pugscode.org/ | nopaste: sial.org/pbot/perl6 | pugs: [~] <m oo se> (or rakudo:, kp6:, smop: etc.) || We do Haskell, too | > reverse . show $ foldl1 (*) [1..4] | irclog: irc.pugscode.org/
Set by TimToady on 25 January 2008.
00:04 eternaleye joined
spinclad it looks likely that it would impose numeric context on both sides, ie. try to interpret the strings as numbers. 00:07
Eevee I suppose that's really the question then 00:17
will strings still automatically cast to numbers
TimToady it depends simply on how the default infix:<->(Any $x, Any $y) is written 00:22
unless someone explicitly writes infix:<->(Str $x, Str $y) 00:23
at minimum, you'll get a couple of warnings out of it 00:24
Eevee well, would hope so
just wonder if they'd be fatal or not
I can't imagine where trying to subtract "apple" is what someone wants 00:25
TimToady we're trying to avoid fatals because we want to be parallelism friendly
so more likely is to return an unthrown exception
if you're hypersubtracting two arrays, you don't want to blow up everything just because of a bad apple or two 00:26
Eevee groan
[particle1 Failure++
00:26 [particle1 is now known as [particle]
TimToady what we have here is Failure to communicte. :) 00:26
*communicate even... 00:27
Eevee well 00:28
damn, that sucks both ways 00:29
if it's fatal, I just lost the entire operation
if it's not, I have to either loop through both beforehand and error-check or loop through the result and look for a nonresult (or forget to do anything and have the exception floating around until who knows when) 00:30
00:32 justatheory left
Auzon Either way, you need to be sure that your array doesn't have any strings in it if you're going to subtract. 00:33
TimToady you're probably going to do some kind of reduce operation at some point, and you just use one that dispatches to a Failure case 00:34
and that can determinate policy 00:35
01:14 kanru left
obra TimToady: I've been hacking at MAD a bit. Would you be ok with everything getting shifted into packages under MAD::? 01:15
01:16 pbuetow left 01:22 BinGOs_ joined 01:27 BinGOs left 01:36 BinGOs_ left
TimToady obra: is fine by me, as long as it's being maintained 01:55
01:56 pascaldamian2 left
mncharity backlogging... laughing... 01:59
"apples" - "oranges"
I guess that should be ļ»æ"apples" <=> "oranges" 02:00
TimToady now one wishes to make a pun on Orange Free State...
mncharity oyyyy
Hmm. "The Perl 6 Joke Book"... 02:02
maybe someone should start a file...
Tene I already have one. ~/irclogs/freenode/#perl6.* 02:03
mncharity looking... 02:04
oh, joke. got it. didn't realize there was a ... . oh. joke. got it... maybe 02:05
;)
02:05 justatheory joined
mncharity ok, don't get it. explain? 02:08
Auzon All his logs are The Perl 6 Joke Book 02:10
mncharity not psyched to hack STD_red this evening. :/
allbery_b for some reason I thought TimToady said that 02:11
(about irc logs)
TimToady wonders if irc is short for "I recall correctly"
mncharity finish sync'ing std.rb; cross fingers; attempt parse of "most of" t/; debug; try again; achieve "finally have a second (non-pugs) parser for t/", one willing to talk about objects. fork elf_f. attack tests. smoking... 02:13
TimToady I think STD5 is getting pretty close to functional 02:14
mncharity :) !
TimToady was astounded to discover that tre doesn't match "\n" against [\n\f] 02:15
Auzon STD5 is the Perl 5 translation of STD.pm?
TimToady in particular, it's the *automated* translation
mncharity \n\r?
obra TimToady: I'm trying to get MAD to the point where it will actually be run as part of the 5.10 tests by default
TimToady no, was trying to match \v
mncharity oh, [\n\f] eeep 02:16
obra "force others to maintain it through trickery"
TimToady getting all the ws, unv, vws rules tidy is a pain if you're trying to do LTM right
well, there are some unicode chars in there too
mncharity ah 02:17
TimToady but it's really hard to make much progress parsing Perl if you can't recognize vertical whitespace...
obra++
*obra++
mncharity re pain, yeah. hmm. /me wonders about implication for maintenance and modification... 02:18
TimToady sorry, C joke...
mncharity specifically, a C ++ joke.
TimToady rofl
obra *snicker* 02:19
Eevee ha
mncharity :)
TimToady well, the main point of STD5 is to work through all the implications of real LTM
and specifically autogenerated LTM
obra is trying to decide what the right thing to do about the XML::Parser dependency is
mncharity re point, nod 02:20
TimToady MAD doesn't have to spit out XML, it only has to spit out something recognizable by nomad, or whatever lib it's gotten itself turned into
so while I can cheat to some extent, I often choose not to
last night and this morning I hacked (hopefully correct) longest-token backoff into the lexer 02:21
so we should be able to put back the commented rules that start with <ident> pretty soon
assuming we don't want to leave it the way it is for performance
should probably be special casing for ident tokens anyway 02:22
but that's just an optimization
obra nods
mncharity re put back, np. one of the advantages of the otherwise painful manual decoupling of STD_red from STD.pm is I've just kept them in. 02:23
TimToady main thing is to preserve correct semantics, which probably means I should break apart term:name now just to force the issue
obra I don't actually have a huge issue with the xml-as-metaformat. It's more of a "the existing Nomad side of things would require refactoring to not use XML::Parser."
But cheating and creating a baby-mad-xml parser seems plausible
TimToady well, if I'd done it a few months later I'd probably have tried to use yaml instead 02:24
obra nods
mncharity re XML, I note my startlement at the poor performance of yaml (hmm, was it just ruby's, p5, or both (syck)). if performance matters, might consider a foo(bar(...,"...",3,...),hee(...)) function form which can just be eval'ed.
TimToady but yaml has its own issues...
02:25 obk joined
TimToady that could also be made to work 02:25
and has a better chance of passing odd characters that upset xml/yaml
mncharity oh yeah. that, and other "interesting" little bugs. 02:26
obra Yep. I sent a patch to p5p today for the first of those. 02:27
First (two lines of) C I've written in years.
mncharity it seems clear no one is slamming syck with large volumes of generated end-to-end tests. 02:28
*yet 02:29
02:29 obk left
TimToady well, not with unicode in it. :) 02:29
obra :) 02:31
obra mails off a 400k diff to p5p
mncharity even non-unicode. I'm getting fuzzy on the details, but there was the "if the field value is string x, then let's change the field name..."
"ļ»æ(two lines of) C"... ļ»æ"400k diff"...
mncharity thinks "line wrapping"
TimToady thinks "tab expansion" 02:35
mncharity oo, good one :) 02:36
TimToady but probably just mostly Perl, not C
02:37 justatheory left
mncharity sent/sending implies to different patches, but i was slow to notice that... 02:37
*two even
one would be odd
TimToady it's the parody bit 02:38
mncharity EECC 02:39
hmm. EECC! 02:40
ok. let's see... focus...
obra Ah. a nice tidy 57k once compressed
TimToady that's cheating 02:41
though probably a much better measure of actual information content 02:42
[particle] ...as if toke.c isn't compressed.
mncharity if programmer productivity is very roughly constant in lines of code... could you write the compressed form and be 10x more productive?
[particle] off &
Auzon See you, [particle]
TimToady: Interesting idea. Compressed LOC as a new metric? 02:43
TimToady well of course, that's why my editor is cat >/dev/kmem
why do you think it always says "decompressing" when you're booting up linux?
mncharity is somewhat curious if # vi /dev/kmem works... but not enough to try 02:44
obra note that I moved every file in mad/ to saner locales.
so really, I probably only touched about 40 loc non-programatically
TimToady Auzon: yes, but then you'll get comments filled with random passages from classic books
oh, wait, I already do that...
mncharity lol 02:45
Auzon Eh. Most code metrics are flawed.
If not all.
TimToady that's because most code is flawed, I imagine
Auzon Well, to assess productivity :P
mncharity I'd forgotten about the pugs source quotes/poems. That was fun.
obra was pleased. clkao and I went for a week's hackathon. sloccount told us we'd worked for 9 months and it had cost $150,000 to create our output 02:46
mncharity ļ»æAuzon: smaller better?
Auzon That's my opinion on code, yes. 02:47
mncharity "dramatically improve productivity... remove macros from your language..."
"Java's secrets to productivity..."
beans!
I was kind of disturbed the next morning when I realized I had been making "loc is good"-ish irc arguments the night before. *Late* the night before. 02:49
"I am sorry for the length of my letter, but I had not the time to write a short one." Blaise Pascal 02:50
obra And now to see what the porters' wisdom on my hackjob is
But first, Battlestar Galactica. 02:51
mncharity hacking p5! it has everything! ... suspense! ...
obra Thanks for the san-check, TimToady. (madness is still measured on the sanity scale, after all)
mncharity but is it a logarithmicĀ order-of-magnitude scale... 02:52
obra Where does Sage's MAD->6 live?
mncharity ļ»æ"I am sorry for the length of my program, but I had not the language to write a short one." Ben Bitdiddle 02:54
"Perl 6 wasn't ready yet."
ok. brain frie. maybe a couple of hours of p6 tomorrow. 02:55
*fried
good night all. been fun. &
Auzon see you.
02:55 mncharity left 02:57 alester joined
TimToady obra: ./misc/pX/Common/P5_to_P6_Translation 02:59
obra Thanks. There's some good stuff there that should migrate upstream into p5 core. (svn.pugscode.org/pugs/misc/pX/Commo...iption.txt in particular) 03:29
lambdabot tinyurl.com/69z2xd
03:41 syle left 04:01 justatheory joined
Eevee time to take a shot at this test migration thing 04:07
04:09 justatheory left 04:10 wknight8111 left 04:12 vaughn left 04:24 justatheory joined 04:26 justatheory left 04:34 PerlJam left 04:35 wolverian left, pugs_svnbot left, Juerd left 04:37 dalek left, pmichaud left 04:42 Psyche^ joined 05:00 Patterner left, Psyche^ is now known as Patterner 05:10 alester left 05:19 alanhaggai__ left 05:20 alanhaggai joined 05:22 alanhaggai left 05:33 yewenbin joined, Psyche^ joined 05:44 Patterner left, Psyche^ is now known as Patterner 06:04 Alias_ joined 06:46 rindolf joined 06:56 yewenbin left 06:58 meppl joined 07:05 alanhaggai joined 07:14 Schwern joined 07:17 baest left, baest joined 07:37 icwiener joined
TimToady hmm, looks like feather is down... 07:40
pugs: say "howdy" 07:47
exp_evalbot OUTPUT[howdyā¤]
07:50 zamolxes_ joined 07:54 kane_ left
rindolf Hi TimToady 07:55
TimToady: good localtime() 07:56
07:58 zamolxes left 08:08 yewenbin joined 08:11 BinGOs joined 09:10 kane_ joined 09:25 schmalbe joined 09:32 iblechbot joined 10:02 Alias_ left 10:03 Alias_ joined 10:14 Alias__ joined 10:16 Alias_ left, Alias_ joined, Alias__ left 10:17 Alias_ left, Alias_ joined 10:23 wolv joined 10:42 chris2 joined 11:19 wolv left 11:46 rindolf left 11:52 alanhaggai left, dalek joined 11:53 wolverian joined 11:56 pmichaud joined, Juerd joined 11:57 PerlJam joined 12:17 dalek left 12:19 Juerd left 12:20 wolverian left, PerlJam left, pmichaud left 12:26 pbuetow joined 12:49 lerog joined 13:17 charsbar left 13:24 charsbar joined 13:32 schmalbe left 13:59 yewenbin left, yewenbin joined 14:01 dalek joined, wolverian joined, Juerd joined, Juerd left, Juerd joined 14:13 alester joined
moritz_ re 14:27
14:30 jan_ joined 14:31 wknight8111 joined 14:36 dalek left, dalek joined 14:39 yewenbin left 14:46 alester left 14:52 elmex joined 14:54 vaughn joined 15:16 Torment joined 15:27 ChanServ sets mode: +o diakopter, diakopter sets mode: +o TimToady, diakopter sets mode: -o diakopter 15:32 Jedai left 16:06 icwiener_ joined 16:21 Juerd left, Juerd joined 16:22 riffraff joined 16:23 icwiener left
riffraff hi 16:23
16:23 eternaleye left 16:57 rindolf joined 17:19 ujwalic joined 17:29 pmurias joined
pmurias pugs: say "evalbot works" 17:31
exp_evalbot OUTPUT[evalbot worksā¤]
pmurias pugs: say sum(1,3,4)
exp_evalbot OUTPUT[8ā¤]
pmurias shouldn't the sum function be removed?
as there exists [+]
?
moritz_ is it still in S29 17:32
?
I don't see it there
pmurias neither do i 17:33
moritz_ so it's just a pugs artifact
pmurias but is tested in t/builtins/lists/sum.t
moritz_ that's one more piece of work for Auzon++ to remove ;-) 17:35
actually we should have a test file with deprecated builtins
that tests for the non-existance of a sub 'sum'
or you can just remove it for now, if you feel comfortable with it 17:36
pmurias i'm not sure we should check for non-existance of anything 17:37
allbery_b any complaitn implementation is allowed to provide more than 6.0.0-STD 17:38
moritz_ allbery_b: is that officially specced?
allbery_b I believe so, that's the point of having the -STD stuff
it's kinda the language spec version of 'does' 17:39
moritz_ I'd find it very confusing if my own 'multi sub sum(...)' isn't called on some implementations because there is a builtin already
allbery_b: but deviations from the standard must be declared
allbery_b: like 'use v6_with_sum;' or whatever
allbery_b: extensibility doesn't imply that you may have any extensions in an implementation of the standard 17:40
17:44 justatheory joined 17:58 b_jonas joined 18:11 Schwern left 18:35 Auzon left 18:44 eternaleye joined 18:55 jferrero joined 18:58 [particle]ventus left 19:00 [particle] left 19:05 eternaleye left 19:08 eternaleye joined 19:09 [particle] joined 19:32 jan_ left 19:57 PerlJam joined 20:17 Khisanth left, Khisanth joined 20:19 Alias_ left 20:35 riffraff left, jferrero left 20:37 jan_ joined 20:45 jferrero joined 20:49 BinGOs_ joined 20:54 BinGOs_ left, BinGOs_ joined 20:58 lerog left, lerog joined 20:59 riffraff joined 21:01 BinGOs left 21:06 BinGOs_ is now known as BinGOs 21:11 chris2 left 21:14 eternaleye_ joined 21:16 eternaleye left 21:24 jferrero left 21:33 icwiener_ left 21:44 rindolf left 21:51 pmurias left 22:24 eternaleye_ left, eternaleye_ joined 22:26 b_jonas left 22:38 BinGOs left 22:58 iblechbot left 23:04 Beschwa joined 23:09 elmex left 23:23 Beschwa left 23:25 riffraff left
meppl good night 23:52
23:54 meppl left