svk sy -s 12103 //mirror/pugs | 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 11 August 2006.
dduncan success! 00:18
thank you cmarcelo, Pugs builds now
for me 00:19
cmarcelo dduncan: nice =) 00:20
diakopter trying on macintel...
Patterner the compile time for pugs on an old powerbook is three days (with interruptions) by now... 00:21
diakopter optimized or un 00:22
00:22 nekokak joined
Patterner guess :) 00:22
diakopter optimized
Patterner 10 points for diakopter. 0 points for the UK. 00:23
diakopter lots of warning while making Judy
Patterner no wait. wrong directory. it's ghc. 00:24
diakopter 6 min. to compile on macintel, unoptimized. 00:25
00:26 rindolf joined
diakopter anyone around who feels like handing out a task (for me to do)? 00:27
Patterner macintel != old pb 00:29
(i'll upgrade when they stop exploding)
diakopter 's has stopped exploding
Patterner good for you :) 00:30
diakopter ok, no assigners around... any suggesters, even? any ideas on what I could work on? I can do perl5 and I've read the Apocalypses, Synopses, Exegeses, Apocrypha, and some Haskell tutorials. 00:33
I'm somewhat familiar with the source tree 00:34
Patterner translate the docs to klingon 00:35
diakopter - isn't there a p5 module that does that?
audreyt diakopter: there's a TASKS file :) 00:55
diakopter sort of
audreyt - good morning 00:56
audreyt hi
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 00:58
audreyt diakopter: are you interested in coding up runpugs? :) 01:01
I mean trypugs
diakopter yes but I have a Module::Build question for you before I get started on something
audreyt er, ok, but I know very little about M::B
diakopter this isn't you? www.perlfoundation.org/gc/grants/20...rijus.html 01:02
lambdabot Title: 2003 Grants - Autrijus Tang - The Perl Foundation
diakopter oh, I see, mostly not done by you on that part; nm 01:03
audreyt yeah. I reanimated MakeMaker with Module::Install instead of joining the shiny Module::Build crowd... 01:06
diakopter yes. how far did cwest get with the Jifty implementation... is it in pugs svn? 01:08
audreyt no. I think it fizzled somehow
diakopter you're about done with Jifty, right?
or are there lots of ambitions still to be realized 01:09
I read the ppt about it
audreyt I think a generic line-based process harness to the web may be the right answer
but I can't quite find something like that on CPAN 01:10
diakopter are you planning on running this on feather?
audreyt diakopter: Jifty is my day $job :) and I need to run in ~1hr to $client side and hack more jidty
diakopter: yes, perhaps in a chrooted quota'ed place. or I can run it in a freebsd jail somewhere, like pugscode.org
diakopter my Module::Build problems appeared when I was installing SVK in order to host a SVN::Web view of pugscode diffs 01:12
mostly for my own benefit. I learn well from reading diffs
the client side is trivial, obviously. I could do a super-quick implementation in WebGUI, since that's the web app framework with which I'm familiar 01:14
super-quick meaning compared to jifty (for me) 01:15
audreyt oh cool
# rt.openfoundry.org/Foundry/Project/.../pugs/log/
lambdabot Title: OpenFoundry(č‡Ŗē”±č»Ÿé«”é‘„ōæ½xE9ōæ½x80 ōæ½xE5 ōæ½xB4-å°ˆę”ˆé–‹ē™¼å¹³å°) -- OSSF
audreyt sure, whatever works :)
01:19 rlb3 joined
svnbot6 r12235 | markstos++ | ChangeLog updates 01:23
audreyt (bbiab) 01:26
svnbot6 r12236 | lwall++ | s:g/FIRST/START/ (Note, didn't change firstBlock storage, since we still 01:44
r12236 | lwall++ | have a FIRST. START doesn't need blocktrait storage since it's inline code.)
TreyHarris "may occur only at the top level of such a loop block". so that means you can put it immediately within the outermost braces of the loop you want it to refer to, but it doesn't mean you can't have a nested loop and control that with another NEXT, right? 01:54
01:54 diakopter joined
TreyHarris So... for @a { if $x { NEXT { ... } } } is invalid, but for @a { for @b { NEXT { ... } } } is valid, with the NEXT block referring to the for @b? 01:55
01:56 JarJarBinks joined
TreyHarris &'s to uncommute 02:00
dduncan fyi, I've just started my first smoke since 6.2.12 was released ... 02:04
or at least, thought I did... 02:05
cmarcelo et al, it seems the build process is still borked 02:06
cmarcelo dduncan: what's happening?
dduncan while Pugs built fine, subsequently trying to say 'make smoke' caused some sort of make to recheck itself, and that deleted some files 02:07
eg, it now says: /usr/bin/ld: table of contents for archive: third-party/installed/libJudy.a is out of date; rerun ranlib(1) (can't load from it)
also, the pugs executable disappeared
cmarcelo err.. =P still have the entire output to paste? 02:08
dduncan and the previously built Pugs did execute too
I only have the output due to make smoke, since the make worked before
but will supply that
cmarcelo fine
pasteling "dduncan" at 24.69.53.198 pasted "'make smoke' causing pugs to be deleted, after successful 'make' with no interim pulls" (263 lines, 18.2K) at sial.org/pbot/18999 02:10
dduncan still, that's rather telling by itself 02:12
good luck figuring it out
cmarcelo i'll try to reproduce it here.. also, I don't know why it always should re-compile (or try to) pugs when making smoke =P 02:14
dduncan: here it goes fine. 02:20
=(
dduncan I have just pulled the last 2 revs and will try again 02:22
but without a prior cleaning
cmarcelo ok. 02:23
02:26 diakopter joined 02:29 SamB is now known as boo, boo is now known as SamB
dduncan okay, so after running 'svk pull' and 'make', the latter ends with this as before: /usr/bin/ld: table of contents for archive: third-party/installed/libJudy.a is out of date; rerun ranlib(1) (can't load from it) 02:30
cmarcelo dduncan, go third-party/judy/Judy-1.0.3 and do a make clean there, then back to pugs root, try make... 02:33
dduncan will do 02:34
cmarcelo dduncan: something? 02:44
dduncan yes, that worked
however, simply running make again will bork it 02:45
by contrast, if there are no source changes since a prior make, then make should have no lasting effect, and that's what it used to do
the way this is behaving is ... 02:46
cmarcelo yep.. we could prepend make-call in util/build_pugs.pl with a make-clean.. but that will recompile judy everytime.. =P
dduncan its like the old copy is always deleted, and then another one remade only conditionally
perhaps the removal and remake should be together in the same conditional? 02:47
cmarcelo dduncan: smoke always started a recompilation of Pugs, which resulted in no file being recompiled (all skipped) but the linking step always happened...
dduncan this is acting like the make process is deleting some interim file at the end that it shouldn't be deleting
and the only way that comes back is a Judy make-clean 02:48
plus make
are you saying this behaviour doesn't happen to you? 02:49
cmarcelo dduncan: lets try something else.. do a make clean in judy then make pugs... this should work... THEN, comment in util/build_pugs.pl the judy stanza and make..
doesn't happen here.. make, then make smoke, goes fine.
02:49 Bit-Man joined
dduncan I will try editing that file a few times to see what changes do or don't make the problem go away 02:55
cmarcelo dduncan: ok. looks like when make(in judy) have nothing to do here, it really doesn't break things and the same files get copied in the right places... this is not the case for you. maybe avoiding doing anything at all when libJudy.a is already in place can solve the problem.. 02:58
[makes sense?]
dduncan I think so
03:00 neoesque joined
cmarcelo wanna try it there and see if works? 03:01
dduncan I still have to figure out exactly what to do
cmarcelo dduncan: do you program in perl? test for installed/lbiJudy.a existence, if its the case, do nothing, else, do our steps including a make clean... 03:02
dduncan yes, give me time here ... 03:03
cmarcelo ok, np
dduncan seems to be working ... I will commit a patch that does this existence test 03:13
cmarcelo dduncan: great! now we must remember this when updating C judy code =o) [need to purge libJudy.a to update it] 03:14
dduncan but I was testing for third-party/judy/Judy-1.0.3/src/obj/.libs/libJudy.a existence 03:15
03:15 mako132_ joined
dduncan because that was the original copy 03:15
cmarcelo dduncan: ok. but even so, if we change a .c in judy tree, it'll be recompiled?
dduncan probably not 03:16
cmarcelo ok. but by now judy C tree won't change much, so commit your patch...
dduncan I just add an if-block around the 5 lines starting at build_pugs.pl line 98
svnbot6 r12237 | Darren_Duncan++ | updated util/build_pugs.pl to only go through the Judy make process if third-party/judy/Judy-1.0.3/src/obj/.libs/libJudy.a doesn't exist ... otherwise, any 'make' following the one that builds Judy will cause a Judy object file to disappear and not be remade, leading to Pugs build failure ... on some systems like mine (OS X) anyway 03:21
03:28 buubot joined
cmarcelo audreyt: i'll let blog activities for tomorrow, after the build process stabilize.. besides doing what i can to unbreak build proc, i'm investigating the Judy.Hash "related" segfaults (is gaal looking onto this too?).. 03:32
(finally..) sleep & 03:33
dduncan turns out I had to install 'Task-Smoke' and 'Best' since last month, but ... 03:44
now that that's done, my smoke is underway
04:38 SSelva joined 04:46 traecer joined, penk joined 05:13 kanru joined 05:18 iblechbot joined 05:27 agentzh joined
agentzh wow, it's very strange that perl 5's "glob 'Judy*/*.o'" doesn't work properly on particle's machine. :/ 05:28
okay, i'll try to tidy up the output while compiling judy on win32 in order to make it less "spectacular". ;-) 05:30
last night's work is mostly a hack. 05:31
*was
gaal re synopsis patch r10944 - what about nested loop blocks? 05:59
05:59 xinming joined
agentzh yes, TimToady's wording is ambiguous there. 06:00
gaal and, what is the definition of "loop" for this purpose? what about user-defined control structures?
svnbot6 r12238 | agentz++ | tidied up the outputs while building Judy on Win32.
r12238 | agentz++ | adjusted build_pugs.pl a bit.
gaal also, having both START and BEGIN will cause, uh, no end to confusion... 06:01
agentzh i'm not sure if r12238 would break unix build. please test.
gaal: *nod*
gaal agentzh: testing.
agentzh gaal++ 06:02
gaal should I clean anything before testing?
agentzh gaal: no. 06:03
06:03 marmic joined
gaal works. 06:03
agentzh incremental make is sufficient.
great
gaal: is it necessary to compile Judy everytime? 06:04
06:04 weinig joined
gaal I don't think so! chase dependencies like 'make' does. 06:04
agentzh the unix part doesn't compile Judy if libJudy.a is already there.
gaal Well, good, if the timestamps are reasonable 06:05
that is if the .a is newer than the .hs etc.
agentzh gaal: currently there's no 'make' logic for win32. :/
writing a Makefile for Win32 Judy will be a lot of work i'm afraid.
gaal TimToady: I know you don't like longish names, but how about CLOSURE_INIT for START? 06:06
agentzh: the .c or the .hs part?
agentzh gaal: i think TimToady dislikes underscores.
gaal I know he does 06:07
agentzh the .c part.
gaal agentzh: there's a Makefile.am there: a windows makefile should be generated by autotools, *if* you're using the gnu toolchain to build... 06:08
another alternative is SETUP {...}
agentzh looking
gaal agentzh: what builds it on w32 now?
agentzh gaal: .bat 06:09
gaal ah.
where?
agentzh a quite long .bat translated by hand from sh scripts last night.
gaal: a sec...
gaal: third-party/judy/Judy-1.0.3/src/hs_build.bat 06:10
transcoding hs_build.bat to nmake's Makefile is not an easy task. :) 06:11
gaal wow, yes, I see
agentzh given the C part of Judy compilation is pretty fast on my machine, i'd say it's okay to rebuild Judy everytime the Win32 user types "nmake" or its families.
it's just a few seconds. 06:12
the bottleneck is other parts of the pugs build system.
scw Could any one help me on the problem stated in the last section of 0rz.net/691Gv ? 06:13
lambdabot Title: Yet Another Pugs Hacker's Blog: August 15th, 2006 -- Last Day of SoC
gaal scw: why not emit the annotation? 06:15
what's the "string" function? 06:16
06:17 weinig joined
gaal oh, it's a parsec builtin. 06:18
scw: what does your parser construct? from the type annotation it looks like it builds a string?
scw gaal: I couldn't emit the annotation since I don't know the type 06:20
svnbot6 r12239 | agentz++ | [third-party/judy/Judy-1.0.3/src/hs_build.bat]
r12239 | agentz++ | removed a huge amount of annoying comments from the source.
scw gaal: It will, as the type tells, return a "parser state" with a string 06:21
06:22 Aankhen`` joined
gaal scw: what's the minimal stuff I need to do to get a sample output from this? 06:27
pasteling "gaal" at 192.115.25.249 pasted "rule2parsec.pl failure" (8 lines, 763B) at sial.org/pbot/19005 06:32
06:34 crem_ joined
gaal scw: ping? 06:35
svnbot6 r12240 | agentz++ | [t/syntax/comments.t] 06:39
r12240 | agentz++ | fixed the "unspace in comment" test to reflect recent changes to S02. TimToady++
06:40 kane-xs joined
dduncan first smoke this month seems to be a success ... 06:40
agentzh dduncan: have you put the smoke report somewhere? 06:41
later & 06:42
06:42 agentzh left
dduncan yes ... its on the smokeserv as usual 06:42
spefically its the: Haskell runcore, repository snapshot / darwin of r12237 that took 136 minutes 06:43
gaal scw: it's not clear to me what LiteralPatch is against. My version of Pugs.Parser.Literal doesn't match that patch 06:44
dduncan so the execution speed seems unchanged from a couple months ago ... not sure if it was supposed to with recent updates or not 06:45
gaal could you give a command line that emits a sample parser please?
scw gaal: I use perl -I../../../perl5/Pugs-Compiler-Rule/lib -I../../../perl5/Pugs-Grammar-MiniPerl6/lib rule2parsec.pl Literal.grammar 07:05
gaal: the patch should be applied to the generated .hs file 07:06
gaal: but according to the error message you got, the Pugs::Grammar::MiniPerl6 part seams not the main reason
gaal: You may need to run 'make' under perl5/Pugs 07:07
gaal: You may need to run 'make' under perl5/Pugs-Compiler-Rule
07:15 SSelva left
gaal scw: ACK. installing prerequisites... 07:23
07:29 baest joined
gaal scw: looks like you need the following declaration in scope: 07:34
instance MonadState RuleState RuleParser where
get = getState
put = setState
(lifted from src/Pugs/Parser/Types.hs)
scw gaal: so why does type annotation work? 07:40
07:40 cmarcelo joined
scw gaal: and, how can an instanciation declaration be exported? 07:42
gaal scw: my bad, you're importing Pugs.Parser.Types already. hmm. 07:46
how do you compile the generated .hs code?
scw: the problem requiring your annotation is that haskell can only infer type up to a paramaterized parser: 07:52
@type option
lambdabot Not in scope: `option'
gaal @type Text.ParserCombinators.ReadP.option
lambdabot forall a. -> Text.ParserCombinators.ReadP.ReadP a -> Text.ParserCombinators.ReadP.ReadP a
gaal you need to supply the "a", otherwise the type of your function is ReadParser a 07:53
as for how *you* infer it's a String... good question :) you either do type inference on the (Perl) rule, or require an annotation there. 07:54
08:01 cm joined 08:08 nekokak_ joined
scw gaal: But doesn't 'option "" ' part make a be [Char] = String ? 08:11
Hmm... however, currently I assume ALL rules without their own production rule have type ReadParser String 08:12
08:20 SageLT joined 08:24 bernhard joined 08:56 drrho joined 08:58 Southen joined 08:59 xinming joined 09:15 dduncan left 09:19 traecer left 09:24 chris2 joined 09:30 Southen joined 09:33 HXanadu joined 09:36 buetow joined 09:53 cyberplaque joined 10:05 cyberplaque left, mauke_ joined 10:10 mauke_ is now known as mauke 10:13 mtve joined
svnbot6 r12241 | audreyt++ | * t/oo/roles/basic.t: Now that roles work in compile time, 10:13
r12241 | audreyt++ | let them take place in compile time.
r12242 | audreyt++ | * Fix build system for GHC 6.5 trunk 2006-07-25 and later, 10:16
r12242 | audreyt++ | which incoporated Simon Marlow's implementation of
r12242 | audreyt++ | "Generalise Package Support", a.k.a. multiversioning.
r12242 | audreyt++ | We now always install Pugs into third-party/installed/,
r12242 | audreyt++ | and link Main.hs against it.
10:22 cmarcelo left 10:34 plisk joined 10:40 baest joined 10:51 evalbot_12389 joined
svnbot6 r12243 | audreyt++ | * agentzh++ noted that "\13\10" should mean simply "1310", 10:55
r12243 | audreyt++ | but that broke Net::IRC, and therefore broke evalbot.
r12243 | audreyt++ | Fixed to say "\d[13,10]" instead.
10:58 evalbot_23394 joined
audreyt I wonder if "\0123" is a warnable offense 10:58
svnbot6 r12244 | audreyt++ | * evalbot-loop: make it svn/svk agnostic
r12245 | audreyt++ | * it's chat.freenode.net not irc.freenode.net
audreyt ?eval "moose"
evalbot_23394 "moose"
11:02 foo\ joined 11:06 agentzh joined
agentzh audreyt: should S02 explicitly outlaw the "\ddd" notation? 11:07
11:07 TimToady joined
audreyt I think so as it departs from perl5 11:07
agentzh what if fire off a mail to p6l? 11:08
audreyt similar to Initial C<0> no longer indicates octal numbers by itself. You must use
agentzh oh, TimToady is here.
11:08 iblechbot joined
agentzh audreyt: *nod* 11:08
i must use the 0o prefix. 11:09
agentzh has sent the mail to p6l. 11:16
11:17 cmarcelo joined 11:22 diakopter joined
agentzh ?eval $x = do 3+2; 11:22
11:22 evalbot_23394 is now known as evalbot_23396
evalbot_23396 Error: Undeclared variable: "$x" 11:22
agentzh ?eval my $x = do 3+2; 11:23
evalbot_23396 \5
agentzh audreyt: is "do EXPR" a valid expression?
my gut feeling is "yes", given that EXPR can also be a statement.
err... from the grammatical point of view. 11:24
audreyt yes it is.
agentzh thx!
agentzh now feels the need to start another thread on p6l. 11:26
11:28 diakopter joined
svnbot6 r12246 | agentz++ | [t/syntax] 11:34
r12246 | agentz++ | recovered hyper_latin1.t which is *not* "meaningless".
agentzh audreyt: "do { $a } while $b" currently hangs on pugs. i was expecting a parse error or a warning or something like that. 11:38
sorry, i mean "do { $a } while $b;" 11:39
cmarcelo audreyt: make needs a prefix or something now, it's trying to installl things in /lib and /bin
audreyt oy 11:40
11:41 pbuetow joined
cmarcelo (it was a question...) 11:41
agentzh dodges.
audreyt agentzh: the "repeat" thing is new, we havn't implemented it 11:43
svnbot6 r12247 | audreyt++ | * fix --prefix.
agentzh audreyt: okay.
11:43 int-e joined 11:56 bpphillips joined 12:03 dmsantam joined
pasteling "agentzh" at 210.22.200.67 pasted "Pugs doesn't run any of these two tests. could anyone fix the pugs src/ for me?" (26 lines, 357B) at sial.org/pbot/19009 12:21
audreyt agentzh: looking into it 12:25
agentzh audreyt++
audreyt agentzh: fixed 12:28
svnbot6 r12248 | audreyt++ | * infix:<is> is now gone. Runtime infix is doesn't make any sense anyway. 12:29
agentzh yay!!!
up'ing
12:32 Bit-Man joined
audreyt mm, GHC-trunk indeed solved the exponential compilation problem as they promised 12:32
s/compilation/compilation time/
svnbot6 r12249 | cmarcelo++ | * "make clean" before "make" in Judy to force regeneration 12:38
r12249 | cmarcelo++ | of libJudy.a. Also re-copy file to third-party/installed
r12249 | cmarcelo++ | everytime, because it may be lost in cases that
r12249 | cmarcelo++ | third-party/judy/Judy-1.0.3/src/obj/.libs/libJudy.a is
r12249 | cmarcelo++ | preserved.
agentzh oh, pugs build fails again... :(
12:38 Bit-Man_ joined
audreyt oh no :/ 12:38
nopaste the error after r12249?
pasteling "agentzh" at 210.22.200.67 pasted "Win32 build failure" (152 lines, 6.8K) at sial.org/pbot/19010 12:39
agentzh even "nmake realclean" didn't help. :/ 12:40
audreyt hm
agentzh i'll update to r12249 and try again. 12:41
audreyt fixed
r12250
agentzh audreyt++
trying out r12250... 12:43
svnbot6 r12250 | audreyt++ | * fix legacy build dirs
12:43 elmex joined, cm left
agentzh wonders why his machine suddenly slows down while building pugs. 12:46
still failed. i'll check out build_pugs.pl myself to make sure it's not my fault. 12:48
audreyt failed with the same thing?
I changed Makefile.PL so it will rm
third-party/fps/fps.cabal
is that file still there?
cmarcelo agentzh: ghc eats lots of mem, maybe the system is using swap to help, and it slows down a bit..
audreyt if you didn't run Makefile.PL 12:49
maybe unlink it by hand and make again?
agentzh cmarcelo: it was not swapping at all.
audreyt: no third-party/fps/fps.cabal. 12:50
audreyt agentzh: but your error log says
Setup.exe: Multiple description files found. Please use only one of : ["fps.cab
al","pugs-fps.cabal"]
agentzh this time it complaints "Setup.exe: cannot satisfy dependency HsJudy-any".
audreyt oh. better.
12:50 Limbic_Region joined
audreyt nopaste the entire log again? 12:50
agentzh okay
pasteling "agentzh" at 210.22.200.67 pasted "New outputs" (42 lines, 2K) at sial.org/pbot/19011 12:51
12:51 rindolf joined
audreyt agentzh: weird... rm -rf third-party/installed/ 12:53
and make again?
rindolf Hi audreyt
agentzh audreyt: k 12:54
12:55 Bit-Man__ joined
agentzh audreyt: that doesn't help. :/ 12:55
audreyt I need the earlier error messages about fps 12:56
agentzh audreyt: a sec
auderyt: during "perl Makefile.PL"? 12:57
12:57 lisppaste3 joined
agentzh *audreyt 12:57
12:58 Bit-Man___ joined
Limbic_Region hrm - it sounds like someone is having build problems and amazingly enough - I think my Win32 is working just fine this morning with a current rev 12:58
12:58 Bit-Man___ is now known as Bit-Man
Limbic_Region or at least it is past building pugs and is on to the post modules 12:58
Limbic_Region does his happy dance
pasteling "agentzh" at 210.22.200.67 pasted "outputs of "nmake unoptimized"" (161 lines, 7K) at sial.org/pbot/19012
agentzh Limbic_Region: :) 12:59
Limbic_Region: does the latest revision work for you?
12:59 Bit-Man___ joined
agentzh i want to make sure if it's specific to my machine. 13:00
13:00 larsen joined
agentzh wow...a lot of people are joining #perl6... 13:00
dmsantam yeah, interest is growing :) 13:01
agentzh nice!
i wonder if the build errors are caused by my judy-on-win32 patch. but it worked for both Limbic_Region and me. 13:02
sorry, bedtime already...
my machine will be shut down with force within 30 sec...
see you tomorrow &
13:03 agentzh left, kolibrie_ is now known as kolibrie
audreyt oy :/ 13:03
and I thought I just fixed it
rindolf audreyt: ping. 13:04
audreyt pong
rindolf audreyt: hi.
audreyt: I've encountered a strange warning with Jifty and Jifty-DBI from trunk.
audreyt: also the tutorial used to be out-of-date. Maybe still is.
I'm in #jifty. 13:05
svnbot6 r12251 | audreyt++ | * another try...
audreyt rindolf: answered in #jifty 13:07
13:07 Qiang joined
gaal yay! SPJ writes that he fixed HEAD "so that both class and instance declarations are allowed in hs-boot files. But not 'deriving'! 13:08
pasteling "cmarcelo" at 200.245.119.10 pasted "strange error parsing Prelude with pugs and GHC 6.5 -- compilation is fine, pugs is generated." (12 lines, 1.4K) at sial.org/pbot/19013 13:09
audreyt gaal: cool!
cmarcelo: fixed 13:10
13:10 fglock joined
svnbot6 r12252 | audreyt++ | * remove bogus use of runtime "is" in Prelude. 13:11
r12252 | audreyt++ | Reported by: cmarcelo++
13:12 vel joined
cmarcelo checking 13:12
13:26 Bit-Man joined
cmarcelo audreyt++ 13:26
ghc 6.5 compiled pugs here..
audreyt: JudyHS was segfaulting prelude generation on you? or was another test? 13:27
audreyt prelud e gen 13:30
13:30 lichtkind joined
audreyt for gaal and me and luqui and on feather too 13:30
cmarcelo ok 13:33
svnbot6 r12253 | fglock++ | Pugs-Compiler-Rule - fixed some tests for 'return object handles stringification' 13:34
13:36 theorbtwo joined
svnbot6 r12254 | fglock++ | Pugs-Compiler-Rule - fixed a 'bool' test in Match.pm 13:40
r12255 | fglock++ | Pugs-Compiler-Rule - ChangeLog update 13:44
r12256 | polettix++ | [src/Pugs/Prim.hs] 13:47
r12256 | polettix++ | Made quotemeta work like Perl 5: escape only happens for Unicode
r12256 | polettix++ | characters up to 256.
r12256 | polettix++ | [t/builtins/strings/quotemeta.t]
r12256 | polettix++ | Extended test-suite to cover all 256 Unicode characters (but only
r12256 | polettix++ | in the ASCII part, the test for EBCDIC is still the old one in
r12256 | polettix++ | Perl 5.9.3 quotemeta.t).
froh-doh
.oO( my logs are possibly too verbose? )
wolverian you're polettix? not at all. 13:48
froh-doh wolverian: yep, I am :) 13:49
wolverian descriptive commit messages are _good_
froh-doh ok then, I hope svnbot6 won't be kicked off for flooding :) 13:50
audreyt iblech has implemented throttling :) 13:54
fglock mm - is '{}' a sub or a hash? 13:58
audreyt hash 13:59
fglock thanks
audreyt np :)
14:02 xdg joined
rgs perl 5.9.4 is on its way to CPAN 14:07
svnbot6 r12257 | fglock++ | v6 - '{}' is a hash again
audreyt ! 14:08
rgs++ rgs++ rgs++
fglock how is '$?SELF' different from 'self' ? 14:13
audreyt $?SELF is gone 14:15
:)
fglock ok! 14:16
gaal 1 14:18
rindolf Hi rgs 14:26
rgs hi rindolf 14:27
fglock markstos: re CGI-Application, a few problems were fixed; it currently fails in line 174 '\$body' is not implemented yet in v6.pm
rindolf rgs++ # perl 5.9.4
14:29 justatheory joined 14:30 cjeris joined 14:33 Daveman joined 14:37 mdiep_ joined, Limbic_Region joined
svnbot6 r12258 | fglock++ | v6 - ChangeLog update 14:40
clkao fglock: is it passing tests? 14:41
rindolf Hi clkao 14:42
fglock clkao: it seems to run the same tests - I didn't run the whole suite 14:47
clkao mmm, it fails quite a bit for me
svnbot6 r12259 | fglock++ | v6 - t/00-grammar/01-pcr.t - as(Str) changed; removed a test 14:58
15:00 larsen joined
TimToady audreyt: you said repeat {...} doesn't work yet, but I believe it does. 15:17
and I think \123 should be illegal now rather than 123. 15:18
15:30 kanru joined
[particle] \1 is already illegal as per S05 15:31
ingy hola 15:46
cmarcelo audreyt gaal : I changed src/Pugs/AST/Internals.hs to use Judy.Hash instead of StrMap and can't find the segfault in prelude nor the t/builtins/bool/defined.t (which was segfaulting here before). tested 6.5 and 6.4.2 here and in feather too. am I missing something? 15:47
16:01 fglock joined 16:03 Kattana joined 16:04 jferrero joined
audreyt [particle]: that's inside patterns 16:05
not qq
TimToady: oh right, you made it work :)
TimToady: 0123 is warning but \123" is error?
cmarcelo: src/Pugs/Internals.hs surely?
(it's the interning that fails not hashes)
cmarcelo well, i was getting segfaults with hashes too.. for IHash changing Pugs/AST/Internals.hs is enough.. 16:07
[particle] audreyt: true. but currently pge interprets \1 in a pattern as 1. so that'll have to be changed, too.
cmarcelo where's interning stuff?
audreyt [particle]: indeed
cmarcelo: src/Pugs/Internals.hs grep for "table"
[particle] anyway, it's been cleared up in the spec, so i'll change pge/perl6
audreyt particle++ 16:08
cmarcelo: or rather, grep for NOINLINE
cmarcelo ok 16:09
svnbot6 r12260 | fglock++ | v6 - fixed ' @array.exists(20) ' 16:20
16:25 prefiks joined 16:28 larsen joined 16:33 pdcawley joined
svnbot6 r12261 | audreyt++ | * Pugs.Lexer: "\123" is now a warning. 16:35
16:35 Psyche^ joined
[particle] audrey: warning, not error? 16:38
svnbot6 r12262 | fglock++ | v6 - fixed autoquoting in '=>'
audreyt [particle]: yeah. 16:42
16:42 xdg joined
audreyt $ perl -we '"\9"' 16:43
Unrecognized escape \9 passed through at -e line 1.
$ perl -we '"\19"'
Illegal octal digit '9' ignored at -e line 1.
in perl5 these are warnings; not sure about making them parsefails 16:44
svnbot6 r12263 | fglock++ | v6 - fixed hash emitter
[particle] i see, thx
16:47 Psyche^ is now known as Patterner
lichtkind [particle] are you in any way related to the ParTiCLe TCL interpreter? 16:48
TimToady I think the \digits syntax has to be reserved for future metacharacter use, and not be treated as non-meta, or quotemeta breaks.
0123 is an exception because we're hoping to make that mean decimal someday 20 years from now or so. 16:49
[particle] lichtkind: i am distantly related. i've committed a few patches there :)
TimToady but generally if you write something that can't mean anything meaningful it should fail.
audreyt mmm. 16:50
I was rather hoping that "\123" can mean "\d123" in some distant future too...
TimToady decimal characters are rather useless under Unicode 16:51
svnbot6 r12264 | audreyt++ | * Pugs.Lexer: improve the \123 warning a bit by mentioning
r12264 | audreyt++ | \d123 and \o123 as disambiguators
audreyt you mean people usually use hex notation for unicode? 16:52
&#xABCD instead of &#12790
TimToady all the docs use U+ABCD rather consistently. 16:53
there are also cultures (seen it at work) where \12 is equivalent to \x12.
audreyt oy.
TimToady hardware people, mostly 16:54
16:54 trym joined
TimToady so I think we really have to suppress \123 for now completely. 16:54
[particle] what about \\0\d+ ?
TimToady dead by the same criterion. 16:55
[particle] great. consistency++
however, i'll miss \015
TimToady \o15
[particle] yep, it'll work. i'll just have to adjust my fingertips
audreyt it's in. 16:56
svnbot6 r12265 | audreyt++ | * Pugs.Lexer: "\0123" is now as dead as "\123" -- both fatal parsefails.
[particle] train the left ring finger to do the right's work
TimToady audreyt++
audreyt :)
TimToady audreyt.sleep--
make sure \0 still works though... 16:57
[particle] wonder what the lines of code to hours of missed sleep ratio is...
eeh? why not \d0 or \o0 ? 16:58
TimToady those work too.
[particle] why not *force ...
TimToady why?
0 is the same in any radix.
audreyt so "\015" does but "\0fff" works?
s/does/dies/
TimToady yes
[particle] \0123 is octal in perl5 so how would \0fff "work"? <null> followed by 'fff'? 17:01
TimToady ohh, looks like latest smokes are 10% faster.
audreyt [particle]: exactly
[particle] ok
audreyt what about "\000" ?
TimToady illegal
[particle] why, it's the same in any radix?
TimToady \0 is just too ubiquitous to kill 17:02
[particle] my fingers need adjustment
TimToady and is very important in C culture.
[particle] DEATH TO C!
TimToady sure, but by subverting, not by brute force. 17:03
[particle] subversion++ ;)
audreyt *** Exception: Error: \00 is ambiguous: write as decimal \d00 or octal \o00 instead
ingy svn rm C
audreyt that error message itself is kinda bogus
[particle] ingy++
audreyt: agreed 17:04
ingy are there any good midi modules for Perl 6?
TimToady what if they meant \d[0]0 ?
ingy is in a band now
TimToady citizen's? rubber? 17:05
ingy The Strange Attractors
TimToady Band in Boston? 17:06
[particle] whirs chaoticly
17:07 SageLT joined
svnbot6 r12266 | audreyt++ | * "\0" is made legal again, though "\00" "\01" etc are all still failures. 17:07
[particle] looks (from S02) that \08 is legal 17:08
TimToady that one's a close call
[particle] it's clear to me in the spec
Only C<\0> remains, and then only if the next character is not in the range C<'0'..'7'>. 17:09
TimToady that's because that's how I called the close call
yes, I wrote that this morning.
[particle] right, just making sure you mean what you wrote
TimToady I always mean what I right. :)
[particle] and making sure audrey's implementing the right thing
write on!
TimToady she needs to be whacked upside the head occasionally, but she's a good kid. 17:10
audreyt \08 very seldomly means \d[0]8 though...
perl -we '"\08"'
Illegal octal digit '8' ignored at -e line 1.
TimToady \0 seldomly occurs in a string at all except at the end.
unless it's binary data
well, \0 doesn't introduce octal anymore, so that wouldn't work 17:11
audreyt having "\08" and "\09 be treated specially whilst 08 and 09 are not, sounds strange
TimToady sure, but the usual error will be with a normal octal number, not a "transcendental" 17:12
17:12 hexmode joined
audreyt k. so you think "\08" is just fine? 17:12
TimToady I think in the future it's likeliest to mean \0 + 8 17:13
audreyt and in that future, why wouldn't "\01" means \0 + 1?
TimToady only if we make it illegal for 20 years or so
a generation of C programmers has to die off first 17:14
17:14 Corion joined
TimToady and maybe several other languages 17:14
audreyt so for now "\08" is a warning, but "\07" is a error?
and "\0f" is silent passthru?
[particle] we can help with that. perl6, the one true language :)
TimToady I'd just treat \08 like \0f 17:15
audreyt 'k.
[particle] \0 followed by oct => error, hex => passthrough
TimToady s/hex/other/ 17:16
[particle] roger.
svnbot6 r12267 | fglock++ | v6 - fixed some hash slice cases; t/data_types/hash.t passes 5 tests more
TimToady well, passthrough means you send the backslash through too to me
so it's more like terminate-sequence-after-9 17:17
*0
audreyt k, all done. I'll go sleep :) 17:18
obra night 17:19
TimToady nite, thanks!
audreyt thanks too :)
[particle] sure, just like \n
svnbot6 r12268 | audreyt++ | * Pugs.Lexer: "\08..." and "\09..." are treated as "\0" and then "8..." or "9...".
nothingmuch audreyt: when using PAR Weak references are not implemented in the version of perl starts happenning... is -B the wrong option or just not working? 17:26
17:30 kanru2 joined 17:31 kanru2 is now known as kanru
audreyt nothingmuch: it worksforme. does the resulting .par or .exe file have lib/auto/List/Util/Util.so ? 17:32
(or .bundle on darwin;) you can use unzip-l to check
nothingmuch: also, you want tsee on irc.perl.org for PAR questions, or [email@hidden.address] :) 17:33
nothingmuch yep, it does
i googled first
no go
audreyt try PAR trunk?
which platform?
nothingmuch osx
audreyt are you using .exe mode or .par mode?
nothingmuch i'll try it later, methinks
exe i think
audreyt -B is noop on .exe 17:34
nothingmuch ah
well, it's not urgent, i'll punt for now
audreyt PAR trunk implements .bundle pre-extraction
which may help. but I can't duplicate this error
nothingmuch i'll give it a whirl
audreyt 'k. good night :)
17:34 plisk joined 17:38 mdiep joined 17:41 hcarty joined 17:46 hcarty left
svnbot6 r12269 | fglock++ | v6 - fixed '$hash.isa' - t/data_types/hash.t passes 30/55 17:56
17:56 justatheory joined 18:16 buetow joined
svnbot6 r12270 | fglock++ | perl5/ - 'cp Perl6-Value Pugs-Runtime-Value' to be Moosified for v6.pm 18:20
r12271 | fglock++ | Pugs-Runtime-Value - renaming files 18:26
r12272 | fglock++ | Pugs-Runtime-Value - renaming files 18:29
18:34 shachaf joined
svnbot6 r12273 | fglock++ | Pugs-Runtime-Value - renaming classes 18:41
18:43 justatheory joined 18:46 bernhard joined 18:49 cjeris joined, silug joined
svnbot6 r12274 | fglock++ | Pugs-Runtime-Value - start merging with Perl6-Container 18:54
r12275 | fglock++ | Pugs-Runtime-Value - merged Container modules, MANIFEST 18:57
r12274 | fglock++ | Pugs-Runtime-Value - start merging with Perl6-Container
18:57 shachaf_ joined 18:58 wazzuteke left 19:04 justatheory joined
svnbot6 r12276 | fglock++ | Pugs-Runtime-Value - some container moosification 19:06
r12277 | fglock++ | Pugs-Runtime-Value - fixed Makefile.PL 19:08
19:21 luqui joined
svnbot6 r12278 | fglock++ | perl5/ set svn properties 19:32
19:50 zgh joined 19:51 SageLT joined 19:55 frederico joined 20:13 fglock left 20:19 [particle] joined 20:22 bpphillips left 20:31 fglock joined 20:38 autark joined 20:50 crem joined
svnbot6 r12279 | fglock++ | Pugs-Runtime-Value - started s/use Moose/use v6-alpha/ 20:50
21:01 shachaf_ joined
TreyHarris ?eval for (a => 1, b => 2, c => 3) -> $x,$y { say "$x: $y" } 21:05
21:05 evalbot_23396 is now known as evalbot_23466
evalbot_23466 OUTPUT[a 1: b 2 ] Error: Undeclared variable: "$y" 21:05
TreyHarris ?eval for (a => 1, b => 2, c => 3) -> $x;$y { say "$x: $y" }
evalbot_23466 OUTPUT[a 1: b 2 ] Error: Undeclared variable: "$y"
TreyHarris ?eval for (a => 1, b => 2, c => 3) -> $x { say "$x.key: $x.value" } 21:10
evalbot_23466 OUTPUT[a 1.key: a 1.value b 2.key: b 2.value c 3.key: c 3.value ] undef
TreyHarris ?eval for (a => 1, b => 2, c => 3) -> $x { say "$($x.key): $($x.value)" }
evalbot_23466 OUTPUT[$(a 1.key): $(a 1.value) $(b 2.key): $(b 2.value) $(c 3.key): $(c 3.value) ] undef
TreyHarris ?eval for (a => 1, b => 2, c => 3) -> $x { say "$x.key(): $x.value()" } 21:12
evalbot_23466 OUTPUT[a: 1 b: 2 c: 3 ] undef
21:13 [particle] joined
svnbot6 r12280 | fglock++ | Pugs-Runtime-Value - new module Pugs::Runtime::Unboxed - defines semantics for unboxed values 21:23
21:24 weinig_ joined 21:25 justatheory joined 21:27 weinig joined 21:41 frederico joined
svnbot6 r12281 | fglock++ | Pugs-Runtime-Value - Pugs::Runtime::Value is pure-p6; TODO: Capture, List 21:41
21:42 SageLT joined 21:45 Bit-Man joined, bcorn joined 21:46 saorge_ joined 21:54 weinig is now known as weinig_ 21:55 weinig_ is now known as weinig 21:59 lisppaste3 joined 22:05 evalbot_23466 is now known as evalbot_23471
fglock Pugs::Runtime::Value is under way - use.perl.org/~fglock/journal/30643 22:07
lambdabot Title: Journal of fglock (5723)
22:07 fglock left 22:43 justatheory joined
broquaint Anyone know if there's any been any effort to get perl5 implemented for parrot yet? 22:44
leo broquaint: Wow. I had no idea. I'm glad I've read that. What kind of audience were 22:49
oops sorry again: www.poniecode.org/
lambdabot Title: Project Ponie -
leo but that's pretty dead, as the interest seems to have declined 22:50
broquaint Dang. 22:51
Thanks anyhow.
I'm looking to write some code to convert PPI -> PIR.
leo there are a lot more ways to get perl5 running with P6 than just the one specced with the ponie plan 22:52
PPI would be one of that :)
broquaint But I vaguely recalled that there was a beginning of a perl5 implementation with the types and what not.
If there's nothing knocking about I guess I'll just have too bootstrap myself :) 22:53
leo yes, there's even a repo of P5 PMCs somewhere
broquaint Intriguing ...
22:55 zgh joined
leo please mailto nicholas clark for details 22:55
broquaint Right oh, thanks again :) 22:57
22:59 justatheory joined
lichtkind hejho leo 23:02
leo hi lichtkind 23:03
lichtkind leo im the editor guy , its first time i see you there so i had to start some handwaves 23:04
leo hi - nice to meet you here :) 23:05
23:06 rashakil joined
lichtkind yeah how does parrot evolves, i read the ml but thats not the same as listen from horses mouth :) maybe im part of the team i just began to write perl tutorial, now in english too 23:07
i mean perl6 tutorial
leo parrot is evolving steadily 23:08
23:10 rashakil joined
lichtkind leo some catche phrases for my perl6 talk in frankfurt in september ? 23:12
catchy
leo lichtkind: I've no clue - you know what the talk will include 23:14
mugwump hey leo any plans on adding hypervisor support to the parrot virtual machine? 23:15
:->
23:16 zgh left
leo mugwump: after specs have settled, why not ;:-) 23:16
mugwump it might make it easier implementing kernels etc with it that way 23:17
leo but first is STM implementaion, which is more reasoable
23:17 diakopter joined
mugwump yes STM sounds very important 23:18
leo it's on the way - stay tuned on #parrot
23:18 zgh joined
TreyHarris is there a doc on STM syntax as planned in perl 6? all i've seen is audrey's ruminations on her blog 23:23
lichtkind leo comon you know what i mean, i really read the announcements and some more stuff, and dont whant to bother you more than needed but tell me at least where can i read more about recent progress 23:25
leo more is e.g. at www.parrotcode.org/misc/parrotsketch.html 23:28
lambdabot Title: Parrot Sketch - parrotcode
svnbot6 r12282 | Sage++ | Adding some missed nodetypes (apparently I missed most of op_g* for some reason). 23:30
r12282 | Sage++ | Unknown nodes much less prevalent in tests now.
r12283 | Sage++ | Adding some missed nodetypes (apparently I missed most of op_g* for some reason).
r12283 | Sage++ | Unknown nodes much less prevalent in tests now.
23:32 Revision17 joined 23:39 lollan joined
svnbot6 r12284 | Sage++ | Documentation on the new -N option, which stops translation, just for testing. Makes it easier to find Unknown nodes from missing nodetyps as opposed to problems in the code. 23:45
23:45 shachaf_ joined 23:46 lollan joined
TreyHarris STM from the language perspective is more than just C<let>, right? 23:53
23:56 justatheory joined
pasteling "TreyHarris" at 207.171.180.101 pasted "port install problems with ghc 6.5" (37 lines, 1.6K) at sial.org/pbot/19030 23:58
TreyHarris if anyone can help me with my ghc problem i just pasted, i'd be grateful

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!