verifying the code runs not just compiles
audreyt - what about system() - shouldn't that be setting $! too? 16:36
audreyt sure
they all do now
Limbic_Region k
so now go to sleep 16:37
audreyt obliges
:)
Limbic_Region sleep well
TreyHarris night audre
urgh
audreyt
16:37 justatheory joined 16:39 justatheory joined 16:43 justatheory joined
svnbot6 r12567 | fglock++ | * v6.pm 0.014 16:46
16:48 masak joined 16:49 SageLT joined
Limbic_Region doesn't have the heart to tell audreyt that $! still isn't set if system() is wrapped inside an eval { } 16:51
svnbot6 r12568 | fglock++ | v6 - MANIFEST had some extra test files
audreyt you mean try{}
sure, because try{} resets $!
so you need to simply run system
and then check $!
without using try{}
Limbic_Region well, no I was using eval because I forgot about try -
audreyt same... eval"" resets $! as well 16:52
Limbic_Region no complaints about using eval btw
audreyt so don't use either :)
cmarcelo audreyt: is "toRef i = return $ unsafeCoerce# i" too abusive? considering its just for Int..
Limbic_Region but system produces all kinds of undesireable output so `` it is
audreyt cmarcelo: is Value 32bit always, or is it native length? 16:53
cmarcelo: as long as it's native length then unsafeCoerce# is fine
cmarcelo native
Limbic_Region grrr
when I run ./pugs -c examples/cribbage_scoring.pl $! = 1 16:54
cribbage_scoring.pl doesn't parse correctly
when I run ./pugs -c examples/progressive_powerset.pl $! = 1
16:54 weinig is now known as weinig|away
Limbic_Region progressive_powerset does parse correctly 16:54
$! isn't a lot of help there
so just drop the -c 16:55
that just feels wrong but grrr - I am going to finish this today
wait - that is showing $! in both cases too 16:56
cmarcelo audreyt: what was blowing evalbot, the 1+1? there's a simpler test (other than running my own evalbot) I could make here to check before commit? 16:58
17:02 merlyn joined
audreyt cmarcelo: any expr at all, 1+! would do 17:05
./pugs -e1
would do
./pugs -e 'sub f {}; f()'
to be utterly sure
cmarcelo ok.. i'm looking at HashIO now, then I'll test it.. 17:06
Limbic_Region audreyt - should `pugs.exe -c examples\foo.pl` work? It currently only works if I do examples\\foo.pl (note double backslash) 17:07
audreyt Limbic_Region: it has never worked in perl5 17:10
so I doubt it would suddenly work in perl6
(`` is a qq-interpolator) 17:11
g'nite :) &
17:12 hexmode joined 17:21 Lorn joined 17:29 b_jonas joined 17:37 jferrero joined 17:38 holoway joined, b_jonas joined 17:41 foo\ joined
Limbic_Region wooo hooo finally 17:41
svnbot6 r12569 | Limbic_Region++ | audreyt++ for the assistance 17:42
Limbic_Region not pretty but it works - I would truly appreciate some advice or cleaning up the workarounds in rev 12569
17:42 kanru joined 17:49 marmic joined
fglock is S03, list infix ';' is in the same category as 'ōæ½xA5 <== ==>' ? 17:50
17:56 meppl joined
Aankhen`` Anyone know what it means if SVK says "Incomplete data: Delta source ended unexpectedly" when syncing? 17:58
prefiks Aankhen``: look at topic 17:59
gaal Aankhen``: svk sy -s 12103 //mirror/pugs
svnbot6 r12570 | audreyt++ | * Pugs.Parser.Operator: Resync with S03's optable, adding
r12570 | audreyt++ | support for ?& (the non-short-circuiting boolean AND operator).
r12570 | audreyt++ | * Also parse for "ff" and "fff" but they are not yet implemented.
r12570 | audreyt++ | The old "till" form is now gone.
Aankhen`` Ah.
prefiks: Thanks. *sheepish grin*
gaal or 'svn.openfoundry.org/pugs' instead of //mirror/pugs if that gave an error
the topical solution didn't workforme, fwiw
Aankhen`` I seem to have an up-to-date SVN::Mirror.
gaal and: Aankhen``! hi! :-) 18:00
Aankhen`` Hihi gaal. :-D
Ooooh, guess what?
I leave India in 9 days!
gaal where to? :)
Aankhen`` Canada.
gaal excellent! lots of moose in Canada. 18:01
Aankhen`` Aye. =)
The SVK command worked, thanks. :-)
gaal moose!
Mein Luftkissenfahrzeug ist voll von den Aalen.
Limbic_Region so of the 158 example scripts in examples/, 46 fail to parse correctly - not sure if that is better or worse than expected 18:03
svnbot6 r12571 | cmarcelo++ | * HsJudy: Fixes usage for negative numbers as keys and values.
Limbic_Region s/dates/dated/ 18:04
Limbic_Region begins trying to cleanup the non-parsing examples though his perl 6 knowledge is quite dates
Aankhen`` That was weird.
[23:34:16] <Limbic_Region> s/dates/dated/
[23:34:17] * Limbic_Region begins trying to cleanup the non-parsing examples though his perl 6 knowledge is quite dates
Is that how you sent them?
gaal Limbic_Region: your Perl 6 is better than that of most people I know. 18:05
go for it :)
Limbic_Region I think so
I think so to Aankhen``
gaal predictive error correction!
Limbic_Region gaal - your help cleaning up some of the workarounds in t/examples/all_parse.t would be appreciated if you have tuits
gaal Limbic_Region: not really, sorry
Limbic_Region Aankhen`` - no, I sent them in the reverse order but the client I am using is likely sending in incorrect order 18:06
Aankhen`` Oh, I see.
Limbic_Region CGI:IRC is not the best client in the world
but it is the only way I can be here from $work
Aankhen`` All ports blocked? 18:07
Limbic_Region Aankhen`` - well, obviously not http/https but those have to go through a proxy - so for all intents and purposes - yes 18:08
Aankhen`` Hrm.
So they have to specifically whitelist ports for them to be open?
Limbic_Region Aankhen`` - I work for the U.S. Government - specifically one that focuses on security and secrecy 18:09
nuff said
Aankhen`` Heh.
gaal there's an alternate U.S Government that doesn't focus on security and secrecy? 18:11
[particle] yes, it's in iraq 18:12
Aankhen`` LOL.
TreyHarris gaal: lol, I didn't realize you were at a half-hour offset timezone. I went back almost a day looking for that comment from Limbic_Region at :34 past one of my hours :-) 18:13
gaal I... don't believe I am!
IDT is GMT+3
TreyHarris gaal: i meant Aankhen``, I tab-completed once too many times 18:14
Aankhen`` Heh.
Yeah, that half-hour has thrown off many people. 18:15
gaal noticed a bug in IBM's "timzezone" wallpaper
bbiab &
TreyHarris Limbic_Region: I've noticed that many *tests* are still using antediluvian pre-Synopsis Perl 6. 18:16
clkao ooo
$ ab -clkao- [~/work/pugs/perl5/Pugs-Compiler-Perl6] time perl -Ilib -It -MTest -e1
TreyHarris I've been cleaning them up as I come across them
clkao 1..0
clean compile
0.761u 0.051s 0:01.26 64.2% 10+6498k 0+0io 0pf+0w
Limbic_Region TreyHarris - well, do you happen to know if subtypes are currently parseable using Pugs? 18:19
examples\hq9<plus> is currently failing to parse - I thought because s/subtype/subset/ but even that is still failing to parse - so I am thinking pugs hasn't implemented yet? 18:20
TreyHarris Limbic_Region: doesn't seem to be, just cut-n-paste Larry's first example in S12, parsefail 18:21
Limbic_Region TreyHarris - these aren't tests, these are examples
the test is if they parse/compile or not
TreyHarris Limbic_Region: i'm sorry, I don't understand what you're saying. If it doesn't parse, it doesn't parse, example or test, right? 18:22
TreyHarris is still not caffeinated enough so may be a bit slow :-) 18:23
Limbic_Region TreyHarris - there are 158 example scripts which we are not testing and therefore, not keeping up to date. I proposed adding a new test script that would iterate through them testing if they compile or not. 18:24
TreyHarris yes, i understood that... but i thought we were talking about subset not parsing. 18:25
oh, were you responding to my comment "I've noticed that many *tests* are still..."?
Right, I know it's different
Limbic_Region I am - I was trying to fix examples that were fixable - in this case it is not
fglock clkao: hi
TreyHarris But my point is, if the tests aren't even up to date, it doesn't surprise me that the examples are even farther out of date :-) 18:26
18:26 Eimi joined
clkao fglock yo 18:26
TreyHarris commute &
Aankhen`` Hrm... Agent Zhang mentioned a util/smartlinks.pl, but I can't seem to find it.
18:27 buubot joined
fglock clkao: did you rm Test.pmc first? 18:27
TreyHarris Aankhen``: have you pulled a pugs tree down from svn?
it won't be in a binary build
Aankhen`` I have the source checked out from Openfoundry.
TreyHarris or the last release
do you see the util directory?
Aankhen`` Yes. 18:28
I see a lot of files, but not smartlinks.pl.
fglock clkao: oh, you are just testing '1', ok
TreyHarris it should be at the top level, the same place where Makefile.PL and configure and t are
clkao fglock: ya. really fast
TreyHarris ok
clkao in fact, superfast. almost sane as precompiled pmc 18:29
TreyHarris Aankhen``: run "svn stat -uq util"
(or, if you're inside the directory, s/util/./)
Aankhen`` Uhh... how does that translate to SVK?
TreyHarris do you get an "A smartlinks.pl"?
oh
um.
Aankhen`` Hang on, it might be my fault.
TreyHarris dunno :-)
Aankhen`` I think I forgot to `pull`. *bangs forehead on wall* 18:30
TreyHarris smartlinks.pl was only added as of r12368
ok, really gotta run
&
Aankhen`` Ciao.
18:31 Eimi joined
svnbot6 r12572 | Limbic_Region++ | renamed from mandel-p5.pl to allow test scripts to distinguish 18:31
Aankhen`` Thanks for trying to help, even if the fault was my own cluelessness. :-)
Limbic_Region how do you remove mandel-p5.pl from the repo? I assumed svn del would do it? 18:32
svnbot6 r12573 | Limbic_Region++ | s/subtype/subset/ though still fails to parse due to pugs
Limbic_Region gives up for the day not wanting to screw things up 18:34
svnbot6 r12574 | fglock++ | v6 - pod: command line switches, environment variables documentation 18:36
fglock clkao: v6.pm is doing around 100 lines/s here 18:38
18:39 Eimi joined 18:40 justatheory joined 18:42 justatheory joined 18:45 Eimi joined 18:53 elmex joined 19:07 penk joined
Lorn hi ppl, i install v6.pm here, and work with file.pl but in command line show this error 19:17
perl -Ilib lib/v6.pm -e 'for 1,2,3 -> $x { say $x }'
Can't open perl script "lib/v6.pm": No such file or directory
silug shouldn't that be '-Mv6' not 'lib/v6.pm'? 19:18
TreyHarris Lorn: I think you want -Ilib -Mv6
19:19 weinig|away is now known as weinig
TreyHarris actually, you want -Mv6-alpha 19:19
-Mv6 will complain that you're not running Perl 6 natively
Lorn perl -Ilib -Mv6-alpha -e 'for 1,2,3 -> $x { say $x }' 19:20
Can't open -e for input:
No such file or directory at /usr/lib/perl5/site_perl/5.8.8/Module/Compile.pm line 200.
BEGIN failed--compilation aborted.
gaal hi, quick survey: among those present, who is using GHC 6.4.x? Is a 6.5 snapshot hard to install on your platform? 19:21
TreyHarris Lorn: i'm getting the same thing 19:22
19:22 weinig joined
TreyHarris gaal: i am, and attempting to do a ports install of 6.5 cause such badness--neither 6.5 nor 6.4 would work anymore--that I had to restore from backup :-( 19:22
Aankhen`` gaal: I am using 6.4.2, more out of inertia than anything else. I doubt the snapshots would work happily on Win32 anyway. 19:23
gaal OS/X right?
Lorn TreyHarris: i will try -Ilib lib/v6.pm because cpan, recommend this in SYNOPSIS
TreyHarris audreyt helped me for several hours before that point, too, so it apparently wasn't something easy/obvious
gaal: yep.
gaal Aankhen``: FWIW I'd used earlier 6.5 snapshots with no problem at all from binary dists on Win32
Aankhen`` Okay. 19:24
If it works, and Pugs needs it, I'll gladly install 6.5.
TreyHarris Lorn: yes, that seems to work, though with a number of sub redefinition warnings, and something about "Too late to run INIT block at ..v6.pm line 22" 19:25
gaal Aankhen``: 6.4 forces us to do some ugly workarounds in the hs code. from what I've seen, it takes a little longer to build pugs with it but the resulting pugs is much faster
6.6 should be released within a month 19:26
Aankhen`` The resulting Pugs is much faster with 6.5?
gaal once it does we're anxious to move over to it
Aankhen``: yes.
Aankhen`` Hey, if a new, stable version of GHC is released, and someone reminds me about it, I'll upgrade in a heartbeat.
Lorn TreyHarris: ok, i waiting for fglock for some help, thanks ;) 19:27
19:27 anov joined
gaal Aankhen``: why wait? :) it's a biggish download but it's easy to install on windows 19:27
just remember to update your path :)
Aankhen`` gaal: Eh, 6.5 is not stable, is it?
gaal and you can have more than one ghc installed.
no, but it works :) 19:28
Aankhen`` Heh.
Okay, well, I'll do it in the morning. I don't have the brainpower to do it right now.
gaal for that matter... pugs isn't stable either :;p
Aankhen`` Pugs is special. ;-)
gaal ok, suggestion, start the d/l tonight
Aankhen`` Good idea.
TreyHarris gaal: will 6.4 just stop working soonish? if so, i guess figuring out why i can't get 6.5 working becomes a harder priority--though i note that even haskell.org says there's no good installation for OS X
gaal TreyHarris: audreyt and I are anxious for it, but of course we won't move until everybody's accounted for 19:29
Aankhen`` hc-6.5.20060819-i386-unknown-mingw32.tar.gz
Is that the one I should download?
gaal if you can spare the moose to get it to work, the GHC community will thank you for it 19:30
Aankhen``: I think there are .msi install versions of it.
let me look....
Aankhen`` Not that I can see.
(I'm looking at www.haskell.org/ghc/dist/current/dist/
lambdabot Title: Index of /ghc/dist/current/dist
Aankhen`` )
TreyHarris gaal: well, my working knowledge of Haskell is limited to my first CS course in 1992, and my knowledge of GHC is nil. But I'll poke at it. :-)
19:31 fglock joined
gaal Aankhen``: hmm, that's weird, I coulda sworn the old 6.5 I have wasn't installed from a tarball... oh well, at least that's still a binary dist :) 19:32
Lorn fglock: :) i have problems with v6.pm
Aankhen`` :-)
fglock lorn: in order to use lib/v6.pm as a script, it needs to be there - either cd to the dir, or make a link
'perl -Ilib lib/v6.pm' - with this syntax, it's not looked up in PERL5LIB 19:33
it is a plain script 19:34
Lorn fglock: hmm, in script work fine, but in command line, dont work, the TreyHarris tell me to try this perl -Ilib -Mv6-alpha -e 'for 1,2,3 -> $x { say $x }'
but
Can't open -e for input:
No such file or directory at /usr/lib/perl5/site_perl/5.8.8/Module/Compile.pm line 200.
fglock Lorn: it's a Module::Compile and v6.pm problem - would you like a commit bit to fix it? 19:37
19:40 alinbsp left
Lorn fglock: [pt-br] desculpa, mais eu esqueci que o ōæ½xE9 would :) e nōæ½xE3o entendi o significado da frase[/pt-br] 19:40
fglock Lorn: at the time -Mv6-alpha loads the compiler, -e was not executed yet 19:41
Lorn fglock: ok
fglock Lorn: [pt-br] " vocōæ½xEA quer acesso de escrita no site para resolver o problema? "
TreyHarris fglock: can -Ilib -Mv6-alpha -e'...' be made to work? that other syntax is confusing to me 19:42
19:43 MacVince joined
Lorn fglock: i understood, no, i try to resolv, if i can, i send the patch or upload them (after resolv :) ), ok? 19:44
fglock TreyHarris: ingy and audreyt were working on it
Lorn: ok :) 19:45
TreyHarris fglock: thanks 19:46
fglock meeting & 19:48
19:48 fglock left, fglock joined 19:51 zgh joined
svnbot6 r12575 | cmarcelo++ | * After fixing HsJudy, switch back to IntMap in Pugs.Parser.Operators. 19:51
cmarcelo does evalbot updates by itself or need some trigger? 19:52
?eval 1+1 19:56
gaal cmarcelo: it updates itself. 20:07
unless it blows up! 20:08
audreyt: I just committed to a local branch a version that doesn't #include Pugs.Code.Val, thanks to the snaphot ghc fixes. 20:12
two quirks I've noticed: 20:13
first, I can still think of no way to separate the various primitive Pures to different files. the class ICoercible must be visible everywhere, apparenly *with* methods, but the types for the methods can't show up in the boot file 20:14
because e.g. if you say class ICoercible where asBit :: moo -> PureBit, Bit.hs will refuse to consume that boot file 20:15
I "solved" that by leaving Pugs.Val.Base for now.
the other thing is that consumers needed a Pure class, but I couldn't put the full qualification of the class in the boot (because then it'd need to mention ICoercible, etc.). So I just have in the boot file 'class Pure a' 20:16
I *think* that's all right? 20:17
anyway, this does feel a little better. getting rid of the include is great :) but I wonder if we should petition for a ghc treatment of issue #1 in 6.6? 20:18
not that I can see how that could work.
since the point of a boot file is that you don't yet know what the incoming type really is, so you don't know it really is just a reexport of a type you yourself will have supplied to the consumed module 20:19
-fin- 20:20
cmarcelo audreyt: P.P.Operators should work with IntMap now. my feather builds and works fine. dunno why evalbot didn't came back. I wonder if some of the segfaults you were having are related to this.. 20:22
20:22 fglock joined 20:24 cmarcelo left 20:34 bpphillips left 20:44 coke_ joined
coke_ anyone here interested in CIL as a backend for pugs? hurry up and join #parrot on irc.perl.org 20:44
seen audreyt 20:45
20:48 cj joined
cj audreyt: muahaha 20:48
zgh coke_: is this the same CIL? manju.cs.berkeley.edu/cil/
lambdabot Title: CIL Documentation (v. 1.3.5)
beppu I'm not a bot, but according to my scrollback, audreyt said: audreyt g'nite :) & :at 10:11am PST
cj beppu: I figure audrey has a scrollback :) 20:54
aw man... someone let merlyn in here?
coke_ for this guys thesis, but he seems to think using pugs to target CIL won't do for thesis requirements.
ah well.
zgh careful he's on a cruise, that seems to make him cranky 20:55
cj zgh: he's more interesting when he's cranky :)
zgh hee
webmind env 20:57
oops
coke_ give him a drink, and back away slowly. 20:58
cj coke_: it's funny 'cuz it's true 20:59
I've got hiccups... again
woe am I
coke_ vaguely wishes that all these psuedos were mapped to 'have I ever run into this person at a con' images. 21:00
cj I'll help you out with that one...
colliertech.org/images/friends/me/cj.jpg 21:01
or with the long: colliertech.org/images/friends/me/t...mb-500.jpg
coke_ hurm. nope. 21:02
also nope. 21:03
=-)
zgh coke_: were you at OSCON this year? 21:04
coke_ just yapc::na
zgh some of my wo-workers made it to that one, I only made OSCON, maybe next year ^_^
coke_ I'm actually hoping houston hits next year. Used to live there, nice to visit again.
zgh wo-workers? 21:05
[particle] coke... on #perl6?
[particle] faints
coke_ bite me, jerry!
[particle] get lost, enron!
coke_ you're just jealous that tcl on parrot is going to be done before perl6!
and we've only got 2 programmers!
[particle] thanks to mdiep!
mdiep *blink*
TreyHarris cj: hey man 21:07
21:07 frederico joined
cj hey there TreyHarris 21:09
I missed your talk the other day
sorry :(
pushing out rss code for the plogs 21:10
TreyHarris no prob, if you're on here, i probably didn't hit anything you don't already know :-)
cj I figured. I saw your name in the perl6 book. you're famous :) 21:11
TreyHarris i had to skip given/when, gather/take, captures and regexes anyway because there was no way to effectively cover it in the time alloted. (1 hour is *not* long enough to cover the major changes in Perl 6, even if you just want to say a sentence or two about each and show a code sample)
cj it's like it's a new language or something ;) 21:12
zgh can't believe that his perl4 code won't run in perl6, it runs in ruby! 21:13
coke_ ... there's a perl6 book?
zgh kidding
TreyHarris actually, i could've covered gather/take... dunno why i didn't.
zgh sort of
TreyHarris coke_: very out of date, but yes
coke_ not the parrot one, but the other one? 21:14
cj isn't it called "perl 6 and parrot"?
coke_ ah. dan's book. k
TreyHarris coke_: Sugalski & Randal, Perl 6 & Parrot Essentials
21:14 ludan joined
cj www.amazon.com/gp/product/059600737...11?ie=UTF8 21:15
coke_: I thought it was allison's book ;)
coke_ heh
... you say randal and I think merlyn, not allison. all too fconfusing. =-)
cj I'm giving her the credit because she signed mine
[particle] you should meet allison schwartz! 21:16
talk about confusing...
cj hiccup
21:16 penk joined
cj imagines allison + randal babies 21:16
we should keep that from happening
TreyHarris lol 21:18
cj TreyHarris: have you seen this? 21:20
colliertech.org/~cjcollier/spug.jpg
leo TreyHarris: heh - that book has 3 authors - careful 21:21
[particle] grins 21:22
which author is responsible for all the outdated stuff? 21:23
cj leo: oh, I didn't see you there :)
leo: you going to be up here some time soon? I still need a couple of signatures :)
er, autographs
TreyHarris leo: nah, the other one didn't do anything from what i hear ;-) 21:24
seriously, though, leo, i meant no slight, was just doing the usual citation thing of mentioning only the first two authors in order on the cover 21:25
[particle] sure, leo is short for 'et al' 21:26
cj heh
leo 's considering a new name like 'Abel' 21:27
nothingmuch audreyt: ping
leo [particle]: I'm not short - 1m86 ;)
[particle] :) 21:28
TreyHarris nothingmuch: it's not even 6am there
21:28 coke_ left
nothingmuch TreyHarris: that means she is probably still up ;-) 21:28
TreyHarris no, she went to bed four hours ago
nothingmuch seriously though, audreyt never had any direct correlation with any one time zone
oh, i didn't see that
nothingmuch needs a pause admin... =/ 21:29
cj leo: you could do what I did. marry a girl with a last name that starts with A, and then hyphenate using the alphabetically superior one first
or guy, or whatever
leo cj: thanks ;)
[particle] hrmm. i think pugs should have a -Wall option, and perhaps a -Tang option 21:31
TreyHarris irssi help... why does '/last -regexp audreyt\>\s 5' not pull up anything? i have several days buffered and can see audreyt's last
cj [particle]: both good ideas :) 21:32
TreyHarris hm. i remember a CPAN module that did readline prompts really easily and also switched to STDIN transparently for piped usage... anybody know which one i'm talking about? 21:33
beppu wasn't the old joke something like: if gcc has a -Wall, shouldn't perl have a -Stallman?
cj TreyHarris: no idea. searching for your last works fine for me :)
leo particle: you know, parrot hat for quite a time a -Dan_Sugalski gcc option (iirc)
cj TreyHarris: but not with \>\s 21:34
beppu funny
TreyHarris cj: right. audreyt's nick is mentioned so frequently i was trying to get just her actual comments
svnbot6 r12576 | lwall++ | For some reason the example parser test loop stops early without try {...}.
cj TreyHarris: so I'm guessing that there's a problem with the \>, maybe?
[particle] leo: i had no idea! 21:35
cj TreyHarris: if you strip the \ before the >, that might work
TreyHarris cj: i can't make any backslashed item work. makes me wonder if -regexp is there more as a hope than as a feature 21:36
cj TreyHarris: you can always check. want an svn url? :)
wolverian since when has > been a special character anyway? 21:37
TreyHarris cj: no. i spent 40 minutes i didn't have prior to my presentation trying to get irssi built on a company host so that I could demonstrate evalbot. long enough to make me not want to deal with irssi anytime soon :-)
wolverian: it's not, but i generally escape non-\w's just to be sure, especially when i'm a layer removed from the match 21:38
cj TreyHarris: heh, yeah. it can be a pain sometimes.. 21:39
21:39 cjeris left
wolverian TreyHarris, well, it failed in this case. :) searching for a plain > works fine here, and searching for e.g. \? works as well 21:39
I don't know why \> doesn't work, though. 21:40
21:40 SageLT joined 21:41 justatheory joined
TreyHarris wolverian: it just doesn't like backslashes. "/last -regexp audreyt>.*utterly" works, "/last -regexp audreyt>.*\sutterly" doesn't 21:42
lol, except in finding the comment i just made. so apparenly backslashes are being autoescaped
which rather dilutes the power of having perl regexps available in the first place...
wolverian TreyHarris, yup. 21:59
21:59 prefiks left
cj heh. LNG = Last Known Good... sound good to you folks? 22:08
22:13 kanru joined
svnbot6 r12577 | lwall++ | Cleanup of some hop6 syntax errors. 22:19
r12578 | fglock++ | * v6 - this command line now works (note that there are 2 strings, one 22:29
r12578 | fglock++ | is for loading the module, and one is the perl6 code):
r12578 | fglock++ | perl -e 'use v6-alpha' ' "hello world".say '
22:32 baest joined
svnbot6 r12579 | lwall++ | fixups to snowing.pl 22:32
r12580 | lwall++ | syntax updates for motd-i 22:38
TreyHarris nothingmuch: did you find any pause admins? 22:44
svnbot6 r12581 | lwall++ | syntax updates for rpm_calc.pl
TreyHarris In r12579: 22:46
-sub sleepy{for 1..5500 {;}}
+sub sleepy{for(1..5500) {;}}
why?
svnbot6 r12582 | lwall++ | operator \\ must be named infix:<\\\\> or infix:{qn'\\'} 22:56
Juerd Holy crap 22:57
We have a \\ operator?
wolverian in examples/ 22:58
Juerd dot.kde.org/1156199214/ # Given the size of the KDE project, and the size of the parrot/pugs projects, we're doing quite well with feather's specs, I think :)
lambdabot Title: Immanuel: KDE Introduces New Server
22:59 SageLT joined
svnbot6 r12583 | lwall++ | syntax updates (still doesn't pass though) to 01-10interpolating-functions.pl 23:02
SamB it probably isn't for array update? 23:03
TimToady TreyHarris: for() is a function call, not a statement control.
SamB that doesn't look like a good way to sleep either 23:04
23:05 bcorn joined
TreyHarris TimToady: where is for() defined then? 23:06
TimToady for() is what it used to be. I changed it to for 1..5500 23:07
your diff is backwards
TreyHarris oh... so it is. 23:10
TimToady for() is banned by S04, about line 955
TreyHarris grins sheepishly and goes back to what he was doing
the "r12579" makes it oh so easy to just stick a - right before it in the svn diff command... :-) 23:11
i should write a script so i won't make that mistake again
line 878 of S04. 955 is the last line of the file... 23:13
TimToady .pod? 23:14
TreyHarris yes
TimToady yeah, you're right,
I was reading the wrong part of my vim status line
sorry.
TreyHarris we can both feel sheepish then :-) 23:15
TimToady darn, my first mistake of the year... :)
23:21 mdiep joined 23:22 jferrero joined, b_jonas joined 23:35 markstos joined 23:53 Bit-Man 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!