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.
00:13 xinming left
meppel-san good night 00:16
00:17 meppel-san left 00:26 Zygo- joined, Zygo left 00:27 Zygo- is now known as Zygo 00:33 Zygo left, Zygo joined 00:36 iblechbot left 00:40 wknight8111 is now known as wknight-away 00:59 nipotaway is now known as nipotan 01:04 lyokato_ joined 01:08 Zygo- joined, Zygo left 01:09 ferrer joined, Zygo- is now known as Zygo
cj is the perl6 sprintf API the same as perl5's? 01:10
01:19 cnhackTNT joined 01:24 Zygo left, Zygo joined 01:28 alester is now known as AndyAway 01:34 mattz left
TimToady cj: probably still there, but mostly just use $x.fmt("%03d") 01:42
01:51 devogon left 01:58 kane__ left, kane_ joined 02:00 kane__ joined, kane_ left 02:01 kane__ left 02:02 Zygo left, kane_ joined 02:04 kane__ joined, kane_ left, whooosh joined 02:05 whooosh left 02:10 demq left, demq joined 02:16 Zygo joined 02:28 Zygo left 02:29 Zygo joined, Zygo is now known as ZygosGhost 02:30 ZygosGhost is now known as Zygo 02:32 Zygo left, Zygo joined 02:33 zakame joined
cj TimToady: okay. I'm tring to add something quick'n'dirty'accurate to svn.pugscode.org/pugs/v6/v6-MiniPer...ntime.pm's Main:: 02:35
lambdabot tinyurl.com/385yzo
cj there was supposed to be another 'n in there 02:36
also, I want to inject values into a template 02:39
*multiple* values 02:40
02:42 zakame_ joined 02:45 zakame left 02:48 zakame__ joined 02:49 zakame__ is now known as zakame 02:54 Zygo left, Zygo joined 02:57 kanru left 03:02 agentzh joined 03:05 zakame_ left 03:07 demerphq joined 03:10 gwern joined
gwern hm. I'm trying pugs from svn - is it bad that HsSyck doesn't build because of bytestring problems? 03:17
and pugs-hsregex too... yeah, that seems to kill my build
is pugs updated for 6.8.x, it occurs to me to ask...? 03:18
03:19 dduncan joined 03:21 demq left 03:25 dduncan left 03:35 [particle1 joined
TimToady gwern: I'm still running 6.6.1 myself, and I don't think it's updated to 6.8 yet 03:40
it could use some TLC on that score...
gwern TimToady: how unforunate. I've got pugs-hsregex updated, but HsSyck looks a bit tricky because bytestring has changed a fair bit 03:41
TLC?
TimToady Tender Loving Care
gwern TimToady: ah. do you know who I should send my 'svn diff's to?
03:42 mattz joined
allbery_b I have ghc-pkg problems building pugs, fwiw. 03:43
gwern allbery_b: how did you get HsSyck working? 03:44
allbery_b I didn't, it fails configure on ghc-pkg errors
as I just said
gwern hm. I assumed you meant pugs-pugs, not the libraries it uses 03:45
allbery_b never gets around to checking if bytestring is the right version, ghc-pkg barfs on its parameters
no, the pugs repo
hrm, no, I misremembered. doesn't error on params, just ignores the global package repo because of its params and then falls over because it can't find base 03:46
gwern heh. my problem is that when it trys to build, it wants old functions like copyCStringLen
03:47 [particle] left 03:53 REPLeffect joined
gwern ah, HsSyck wasn't so hard to fix. 'oleg cat says: see? ur type problem wasn't so hard' 03:59
TimToady wouldn't it just be easier to give you a commit bit to pugs repo? 04:03
diakopter checks if CommitBit is running. 04:04
TimToady it is
gwern dunno. pugs seems to work out of svn and I'm not familiar with svn. I guess when I finish, I'll just send one big diff.
04:04 Limbic_Region left
TimToady if you have a commit bit, you can just send it by saying "svn ci" 04:04
just /msg me your email and I can give you a commit bit 04:05
gwern hm. what is pugs-HsSyck useful for? it says it handles loading and dumping YAML, which is some data serialization format?
TimToady yes
gwern alrigh, but what do you actually do with it? is it good for post-mortem debugging or something? 04:06
TimToady it uses yaml to save compiled images of Prelude.pm and Test.pm, at least
speeds up startup a lot 04:07
commit bit sent
it's customary to add yourself to AUTHORS on first commit
gwern ah. so Prelude.pm and Test.pm need to do a lot of data munging and this is used for caching the end products to disk?
TimToady yes
gwern TimToady: aight
04:08 lyokato_ left 04:11 alc joined 04:12 lyokato_ joined
pugs_svnbot r19953 | lwall++ | [gimme5] can now parse integer 42 running STD under Perl 5 04:12
r19953 | lwall++ | [Cursor5,STD5] now strict and warnings clean 04:13
diff: dev.pugscode.org/changeset/19953
lambdabot Title: Changeset 19953 - Pugs - Trac
gwern neat bot. #haskell needs that for hackage uploads!
diakopter the bot runs in pugs. Although a much simpler one could be made to check the rss feeds from the trac (or SVN::Web) site. 04:15
gwern man, now I'm confused. hackage has hssyck 0.4, but the one in pugs is 0.2?
04:18 RayMagini left 04:20 wknight-away left 04:21 RayMagini joined 04:24 zakame_ joined
diakopter [OT] just discovered the VPSes at serveraxis.com. Half the price of what I'm currently paying, and they're Xen, not VZ. 04:27
Tene [OT] I can voich that www.neverblock.com/ has decent xen hosting, but I have no clue what the pricing looks like. 04:29
lambdabot Title: Com.org - Only the best links ...
04:30 jjore_ joined
diakopter lambdabot's DNS lookup got hijacked 04:30
or something
Auzon google.com
lambdabot Title: Google
Tene Huh. Spooky.
diakopter www.endspa 04:31
lambdabot: www.endspamemails.com/
ok...
Auzon Maybe it parsed it incorrectly. 04:32
diakopter re: neverblock I need 8x more RAM than their VPSes provide 04:33
04:36 pcbuetow joined 04:39 zakame__ joined 04:44 zakame left, zakame__ left 04:45 zakame__ joined, zakame__ is now known as zakame 04:46 zakame_ left 04:49 pbuetow left 04:52 zakame_ joined 04:56 kst left
cj diakopter: how does ec2 compare? 04:58
buu serveraxis.com. 05:02
www.neverblock.com/
Uselss bot. 05:03
lambdabot Title: NeverBlock Reliable Virtual Private Servers
Auzon hah
05:04 zakame left, REPLeffect left
gwern in case anyone was wondering, I've made it up to 56 of 109 in the build process thus far 05:21
05:21 xinming joined, pcbuetow left 05:22 lyokato_ left
gwern but I think I will call it a night here 05:23
Tene g'night 05:25
05:35 penk joined 05:41 Zygo left, Zygo joined 05:46 monomorph joined 05:54 njbartlett left 05:55 monomorph left 06:05 lyokato_ joined
spinclad cj: re inject *multiple* value: C< say "blah {$x.fmt('%a')} blah {$y.fmt('%b')} blah"; > may improve on sprintf style in some ways (keeping value, injection point, and format together) 06:23
06:24 zakame joined 06:25 zakame left 06:33 thoughtpolice left 06:40 zakame_ left 06:41 penk left 06:44 zakame joined 06:54 BinGOs left 06:57 fmerges joined 07:05 devogon joined 07:15 BinGOs joined 07:26 Daveman left 07:37 Aankhen`` joined 07:41 Daveman joined 07:51 iblechbot joined 07:55 stevan_ left 07:58 jql left 08:05 njbartlett joined 08:06 penk joined 08:22 fmerges left 08:23 njbartlett_ joined 08:26 fmerges joined 08:33 njbartlett__ joined 08:36 njbartlett left 08:50 njbartlett_ left 08:51 zamolxes joined 08:52 GPenguin joined, Gwern-away joined 08:58 njbartlett__ left 09:07 pravus left, pravus joined, sri_work joined 09:11 hanekomu joined 09:27 cognominal_ left 09:28 cognominal_ joined 09:29 cognominal_ left, cognominal_ joined
pugs_svnbot r19954 | lwall++ | [gimme5] some utf8 coercions to get around yaml yuck 09:38
diff: dev.pugscode.org/changeset/19954
lambdabot Title: Changeset 19954 - Pugs - Trac
09:48 alc left 09:57 cognominal_ left, gwern left 10:05 penk left 10:17 penk joined
ruoso TimToady, have you seen my questions about STM yesterday? 10:21
10:22 cnhackTNT left 10:27 penk left 10:32 nipotan is now known as nipotaway 10:33 Daveman left 10:39 chris2 joined 10:41 masak joined 10:48 zakame left 10:51 wknight-away joined 10:53 Zygo left 10:54 Zygo joined 11:01 ebassi joined 11:05 IllvilJa left 11:11 GPenguin left 11:25 IllvilJa joined 11:29 dcnl joined 11:34 cmarcelo joined 11:35 dcnl left
ruoso After a little more thinking about STM... I think it would be sane for it to be Object-based. Which means that only STM-aware objects are going to pay attention to the transaction... but I don't think side-effects should be tracked... 11:35
the user must know that the code in a contend block is subject to retry, and that only STM-aware objects will be rolled back 11:36
there's actually no way of keeping track of side effects... considering we can have unknown representations 11:38
that can dispatch the method calls however they want... 11:39
if we think this way... we can have a simple "autocommit" property on the current interpreter instance, where a STM-aware object would check and implement the transaction, knowing that the current interpreter is the transaction holder, and probably registering in the interpreter as commit-listener... 11:42
but there's still the issue of making the commit atomic... 11:43
probably implementing a distributed-transaction-like api could solve it
11:46 Helios_ joined
ruoso TimToady, can we have contend inside contends? 11:47
11:47 Helios left, Helios_ is now known as Helios
ruoso if so, the interpreter would have to hold a "transaction" reference... that would be undef most of the time, and at that time, the objects would always auto-commit... 11:51
and this transaction object would implement a distributed-transaction api where each object would hold its own transaction...
and this same api would implement the creation of a sub-transaction... 11:55
but for all this, keeping track of side effects is still impossible
in fact... keeping track of side effects is actually a bad idea... 11:59
considering one STM-aware object might be an ORM objcet
which means that it might actually be using IO to implement the transaction 12:00
which means that there is no actual STM, but simply Distributed Transactions 12:04
where the interpreter is simply the Distributed Transaction Coordinator
12:04 lyokato_ left
ruoso it may actually be more impressive to think about Perl 6 implementing the XA spec than using STM... specially for enterprise software development 12:08
think about web-service objects that transparently integrate the transaction... where the perl code simply looks like a normal set of calls... 12:11
12:12 wknight-away left 12:16 cognominal_ joined
pugs_svnbot r19955 | putter++ | [misc/STD_red] progress. std.rb is now a careless mostly-non-working Match-free version of STD.pm in ruby. next step is s/careless//. 12:21
diff: dev.pugscode.org/changeset/19955
lambdabot Title: Changeset 19955 - Pugs - Trac
mncharity gwern: thanks for your work! :) 12:26
12:34 chris2 left, njbartlett joined 12:40 njbartlett_ joined 12:47 mncharity left 12:56 hanekomu_ joined 13:02 njbartlett left 13:10 njbartlett_ left 13:11 cognominal_ left 13:12 hanekomu left 13:13 cognominal_ joined 13:15 penk joined 13:24 cognominal__ joined 13:32 cognominal_ left 13:36 bphillips left, orafu joined 13:39 baest left 13:40 baest joined 13:44 pbuetow joined 13:54 monomorph joined
wolverian Juerd, I really like perluniadvice. do you have any thoughts on alternative XML processors, e.g. XML::LibXML; how does it handle unicode? 13:57
14:14 baest left 14:15 baest joined
masak why does a Match object have an "underlying result object"? 14:16
14:16 stevan_ joined
masak why aren't the two the same? 14:16
14:20 marmic left
spinclad masak, so it can still have positional and named submatches, no? 14:20
masak the match? hm. 14:22
I guess I'm just philosophically confused about the difference between a Match and a match result
spinclad at least that's how i read the difference: the result object is the match's scalar value
masak ah
ok, that makes sense
also, what exactly is a 'zen slice'? 14:23
is it used outside of regexes?
where is it spec:ed? 14:24
spinclad mentioned in S05:2164 and S09:596 14:25
masak spinclad: thx. I saw the mention in S05 and didn't really understand 14:26
14:26 kanru joined
masak but I'll go read the S09 mention now and maybe it'll get better 14:26
spinclad a 'zen slice' is an empty slice: @x[], meaning @x[*] (all elements) 14:27
masak hm, yes
spinclad sorry, not === @x[*], but only the currently defined indices. 14:29
masak yes
would it be reasonable/desirable to disallow certain deep indexings of $/ based on the structure of the regex? 14:30
14:31 marmic joined
masak e.g. $4 is disallowed if there is no fifth group 14:31
and $0[0][0] is disallowed if there is no triple nesting in the first group
spinclad (ick. still haven't said it right. anyone listening, just see S09 for it.)
masak spinclad: I thought that was it. but I will read S09 carefully 14:32
spinclad s/defined/allocated/
14:33 fmerges left
masak right. 14:33
spinclad re disallow, that sounds right.
masak sounds like a natural consequence of regexes becoming more declarative 14:34
14:34 Auzon left
cj spinclad: is this implemented in mp6? 14:38
spinclad: also, is there an sprintf in p6?
spinclad re sprintf: there will be i expect, though not specced yet. 14:40
:perl6 my $whom = 'world'; say "hello {$x.fmt(%s)}" 14:44
perl6: my $whom = 'world'; say "hello {$x.fmt(%s)}"
exp_evalbot kp6: OUTPUT[hello {.fmt(%s)}␤]
..pugs: OUTPUT[*** ␤ Unexpected "."␤ expecting "::"␤ Variable "$x" requires predeclaration or explicit package name␤ at /tmp/d9CBww4ueO line 1, column 35␤]
..rakudo: OUTPUT[No scope found for PAST::Var '$x'␤current instr.: 'parrot;PCT::HLLCompiler;panic' pc 158 (src/PCT/HLLCompiler.pir:103)␤called from Sub 'parrot;PAST::Compiler;post_children' pc 1117 (src/PAST/Compiler.pir:138)␤called from Sub 'parrot;PAST::Compiler;pirop' pc 1906
..(src/PAST/Compiler.pir:423)␤called from Sub 'parrot;PAST::Compiler;post_child...
cj panic!!!
spinclad re implemented in mp6, you mean .fmt? probably not. i would not count on any existing implementation (pugs, kp6, rakudo...) being complete, particularly in details like this.
(perl6: asks all three) 14:45
cj spinclad: I expected as much. If there is going to be a perl6 sprintf, I would like to add it to mp6 so that kp6's perl5 emitter work can be generalized
TimToady.push("spec sprintf") 14:46
spinclad perl6: my $whom = 'world'; say "hello {$whom.fmt(%s)}" # third try
exp_evalbot kp6: OUTPUT[substr outside of string at compiled/perl5-kp6-mp6/lib/KindaPerl6/Runtime/Perl5/MP6Runtime.pm line 98, <> line 1.␤substr outside of string at compiled/perl5-kp6-mp6/lib/KindaPerl6/Runtime/Perl5/MP6Runtime.pm line 98, <> line 1.␤substr outside of string at
..compiled/perl5-kp6-mp6/lib/KindaPerl6/Runtime/Perl5/MP6Runtime.pm line 98, <> line 1...
..pugs: OUTPUT[*** ␤ Unexpected ")}\""␤ expecting "::"␤ Variable "%s" requires predeclaration or explicit package name␤ at /tmp/8woZAedi8E line 1, column 45␤]
..rakudo: OUTPUT[hello {world.fmt(%s)}␤]
cj you screwed up my irssi :) 14:47
spinclad so, none of them. some don't even do "string {expr}" interpolation
Juerd wolverian: I have very limited experience with XML processors. I recall that XML::LibXML does it pretty well, but if I were sure it'd be on perluniadvice :)
spinclad cj: sorry. i see now i can /msg exp_evalbot for private queries. 14:49
cj spinclad: :)
14:51 rdice joined
spinclad pugs actually wins 14:52
pugs: my $x = 'world'; say "hello {$x.fmt('%s')}"
exp_evalbot OUTPUT[hello world␤]
spinclad the others miss both {...} and $x.method interpolation 14:54
14:54 Lorn joined
spinclad cj: better to leave mp6 fixed i fancy. once kp6 bootstraps (come the day), it would make sense enough to add it there. 14:56
though if it turns out to be easy...
cj -- Macro: AC_INIT (PACKAGE, VERSION, [BUG-REPORT], [TARNAME]) 14:57
Process any command-line arguments and perform various
initializations and verifications.
eep. wrong channel.. sorry
14:58 mattz left
spinclad i think the intent of mp6 is to be strictly low-level: no interpolation at all; if you want to interpolate, you can paste the string together with ~ yourself. 14:59
15:02 mattz joined
cj spinclad: that's what they're doing now... it seems somewhat painful, but I guess it will work :) 15:06
cj makes a patch
15:06 hanekomu_ left 15:10 AndyAway is now known as alester
cj if I wanted to create a class member (say, $newline), how would I a) declare it and b) reference it from a class method? 15:13
15:20 Daveman joined 15:23 GPenguin joined 15:28 bphillips joined 15:29 cognominal_ joined 15:30 cognominal__ left
masak in the following regex 15:41
mm/ mv $0=<file>+ /
what part or feature matches the spaces between files?
[particle1 from STD.pm, token quote:mm { <sym> <quotesnabber(':regex', ':s')> } 15:42
so, mm// turns on :regex and :sigspace 15:43
masak (hm, maybe none, because the subsequent example in S05 is m/ mv \s+ $<from>=(\S+ \s+)* /)
yes
15:43 [particle1 is now known as [particle]
masak does that mean that spaces between <file>s are allowed? 15:43
15:43 fridim_ joined, cmarcelo left
masak in that case, I've underestimated :sigspace 15:44
[particle] i think so, but i'd need to test it to be sure
pmichaud: ping
masak if t'were mm/ mv <file> <file> /, I would not hesitate
15:44 njbartlett joined, cmarcelo joined
masak but it's mm/ mv <file>+ / 15:45
[particle] right
pmichaud pong 15:48
$0=<file>+ doesn't match spaces between files (assuming that <file> doesn't eat spaces) 15:49
[particle] ok, so spec bug
masak that's what I thought
pmichaud where is that particular instance?
[particle] should be [<file><.ws>]+
masak S05:2870 15:50
pmichaud ah, in S05. I thought we had already fixed those.
[particle] S05:2648
would [<file> ]+ work? 15:51
probably not
pmichaud sure, that works
masak :)
[particle] wow. dwim++
pmichaud although I think I'd write it as [ <file>]+ instead 15:55
for clarity it should probably be [ <file> ]+ 15:56
masak would that guarantee that no space is required after the last <file>? 15:57
pmichaud well, the normal meaning of :sigspace is "optional whitespace between tokens"
where it becomes "required whitespace between word tokens"
15:59 orafu left
pmichaud anyway, [ <file> ]+ would capture one or more <file>s separated by spaces. spaces would only be required between pairs of <files> that end/begin with a word character 15:59
[particle] masak++ for spec bug hunting 16:00
pmichaud believe it or not, in this particular case I believe it was proposed that we keep <file>+ even though it's technically incorrect 16:01
I'm checking my design team notes
masak [particle]: I'm trying extra hard to internalize p6 rules right now 16:02
pmichaud: I don't see what would be the point of keeping a regex which has been proven non-useful, even if it's just an example 16:03
[particle] without a note about it, at least.
pmichaud I agree -- that's why I'm re-checking my notes about it
16:09 njbartlett left 16:10 masak left
pmichaud okay, found the thread, now to re-read it. 16:13
16:14 cognominal_ left 16:16 fridim_ left
pugs_svnbot r19956 | gwern++ | This patch updates pugs-hsregex; it should work with both 6.6 and 6.8.2. However, backwards compatibility requires Cabal versions equal to or greater than 1.2 (so you can use the conditional syntax to change build-dependencies on base). 16:16
diff: dev.pugscode.org/changeset/19956
lambdabot Title: Changeset 19956 - Pugs - Trac
pugs_svnbot r19957 | gwern++ | Update HsSyck to merge in the Haddock documentation and the Cabal sdist improvements which constituted 0.3 and 0.4. This version also updates hsSyck to work with modern ByteStrings, bytestring>=0.9.0.1, which is the oldest one packaged with GHC 6.8.x. (You can get this and hsregex on Hackage as well.) 16:19
diff: dev.pugscode.org/changeset/19957
lambdabot Title: Changeset 19957 - Pugs - Trac
pmichaud hmmm, I can't find where we had that discussion in my mail archives. Perhaps it was in a conference call, then. 16:24
afk, lunch & errands 16:28
16:31 jjore_ left 16:47 macae joined 16:52 kyrbe joined 16:53 pbuetow left, pbuetow joined 16:54 pbuetow left, pbuetow joined 16:56 DarkWolf84 joined 16:59 kst joined 17:03 kyrbe left 17:05 cognominal_ joined 17:11 pcbuetow joined, pcbuetow left 17:12 pbuetow left, pbuetow joined 17:13 pbuetow left 17:16 pbuetow joined 17:17 Patterner left, Psyche^ joined, Psyche^ is now known as Patterner
pugs_svnbot r19958 | gwern++ | Update HsJudy's bytestring call, also update for split-base. As before, requires a more recent Cabal so can use conditionals, and as before, this package is now also on Hackage. While I was at it, I did some -Wall cleanup 17:25
diff: dev.pugscode.org/changeset/19958
17:26 pbuetow left
lambdabot Title: Changeset 19958 - Pugs - Trac 17:26
17:27 pbuetow joined 17:34 meppl joined 17:36 testing joined 17:37 testing left 17:38 testing_ joined, testing_ left 17:42 ferrer left, cathya joined 17:51 GPenguin left, blindfish joined
pugs_svnbot r19959 | gwern++ | Cleanup Str.hs and Types.hs, and add necessary bytestring updates 17:53
diff: dev.pugscode.org/changeset/19959
lambdabot Title: Changeset 19959 - Pugs - Trac
17:56 njbartlett joined
pugs_svnbot r19960 | gwern++ | Cleanup UTF8.hs for ByteString update 17:56
diff: dev.pugscode.org/changeset/19960
lambdabot Title: Changeset 19960 - Pugs - Trac
17:57 njbartlett left
pugs_svnbot r19961 | gwern++ | Update GHC 6.8.1 notes with my observations; 17:57
r19961 | gwern++ | Add myself to AUTHORS;
r19961 | gwern++ | Add my favorite book series to READTHEM;
r19961 | gwern++ | Revamp the Pugs.cabal.in. This includes formatting, additional metadata, and adds a dependency on a more recent bytestring (to work with all the preceding patches), and also a Cabal, because we need that oldtime section syntax... for compatibility with GHC 6.6 & 6.8
r19961 | gwern++ | Remove a regexp from util/build_pugs.pl because we can't have backwards compatbility with that old Cabals if we want 6.8.x as well, so it merely messes up the generated Cabal file.
diff: dev.pugscode.org/changeset/19961
lambdabot Title: Changeset 19961 - Pugs - Trac
18:00 mncharity joined 18:13 justatheory joined
mncharity re sprintf, pugs has it. it's in src/perl6/Prelude.pm. 18:23
written a very long time ago to make some pugs t/ pass.
wow, gwern++ 18:24
18:24 meppel-san joined
Gwern-away don't count your chickens until they hatch; dunno how well my patches work for other peoplpe 18:25
mncharity pugs> sprintf("%03d",4)
Gwern-away: regardless, it's nice to see. :) 18:26
6: sprintf("%03d",3)
(that last is ./misc/pX/Common/redsix/redsix -e 'sprintf("%03d",4)' ) 18:27
re sprintf being not spec'ed, wouldn't it fall under the "everything is the same as p5, unless it isn't" default?
18:31 meppl left 18:32 meppel-san is now known as meppl
spinclad pugs: sprintf ('%03d', 4) 18:33
exp_evalbot OUTPUT[*** Insufficient arguments to sprintf␤ at Prelude.pm line 668, column 13-53␤]
spinclad pugs: sprintf('%03d', 4)
exp_evalbot RESULT[\"004"]
rakudo_svn r25904 | particle++ | [rakudo] bringing regex/token/rule support closer to correct -- everything still acts like regex 18:36
spinclad re 'same as p5', yes... though we're more or less planning to move away from that dependency, in the long run.
18:38 njbartlett joined 18:39 njbartlett left 18:40 bphillips left
spinclad s/more or less// 18:40
mncharity :)
spinclad gwern++ # i'll give it a spin after lunch. (still at ghc 6.6.1 this week) 18:42
mncharity contemplates @fate 18:43
spinclad goes off to eat & 18:44
TimToady currently @fate is slightly oversimplified
spinclad lets & contemplate its own @fate
ruoso disturbed after reading about message-passing-concurrency
TimToady, have you seen my comments on STM? 18:45
TimToady yes
ruoso so?
TimToady I just wanted you to think about it, is all :)
ruoso but I mean
TimToady and you did!
ruoso is the idea of not tracking side-effects at all acceptable? 18:46
TimToady I don't see it in the spec anywhere :)
ruoso :)
TimToady I think in the two layer model the upper layer is more STMish and the bottom layer more Erlangish 18:47
you've seen the paper on that?
18:47 rdice_ joined, rdice_ left
ruoso I've seen Joe Armstrong thesis (chapter 2 only, actually) 18:48
not thesis, dissertation
18:48 rdice left, ebassi left
pmichaud TimToady: (from scrollback) there was discussion earlier about mm/ mv <file>+ <file> / in S05 18:48
in particular, the notion that <file>+ doesn't include any separators 18:49
ruoso TimToady, two layer model?
pmichaud should we fix S05 to read mm/ mv [ <file> ]+ <file> / ?
ruoso have his head spinning... so ideas are not getting out in the correct order...
TimToady pmichaud: no, I just put in a note about assuming <file> is smart 18:50
pmichaud okay, that works then. thanks.
I agree that having <file>+ makes more sense from an illustrative perspective
TimToady also decided ** \s+ was okay 18:51
rather than disallowing quantifiers on the right side
ruoso TimToady, what paper?
TimToady I'm trying to remember where it is.
it's a Haskell paper about running transactions on top of a monad that runs an event engine underneath 18:52
ruoso ah... that one I certainly didn't see...
TimToady something you should look at, if we can dig it up 18:53
pmichaud From S05: "Assume that our file subrule requires a real boundary at that point using..." Actually, in order for the <file>**{2} and <file>+ examples to work later, the file subrule needs to eat the separators as well, yes?
ruoso I was thinking, while reading armstrong's dissertation, about how we could mix the erlang's view on the idea of message queue for each object and an OO interpreter view 18:54
TimToady oh, I was gonna change those to <file> ** 2 and forgot 18:55
<.ws> will certainly eat the whitespace as long as <file> guarantees its there
*it's
this is my brain on flu, any questions... :/ 18:56
pmichaud right. perhaps cleaner is to say that "assume that our C<file> subrule handles any separators between elements"
flu--
ruoso but in the end I realised this is a completely different question... and contend also relates to groups of messages... a problem I don't know if message-passing-concurrency solves 18:57
pmichaud the flu around here cost me two weeks (4 days taking care of sick child, 2 days taking care of myself, six days with sick wife in hospital)
ruoso but I'm pretty sure that it's very unlikely that we can have haskell-like stm in Perl 6, considering the support for different representations...
pmichaud time for another errand. I may be a few minutes late to today's phone call 18:59
TimToady www.seas.upenn.edu/~lipeng/homepage/unify.html
lambdabot Title: Unifying events and threads
19:00 zamolxes left
ruoso will print and read in the way home 19:00
TimToady it's not the paper, but it has a link
19:00 Psyche^ joined
ruoso yeah... i got it 19:01
but I think I'm pretty happy with the native-XA idea
instead of real STM
TimToady what is XA? 19:02
ruoso distributed transaction x/open api
TimToady link? 19:03
ruoso en.wikipedia.org/wiki/X/Open_XA
lambdabot Title: X/Open XA - Wikipedia, the free encyclopedia
ruoso but the basic idea is, since we can't presume to know the representation of the objects to implement STM, 19:04
we implement distributed transactions coordinated by the current interpreter
using an explicit API
so the objects that are XA-aware will be rolled-back or commited... 19:05
while everything else will simply be repeated
so every XA-aware object is understood as a resource manager 19:07
the contend block is the Application Program
anyway... 19:08
need to go...
ruoso will backlog, as usual...
19:17 Patterner left, Psyche^ is now known as Patterner 19:19 Aankh|Clone joined
pugs_svnbot r19962 | gwern++ | Move all uses of pugs-HsSyck to simply HsSyck; it's a package useful for more than Pugs, and this makes it easier to synch with Hackage. 19:19
diff: dev.pugscode.org/changeset/19962
lambdabot Title: Changeset 19962 - Pugs - Trac
19:21 rdice joined 19:22 Lunchy left, Lunchy joined 19:29 cognominal_ left 19:35 fridim_ joined, fridim__ joined, fridim_ left 19:36 testing joined, Aankhen`` left 19:37 GPenguin joined 19:40 Aankh|Clone left 19:56 macae left
testing I'm looking for some help on bitwise operations of strings 20:07
is this a good channel for getting help? 20:09
[particle] yes, if you're patient. have you looked at the tests? 20:10
PerlJam testing: and just to check ... this is #perl6, so we're assuming you're talking about perl6 bitwise ops :)
testing PerlJam: thank you, I just checked and I am using perl5.8, I will move to #perl 20:13
TimToady S02 mentions "identifier extensions" in the section describing adverbial 20:21
20:22 jhorwitz joined
TimToady oops, that was a paste error... 20:25
20:26 testing left 20:38 fakename joined 20:46 fakename left
Tene So, is there a way to operate on an entire array without using all(), any(), none(), etc? 20:53
Similar to a junction, but preserves both order and duplicates? 20:54
[particle] for() ?
Tene I'm hoping to have something I can just run function($a,$b)
Which works fine with junctions... 20:55
But fails when there are duplicate results.
For some definition of "fails"
I can do it manually using nested for loops. 20:56
I guess I could define an infix operator and use » and « 20:57
pmichaud map? 21:05
21:16 ron___ joined
pugs_svnbot r19963 | lwall++ | [STD] put regex into its own subgrammar 21:17
r19963 | lwall++ | eliminate assertstopper kludge in favor of virtual stdstopper
r19963 | lwall++ | make ** separator quantifiable
r19963 | lwall++ | override ws to match only if insignificant
r19963 | lwall++ | significant whitespace is now a metacharacter
diff: dev.pugscode.org/changeset/19963
lambdabot Title: Changeset 19963 - Pugs - Trac
ron___ perl6: say "hello" 21:21
exp_evalbot kp6: OUTPUT[hello␤]
..pugs: OUTPUT[hello␤]
..rakudo: OUTPUT[hello␤]
21:22 chris2 joined
ron___ perl6: sub fizzle ($must_fizzle, :$may_fuzzle) { say "definately time to $must_fizzle"; if ($may_fuzzle) { say "ok - also $may_fuzzle" } else { say "hey - no fuzzling" } } fizzle 'drizzle'; fizzle 'drizzle', 'dazzle'; fizzle 'druzzle', 'dazzle', 'chowder'; 21:22
exp_evalbot kp6: OUTPUT[no method 'APPLY' in Class 'Undef'␤ at compiled/perl5-kp6-mp6/lib/KindaPerl6/Runtime/Perl5/MOP.pm line 345␤ KindaPerl6::Runtime::Perl5::MOP::__ANON__('HASH(0x824cabc)', 'APPLY', 'HASH(0x8e54160)') called at compiled/perl5-kp6-mp6/lib/KindaPerl6/Runtime/Perl5/MOP.pm line 169␤
..main::DISPATCH('HASH(0x824cabc)', 'APPLY', 'HASH(0x8e54160)') ca...
..pugs: OUTPUT[definately time to drizzle␤hey - no fuzzling␤definately time to drizzle␤hey - no fuzzling␤*** No compatible multi variant found: "&fizzle"␤ at /tmp/NHLfTg2eq1 line 1, column 212-249␤]
..rakudo: OUTPUT[Statement not terminated properly at line 1, near "fizzle 'dr"␤current instr.: 'parrot;PGE::Util;die' pc 120 (runtime/parrot/library/PGE/Util.pir:82)␤called from Sub 'parrot;Perl6::Grammar;statementlist' pc 15013 (src/gen_grammar.pir:2402)␤called from Sub
..'parrot;Perl6::Grammar;statement_block' pc 12200 (src/gen_grammar.pir:1396)␤called f...
21:23 blindfish left
pmichaud need a statement terminator after the sub's closing } 21:23
ron___ Am a bit of a novice. Anyone: in pugs how come the second call with two parms is accepted but the third with three is rejected ? 21:24
pmichaud because the sub only accepts two args? 21:25
the sub is defined as having one required, one optional (named) argument. 21:27
ron___ sorry - confused by named args but think I understand now.
perl6: sub fizzle ($must_fizzle) { say "definately time to $must_fizzle"; } fizzle 'drizzle'; fizzle 'drizzle', 'dazzle', 'chowder'; 21:35
exp_evalbot kp6: OUTPUT[no method 'APPLY' in Class 'Undef'␤ at compiled/perl5-kp6-mp6/lib/KindaPerl6/Runtime/Perl5/MOP.pm line 345␤ KindaPerl6::Runtime::Perl5::MOP::__ANON__('HASH(0x824ca44)', 'APPLY', 'HASH(0x8e54af4)') called at compiled/perl5-kp6-mp6/lib/KindaPerl6/Runtime/Perl5/MOP.pm line 169␤
..main::DISPATCH('HASH(0x824ca44)', 'APPLY', 'HASH(0x8e54af4)') ca...
..pugs: OUTPUT[definately time to drizzle␤definately time to drizzle␤]
..rakudo: OUTPUT[Statement not terminated properly at line 1, near "fizzle 'dr"␤current instr.: 'parrot;PGE::Util;die' pc 120 (runtime/parrot/library/PGE/Util.pir:82)␤called from Sub 'parrot;Perl6::Grammar;statementlist' pc 15013 (src/gen_grammar.pir:2402)␤called from Sub
..'parrot;Perl6::Grammar;statement_block' pc 12200 (src/gen_grammar.pir:1396)␤called f...
TimToady note that anyone that runs that successfully is not up to spec
[particle] ron___: try "pugs:" for less noise 21:36
TimToady as pmichaud said, the final } is required to have a ; if you're going to continue on the same line
pugs is not correct to assume the ; there
ron___ That only takes one parameter but works with three.
pugs: sub fizzle ($must_fizzle) { say "definately time to $must_fizzle"; } ; fizzle 'drizzle'; fizzle 'drizzle', 'dazzle', 'chowder'; 21:37
exp_evalbot OUTPUT[definately time to drizzle␤definately time to drizzle␤]
TimToady that should also not be packing three args into one according to recent spec 21:38
[particle] pugsbugs
ron___ Am a novice - should I take further action?
TimToady well, that only changed last week, up till then sub foo ($x) made a unary function
now it's just a list operator and fails to bind 3 args to one parameter 21:39
but nobody implements that yet
21:42 ron___ left 21:48 fakename joined 21:57 ron___ joined 22:00 jhorwitz left 22:01 ron___ left 22:07 fakename left 22:17 BinGOs left 22:18 gbacon left 22:20 Schwern joined 22:23 IllvilJa left 22:31 Limbic_Region joined 22:32 wknight-away joined 22:41 cmarcelo left 22:43 cathya left 22:56 alester_ joined 23:00 gwern joined 23:02 njbartlett joined, BinGOs joined 23:03 njbartlett left 23:05 Gwern-away left, Psyche^ joined 23:07 fridim__ left 23:09 fakename joined 23:11 rdice left 23:13 njbartlett joined 23:20 nipotaway is now known as nipotan 23:21 Patterner left, Psyche^ is now known as Patterner 23:22 wknight-away left 23:23 sri_work left 23:28 GPenguin left, fakename left 23:40 wknight-away joined 23:54 alester_ left, monomorph left, Auzon joined 23:58 Jedai joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!