Preflight #1: xrl.us/hxkk | geoffb's column (/. ed): xrl.us/hxhk | pugscode.org | pugs.kwiki.org | paste: sial.org/pbot/perl6 | www.geeksunite.net
Set by autrijus on 9 October 2005.
gaal er, except that it isn't appearing on ther server. putting on feather. 00:00
svnbot6 r7516 | autrijus++ | * disable rules.t correctly without taking lots of parsing.
dduncan are changes post r7512 ending up in 6.2.10, or are they for later? ... if they're for now, every update to that release should have a corresponding ChangeLog update to what r is represented 00:01
separate point ... 00:02
*** PIL2JS needs the Class::Rebless module from CPAN.
should this be considered a warning or error?
autrijus dduncan: warning, I think
dduncan make[1]: *** [pure_all] Error 1
make: *** [build_perl5] Error 2
autrijus hm, but it dies.
dduncan that comes with it, but pugs seems to have been produced anyway
autrijus gaal: thoughts on us taking the drastic measure of bundling Class::Rebless? 00:03
dduncan running the pugs, it says internally that it is the newest r
gaal fscking adsl upload ratios :(
autrijus: nah, put it in the p5 core. :-p
autrijus okay, failing that. 00:04
gaal perlcabal.org/~gaal/smoke_xp.html
autrijus I'm thinking to append to perl5/PIIL2JS/lib.
dduncan is the r# inside the pugs binary, or inside a data file that it reads?
gaal sure, it's a tiny module.
autrijus dduncan: inside
gaal 556 files 10340 test cases: 10340 ok, 0 failed, 1551 todo, 954 skipped and 0 unexpectedly succeeded 100.00%
dduncan then that should mean the make worked then
but the 'error' can confuse users
autrijus dduncan: solved. thanks! 00:05
svnbot6 r7517 | autrijus++ | * add Class::Rebless to avoid frivolous errors.
gaal lol
autrijus :)
gaal ah, the server did of course accept the smoke, i just forgot to look for 6.2.9 :) 00:08
yo other smokers, where are y'all?
dduncan I noticed that all the smokes said 6.2.9
gaal yes, pf2 bug 00:09
known and fixed
s/2/1/ 00:10
okies, I've got $work tomorrow and it's getting late, going now 00:17
happy flying :)
autrijus thank you gaal :) 00:18
putter: the upshot is that embed won't work reliably for PGE until 0.3.1
as the C side calling conv seems broken.
c'est la vie.
gaal (-: zzZZZ &
putter back 00:20
autrijus putter: a bogus check was placed at parrot_pass_args_fromc 00:21
has since been rescinded in trunk
dduncan autrijus, will there be any more preflights, or do you now have enough confidence to just ship from the repository at time?
putter ah. so README or INSTALL fodder?
autrijus putter: nod. 00:22
putter done? or needing to be done?
autrijus putter: needing to be done. can you help?
putter doing...
my other incident was the eval_yaml failure. 00:23
autrijus yeah, seems like a 64bitism
ghc 6.4.1?
putter yes
embedded parrot 030 is broken on all platforms? and the trunk is believed to work? 00:25
autrijus yes. yes. 00:26
putter does this pugs still work with external pre-leo-ctx5 parrot? or are we only >=030 00:29
svnbot6 r7518 | autrijus++ | * Use the newer parrot_load_bytecode API to load PGE
r7518 | autrijus++ | code and avoid the bogus "eval". This only helps in
r7518 | autrijus++ | post-0.3.0 parrot though. sigh.
Megzl Is there a perl6 version to test? 00:32
what the fuck. Perl6 written in haskell?
You people are insane. 00:33
obra ...
dduncan you come from slashdot?
sounds like it, Megzl
Megzl I didn't exactly come on IRC just cause of slashdot. 00:34
autrijus greetings.
Megzl #Perl6 was only a /join away.
But, I did see the slashdot article.
autrijus Megzl: insanity (in moderation) is celebrated around here, yes.
cognominal ho, pugs slashdotted. 00:35
autrijus putter: how's it going?
cognominal I like to read /. threads to learn about all the misconceptions that people have about Perl :)
Megzl Particularly, I read about Geoff Broadwell (Pugs? is his name), and in his little 'description it said:
"He is on a personal mission to prove that dynamic languages are by far the best programming option for almost every purpose,"
autrijus cognominal: important thing, that :) 00:36
dduncan these days they are, when programmer time is more valuable than cpu time, and hardware is cheap
Megzl I came here to find out if you really think that's possible...
I work on 3d code.
How could that possibly be written in an interpreted/non compiled language?
autrijus Megzl: it can't. 00:37
dduncan it could be controlled by one though
Megzl Then Geoff Broadwell's descriptor is incorrect.
autrijus consequently, you'll notice that almost all dynamic languages nowadays are compield.
dynamic is a property of the type system.
dduncan use compiled for high calculation intensive work, and hardware speaking, and dynamic languages can control those parts, be the glue
Megzl When can't you have a compiled dynamic language? 00:38
That compiles to an executable.
I don't take dynamic languages very seriously because they don't compile to a fast executable.
None I've heard of, at least.
I'm really surprised nobody has done it.
autrijus Megzl: Dylan, I hear.
Python with Psyco, too.
Megzl I'm not stuck with C++. 00:39
I only use it cause it's fast. But I don't love the static typing.
cognominal Megzl: sometimes you want to optimize other dimensions than speed
wolverian oh, nice, autrijus is on /. :) congratulations!
(I think..)
autrijus Megzl: Scheme and lisp, from another setting.
Megzl: there are plenty of compiled fast dynamic languages :)
it is a harder problem to compiled them, yes, due to the lack of type information
Megzl I just love C++'s syntax.
autrijus but it has been done and will be done better still :)
dduncan there's a brand new article ... someone want to first it? 00:40
oop, too late
right between those posts
Megzl Can you tell me why dynamic languages are good?
dduncan a lot faster 00:41
Megzl I do tons of dynamic stuff in C++, I have a whole framework built around boost::any.
dduncan to develop
Megzl It does everything I could want.
putter autrijus: do pre 030 parrots work external, or are we strictly >= 030?
autrijus Megzl: that makes it a dynamic language.
dduncan Megzl, think of dynamic languages as a less dynamic language with certain libraries built into the language
autrijus putter: all reasonably recent parrot should work external
Megzl: it's really a gradient... I'm a firm believer of "dynamic with prototyping, static when optimizing" 00:42
dduncan eg, certain things you have as libraries in C are built-in to Perl, such as basic memory management and string libraries and etc
autrijus Megzl: and any language that lets you do that is a Good Thing.
C++ included.
Megzl I have a little bit of a dynamic language I use for my own purposes on top of C++ -- and the syntax is simple. And it compiles to be fast. doing Perl seems like going too far 00:43
with C++ I get the best of both
with Perl I only get the dynamism
Khisanth dduncan: and regular expressions :)
autrijus Megzl: yes, that is very true. which is why perl6 is going for optional staticness.
dduncan indeed, but that's more of a perl thing than languages in general
Megzl ah
autrijus Megzl: it's really like Dylan in that regard. I think you'll like Dylan. 00:44
Megzl well, if perl6 can be static, I'll give it a more serious look this time.
autrijus Megzl: at least, the standard library is sprinkled with type annotations.
dduncan one of perl's strenghths is borrowing good ideas from other languages and improving them
autrijus the user can still elect to write all-dynamic scripts.
wolverian I'm not sure how static perl6 will be, but it will certainly be optionally typed.
dduncan staticness in this case 00:45
cognominal Megzl: Perl6 will targets parrot which is a VM build around a typed assembler
so supporting (alos) staticness is a core goal of Perl6
s/alos/also/
Khisanth hrm Perl6 seems to be targetting several things :) 00:46
cognominal that's why it is long in the making
autrijus ?eval multi add (Int $x, Int $y) { $x+$y }; multi add (Str $x, Str $y) { "$x$y" }; add('Hello ', 'Megzl'); 00:47
evalbot_7518 'Hello Megzl'
mugwump chasing many rabbits down many holes
autrijus ?eval multi add (Int $x, Int $y) { $x+$y }; multi add (Str $x, Str $y) { "$x$y" }; add(21, 21)
evalbot_7518 42
autrijus Megzl: see? :)
Megzl Yes, I see.
The only thing I don't see, is why it can't actually compile to fully optimized fast code under the hood. And I guess that's what you guys are telling me perl6 will do. 00:48
autrijus Megzl: yup :)
Megzl: one of us has just started playing with type inference to generate fast native C code. 00:49
(it's in misc/Blondie/ in the pugs tree.)
dduncan I'll be back in an hour or so, meanwhile smoking r7517 00:51
putter I haven't been using external parrot much lately. how much difference does it make for end users? (providing context for whether they should bother grabbing cvs parrot)
autrijus putter: almost not at all
putter: a slightly slower startup time (1sec)
that's all.
since we keep parrot process harnessed fastcgi-style. 00:52
putter k
autrijus putter: I'm not sure I can get the syck 64bitism fixed for this release. 00:53
esp since my home box of amd64 is currently offline. 00:54
putter: I think we'll leave it failing for you for now, unless you feel strongly about it
otherwise I'm ready to fly.
putter? 01:02
putter pong. ok. documented. checking in 01:05
autrijus excellent!
putter sorr for the delay. english writing braincells on strike 01:07
svnbot6 r7519 | putter++ | INSTALL: document parrot dependencies and eval_yaml() bug
putter actually, does yaml being broken have any larger implications worth mentioning? 01:08
(not that it's clear there are more than a handfull of x86_64 pugs users yet...)
autrijus your mentioning is fine.
okay. fly! 01:09
putter revisits concept of having a ui that points out when his clock speed has decayed unnoticed... 01:11
ah, an airborne pugs... a nice sight... 01:12
svnbot6 r7520 | autrijus++ | * This be Pugs, version 6.2.10. 01:13
autrijus whew. thank-you all :)
autrijus uploadeth
pugscode.org/dist/Perl6-Pugs-6.2.10.tar.gz 01:14
sanity check before I CPAN it?
putter grabbing... 01:15
err 01:16
autrijus ? 01:17
putter waits 11 seconds so tarfile dates are in the past...
autrijus err.
putter err, what happened to cookbook? 01:18
autrijus examples/cookbook? 01:19
what happened?
putter yes, its in my SVN, but not in the tar...? checking...
autrijus it is there to me. 01:20
putter ok, so just some specific files are missing, eg
03dates-and-times
autrijus there were no files in it anyway. 01:21
maybe you did not svn add them?
no files were there in the previous release either.
other notable omissions?
pasteling "putter" at 66.30.119.55 pasted "svn vs release diff" (32 lines, 1.5K) at sial.org/pbot/13592
putter debian?
autrijus checking. 01:22
debian is explicitly not listed. 01:23
by rafl, so I assume it's okay.
other looks harmless.
putter MIME/Base64?
autrijus empty directory.
putter err ext/MIME-Base64/lib/MIME/Base64.pm? 01:24
never mind
autrijus :)
may we uplaod to CPAN? :)
putter k, well, my svn copied work, and it matches (havent checked permissions), but nothing obvious... 01:25
fglock hi all - congrats for the new release!
autrijus 6.2.10 released! xrl.us/hxnb | geoffb's column (/. ed): xrl.us/hxhk | pugscode.org | pugs.kwiki.org | paste: sial.org/pbot/perl6 | www.geeksunite.net 01:25
autrijus thanks fglock! 01:25
putter what happend to t/builtins/system? 01:26
autrijus explicitly delisted - space in directory names - beyond workaround.
putter k
autrijus (MANIFEST syntax problem)
putter then happy pugs in flight. which just has entirely different visuals since I browsed a "Pugs" magazine. 01:27
autrijus okay, uploading :)
putter++ # last-minute sanity check 01:28
fglock putter - re lazy Array: I think there is a way to make it O(1) - I'll try to work on it tomorrow 01:30
putter nifty.
fglock I was away for about 4 days due to work - missied the best part of the release work :(
putter me too
autrijus :) 01:31
putter yaaaaaayyyyyy PUGS! (inflect <cheer>)
autrijus waves goodnight
fglock good night autrijus
putter good night autrijus &
autrijus *wave* &
Khisanth inflect? 01:34
putter said with a particular inflection. should probably have been say "..." :inflect<cheer> ;) 01:36
fglock: fyi, I've a rewrite of EvalX is started last weekend. will hopefully get iā€Œn sometime in the not too distant future... 01:46
s/is/i
dduncan ... back 01:53
um, yay and stuff
fglock putter: we've talked about moving some files around - are there any changes? 01:54
dduncan well, now that the release is out, big changes can be made ... 01:58
putter no... feel free. :)
dduncan now to smoke another camel ... 01:59
its amazing how some coincidences can work together
putter ;) 02:00
putter delights in Earth, at 10^6 meters per mm, being... a blue marble. (13 mm) 02:01
putter does not delight in having to CPS convert, and thus notably uglify, the prolog regexp engine so local() in repeated embedded code clauses works correctly. :/ 02:03
Pyked heya 02:55
03:10 sleepster is now known as typester
putter fyi, Perl6-Pugs-6.2.10 passed make test on FC3 x86_64 w default config. yay. Including eval_yaml(), which I was expecting to fail. !?! 03:19
hi Pyked. you've hit the "night time" of #perl6. things will pick up in another 6 hrs or so. fyi. 03:21
Khisanth putter: that is pretty strange :) 03:27
considering pugs is hacked by people in various timezones
svnbot6 r7521 | Darren_Duncan++ | ChangeLog updated to add the next release template, for Pugs 6.28.0 03:36
putter weak us pacific presence. and most of us east doesnt stick around until the wee hours. and asia folks tend to run late. so europe isnt awake yet, asia has stumbled off to bed, and europe isnt up yet. happens most every night. 03:37
mugwump sure, but after the US goes to bed it's a while before Europe wakes up
stevan hey mugwump long time no see 03:38
mugwump yeah I know. It's been months since I did any large amounts of Perl 6 03:39
your metamodel is looking delightfully simple when I last looked at the pictures
stevan thanks :)
the eigenclass thing threw me for a little while,.. but I think I got my head around it now 03:40
putter the ruby C source is nicely readable. hmm, though maybe not the part you would care about. 03:42
svnbot6 r7522 | Darren_Duncan++ | moved /misc/Rosetta-Incubator to /ext/Rosetta-Incubator since this code should be installable in time for 6.28.0
stevan putter: I am not very good at reading C actually
putter ah
stevan I get lost when I see too many &&& and *** stuff 03:43
but eigenclasses were mostly a conceptual barrier anyway
I needed to see how they fit in,.. and where
dduncan mugwump, stevan, with this new stretch ahead of us up to 6.28.x, we can now start coding some serious object oriented perl 6! 03:44
oh, and I'm on the west coast of NA, so its about 9pm here 03:45
stevan dduncan: yes, that is the plan :)
mugwump almost 5pm here in NZST
and I'm giving an svk talk at 6pm :)
stevan 11:44 here in EST
dduncan to whom?
mugwump australia?
stevan mugwump: nice. I just started using SVK myself
dduncan I'm debating whether to switch to svk from svn 03:46
mugwump oh, 11:44pm in America's EST
stevan dduncan: the nice thing about SVK is that it is so interchangeable with svn
I havent used any of the advanced features of svk yet actually,.. I use it just like it was svn 03:47
mugwump it's 1:46pm in Australia's EST :)
stevan dduncan: there is also a very nice DMG on the svk website too,.. point-click install for Tiger
dduncan stevan, will have to look at that ... does it work with Panther too?
stevan mugwump: I didnt realize australia had an EST too,. however that make sense,...
dduncan they have 3-4 time zones I think 03:48
mugwump Australia and US Timezones are the only clashes
stevan dduncan: never tried it with panther,.., but when I upgraded to Tiger recently it made installing svn much easier (two for the price of one)
because the svk dmg comes with a nicely pre-compiled svn too 03:49
dduncan I plan to upgrade to Tiger post the huge 10.4.3 patch, perhaps waiting to .4 also
mugwump that really is the great thing about svk ... it just works like svn until you start merging
stevan dduncan: why are you waiting?
dduncan for one thing, Tiger had a lot of changes and I wanted to let them stabilize 03:50
I only have one comp, for work and testing, so want to lower my risk
in particular, Tiger adds filesystem changes, which are the highest risk type of change
stevan dduncan: I know the feeling,.. I waited until my recent $work project was done until I upgraded
dduncan moreover, I don't *need* Tiger's new features 03:51
stevan dduncan: thats what CD Burners are for ;)
dduncan I make backups too, but that's not the point
stevan I dont need them either,..
dduncan I also wanted to make sure my software was compatible first ... I'm upgrading my apps first
stevan however,.. I have to say Tiger seemed "zippier"
dduncan just did ghc a few days ago, to 6.4.1
in any event, I do have a Tiger install DVD since it first came out ... part of an ADC mailing 03:53
mugwump, do you plan to make a Tangram in Perl 6 soon ... to go with your data file conversion utility? 03:54
mugwump ok, the plan on that front is to get the T2 (RAD|CASE) tool going on Catalyst + T2 + TT + AJAX on Perl 5... 03:55
then make sure T2::Perl6 works against the lastest MetaModel 03:56
Then I can hopefully write in Perl 5 and 6 at the same time
dduncan so in other words, you have an existing development project in perl 5 to complete first 03:57
mugwump With that other tool I can rope more people in with a bit of luck
more people from my current job :)
dduncan and if more people from your job use it, you can get to work on it at your job more often 03:58
on it or with it
mugwump that's the idea. Also I should be able to use the tool for re-engineering existing applications 03:59
putter stevan: I have urls for some browsable ruby c src files... but I suspect this is a case where independent invention is preferable, so there's greater potential for innovation. 04:00
dduncan in my case, I'm hoping that if any other FOSS projects (eg Tangram) end up using Rosetta when it's ready, then the existing user bases of those projects also become potential income sources for me, in regards to paid support or improvements to my part of what they're using ... one can dream anyway 04:03
stevan putter: URLs are good too,.. it is better to know what they are doing then to repeat their mistakes :)
putter yes, but its often worth leaving such cross checks for after the dev is done... but here they are... 04:05
in www.ruby-lang.org/cgi-bin/cvsweb.cgi/ruby/ www.ruby-lang.org/cgi-bin/cvsweb.cg...c?rev=1.92 www.ruby-lang.org/cgi-bin/cvsweb.cg...?rev=1.179 www.ruby-lang.org/cgi-bin/cvsweb.cg...?rev=1.836 i think those are the key ones. skimmable. 04:08
stevan putter++ 04:09
putter "i think those are the key ones" as in "i'm tired, so I may well have missed some"... :)
stevan object.c and class.c would be the expected places for it,.. 04:10
putter yeah. i thought there was another piece, but i guess I was confused... 04:14
dduncan the new pugs is now on search.cpan.org ... I've downloaded and am now smoking that camel 04:23
putter ok. way past end of day. 04:26
good night all & 04:28
svnbot6 r7523 | Darren_Duncan++ | Deleted the older ext/Locale-KeyedText, which ext/Rosetta-Incubator supersedes 04:57
scook0 I notice we made slashdot again :) 05:47
jql programming languages are underloved on /. 05:56
PHP: sucks. Perl: sucks. Perl6: gonna suck. Python: sucks. Ruby: lets port Rails to X. C#: evil. use C++. C++: sucks, use C# 05:57
jql frowns upon /. 05:58
gaal cheers for 6.2.10
man this night could have been longer
dduncan but now that 6.2.10 is out, the *real* fun begins 06:16
or at least more-real fun
rafl Yay, a release! :-) 06:17
QtPlatypus dduncan: What are the big ticket goodies that 6.2.10 has? 06:18
dduncan see the changelog 06:19
what I mean by fun is what the *next* release, 6.28.0, will bring
full metamodel integration, meaning proper object support, and so we can start writing a lot more un-cludgy perl 6 code 06:20
those are big changes, which couldn't really start until 6.2.10 was out ... that's my reason for more-fun
of course, we can start using those between now and then, so the fun begins within days 06:21
or now
I'm starting to write serious perl 6 now, even if it doesn't run right away, it should by the next release
forsaken whats the learning curve going from 5->6? 06:44
jql the initial curve is search-and-replace 06:46
alot of code will still work with a search-and-replace
Khisanth after that, brain upgrade! 06:48
forsaken hehe 06:49
just curious if it was a good idea to start learning perl6 now, or wait until the parrot version was realeased (so i can use it on other boxes)
scook0 the learning curve will be easier if you already know a few other languages 06:50
because P6 fixes a few things in P5 that other languages already got 'right' the first time
jql perl tends to be a moving target even in the most stable of times 06:51
scook0 forsaken: just be aware that 'production-ready' P6 isn't something you should be holding your breath for just yet 06:52
jql perl6 is still a parasitic language... 06:53
scook0 I love the fact that P6 is willing to "steal from the best with pride" 06:54
revdiablo That's lwall's MO 06:55
=)
scook0 avast, nothingmuch!
jql perl6 is a voyage of discovery and pillaging
nothingmuch morning
is it pirate day II?
jql @Larry is a band if pirates, buring archive networks and raping languages 06:56
scook0 see, that's the thing I don't like about pirate day
nothingmuch this is funny, because I just reorganized the pirate flag I posted on my cubicle on sep 19th
scook0 now, if you talk like a pirate on any other day, people tell you it's the wrong day for it
nothingmuch it's not wrong =)
it's just surprising
jql it's never wrong. RAmen
scook0 :) 06:57
nothingmuch cars and ironed shirts don't get along
scook0 nothingmuch: how so?
forsaken aye
nothingmuch cars have seat belts
and other nassty things with pointy teeth
that make the shirt all crumpled
it's also the way you sit in them
and the fact that I'm not a careful persomnm 06:58
as evident from my typing ;-)
scook0 is there an easy way to match /\s+/ in Parsec? 07:01
`whiteSpace` does \s*, but I need \s+
aha: skipMany1 space 07:03
Khisanth hmm 07:04
nothingmuch: you just need to gain weight or wear smaller shirts! :P 07:05
nothingmuch Khisanth: interesting theory
nothingmuch likes the way parsec is verbose but still usable
we should be able to do that in perl rules 07:06
have aliases to things like \s and *
or +
nothingmuch complains about gaal's blog and rss 07:07
scook0 parsec is absolutely fantastic 07:08
parsing is such a natural fit for monads and combinators
nothingmuch wonders if scook0 has an opinion about PArrow 07:09
nothingmuch has started reading, but at some point had something else to do 07:10
scook0 I've not really looked into arrows
I hear they're fancy, but I don't know much about them
nothingmuch recites: monad is-a arrow, arrow is not a monad 07:11
or something like that
scook0 arrow generalises monad?
nothingmuch yup
for every monad there is a corresponding arrow thingy
but not vice versa
my problem with reading the article is the fact that there were something like 17 different binding operators 07:12
and by the time the guy was making a point, operators were oozing out of my ears
scook0 nothingmuch: got a url?
nothingmuch searches 07:13
i think it was this: www.cs.chalmers.se/~rjmh/Papers/arrows.ps 07:14
scook0 eww, .ps :)
nothingmuch pstopdf 07:16
scook0 or ggv
nothingmuch or lpr 07:17
scook0 last time I used ps2pdf the output was horrible
probably needed to tweak some settings
jql kghostview likes it 07:18
scook0 so I just opened up the .ps in Preview on a Mac -- crystal clear
nothingmuch this seems promising: www.haskell.org/tmrwiki/ArrowsIntroduction
scook0 gah, I still can't get `foo.bar` to work
jql crap, you're right. PDF sucked 07:19
the baseline jumps
nothingmuch preview does ps to pdf conversion internally... i wonder if the tool is exposed
scook0 well, I was using ps2pdf on my Ubuntu box 07:20
versus Preview on the uni mac
dduncan as I recall, postscript is Mac OS X's native GUI language 07:23
scook0 ssh++ (because installing latex on a uni lab computer would not be fun) 07:24
geoffb <rez>
dduncan all the drawing to screen is in postscript, as with the print system, so it should be natural to simply open such files also
geoffb Interesting. I get my 15 seconds of fame on my day off . . . .
dduncan tjat
scook0 when Preview opens a .ps, it saves to a .pdf before displaying it 07:25
dduncan that's also why there is OS-level support for many complex graphical effects, such as "aqua" uses in abundance
scook0 VICTORY! 07:28
(`foo.bar` now parses correctly!)
nothingmuch scook0++ 07:29
scook0 modulo me having broken something else ;)
in current pugs it parses as `foo(bar)` 07:31
my initial patch broke it completely
nothingmuch it was broken before you came
and now it is fixed
scook0 my next attempt turned it into `foo($_.bar)` :)
nothingmuch after you came and fixed it
hehe
scook0 nothingmuch: I'm just about to clean up the patch--care to test it in a bit? 07:32
nothingmuch sure 07:35
scook0 perlbot nopaste 07:37
perlbot Paste your code here and #<channel> will be able to view it: sial.org/pbot/<channel>
nothingmuch svl diff ... | pbotutil.pl -s pugs -m "scook0's patch'
pasteling "scook0" at 220.253.96.162 pasted "Parser patch MkII" (152 lines, 6.3K) at sial.org/pbot/13596 07:38
scook0 huh?
nothingmuch pbotutil - it rocks
scook0 nothingmuch: teatime; bbl &
nothingmuch ciao
hola Aankhen`` 07:47
Aankhen`` Morning nothingmuch. 07:48
scook0 mmm ravioli :)
Khisanth svl? 07:49
nothingmuch Khisanth: errm, svk 07:51
nothingmuch had pesto twice this week
i've been really lazy about dinner
because:
pesto rocks
it doesn't make a big mess
it's fast
scook0 is still using svn, but plans to try out svk some day 07:53
nothingmuch it's a drop in replacement
but you can also do local branching
scook0 so everything I do now with svn works word-for-word? (s:g/svn/svk/)
nothingmuch AFAIK yes 07:55
but ask #svk to be sure
jql ask svk caused a dyslexic seizure for me 07:59
I couldn't quite make sense of it 08:00
scook0 gah, my stubborn fingers refuse to type 'svk' :) 08:01
nothingmuch alias 08:02
jql svn svk svn svk svn svk svn svk svn svk. I don't even touch-type, and it's hard
nothingmuch always gets higih rep for his banal perlmonks posts, but the ones which I really invest effort into are usually ignored 08:03
jql I only answer really really hard questions. I think I have only 2 or 3 posts on perlmonks 08:04
scook0 jql: at least they're both fairly easy to type
revdiablo nothingmuch: That's the way of it... the Monks are weird sometimes.
jql cherry-picks the interesting ones
scook0 every time I type `my` in Perl, I cringe :)
nothingmuch answers 1 - 0.3 posts a day
dduncan why?
scook0 no other two-letter combination involves moving one finger so much 08:05
jql I don't touch-type. it helps 08:06
especially with !@#$%!
I use 2 hands for my
scook0 jql: language! :P
jql right hand hits m, left hand hits y
dduncan not being able to keyboard is nasty ... can't do much of anything without that skill 08:07
jql and my right hand also presses space & shift, while my left moves to the top row for the imminent $@%
scook0 touch-typing was the single most useful thing I learned in high-school
dduncan when keyboarding, the same finger types the 'm' and 'y' anyway, and they aren't far from each other ... on a QWERTY keyboard anyway 08:08
scook0 dduncan: using the same finger makes it harder 08:09
dduncan maybe, but you use the same fingers for many keys anyway, the 'my' isn't special
Khisanth slower :) 08:10
nothingmuch also does two handed my
Khisanth which was one of the design goal of qwerty or something ...
two handed?
dduncan still, you can't get keywords much shorter than 'my', save the math operators
scook0 sure, it's great to read
dduncan the design goal of qwerty was to slow people down
jql keyboard efficiency is why I liked the ./ operator
nothingmuch Khisanth: the design goal of qwerty, minus the myth was - to be fast enough but give good distribution across the keyboard
dduncan so typewriters didn't jam
jql so very easy 08:11
scook0 I guess I'll just have to pick up two-finger my
jp-autark we should rename "my" to "me"
nothingmuch so that your hands are usually hitting distant keys, and the arms don't jam
scook0 jp-autark: me @hearties
jql and rename our to us?
dduncan well, feel free to customize perl 6 in that way ... you can, you know
revdiablo nothingmuch: So it was more about the distribution than the slowness? 08:12
Khisanth jp-autark: too bad the Pirate name is already used for something else :)
nothingmuch revdiablo: yes
dduncan yes
jp-autark :)
nothingmuch it was slightly slower than ideal
Khisanth shiver me $timbers; ..
dduncan its so keyboards don't jam
because hammers close together weren't hit in sequence
nothingmuch but dvorak is, in practice, only about 10-15% faster for people who are pushing the limit of qwerty
dduncan less likely to collide with each other
s/keyboards/typewriters/ 08:13
mechanical ones
nothingmuch dvorak has another advantage besides speed, though - the movement is not as big, so i hear it's healthier
Khisanth old typewriters :)
nothingmuch but Juerd is the person to ask about this
scook0 nothingmuch: are you running the test, btw? 08:16
nothingmuch scook0: ooops... I started build, but got distracted
tee hee ;-)
(burning stuff, trying to hunt a bee, etc) 08:17
scook0 crosses fingers for test-pass
curses, svk.elixus.org is broken 08:18
nothingmuch scook0: i'll let you know if tests are failing due to parse errors 08:19
and i'll also let you know when it's done
scook0 nothingmuch: thanks
confident about this one
scook0 touches wood 08:20
nothingmuch minimizes window and starts working more intensively
dduncan as of r7523, all tests pass on the default haskell backend
current version that is
nothingmuch dduncan: this is an uncomitted patch
dduncan yes, well go all out
this is the time to go all out
scook0 well, it was committed once
then reverted, because it was broken :)
nothingmuch this is really cynical: [email@hidden.address] 'Please help Hurricane Katrina victims !!! ?' 08:21
dduncan url?
jql geez
dduncan a spam? 08:22
nothingmuch dduncan: err, yes
dduncan anyway, good night
nothingmuch darcs devel listmod
good night
scook0 bye dduncan
08:26 Aankh|Clone is now known as Aankhen``
nothingmuch scook0: bad news 08:48
scook0 gasp
what have I broken? 08:49
nothingmuch t/01-sanity/01-tap.................................***
unexpected "j"
expecting operator, postfix conditional, postfix loop, postfix iteration, ";" or end of input
at -e line 2, column 7
***
sorry, i was away for a while
scook0 looks like I'm failing to backtrack somewhere 08:51
hold on
01-tap doesn't even have a 'j' 08:52
looks like my patch screwed up before the first sanity test even began parsing!
pasteling "nothingmuch" at 212.143.92.226 pasted "exact output of ./pugs -Ilib t/01-sanity/01-tap.t 2>&1" (4 lines, 167B) at sial.org/pbot/13597
nothingmuch say '1 08:53
the ' is ticking it off
scook0 say('1') works fine though 08:54
it's a problem with listop-style calls
nothingmuch =/ 08:57
geoffb Geez, someone came in here saying Perl wasn't good for 3D, and noone pointed him to my Perl OpenGL articles? 09:12
awww, I'm hurt . . . .
:-)
nothingmuch when?
geoffb scrolling back, hold 09:13
colabti.de/irclogger/irclogger_log/...sel=63#l90
How be you, nothingmuch? 09:14
nothingmuch i'm good 09:15
you?
geoffb pretty decent 09:18
Tired though . . .
probably headed to bed in near future
Just got kinda jazzed to have one of my articles get noticed, so that kept me up -- and then I decided to catch up with the backlog 09:19
Holy crap. My page views for that blog entry went from ~850 to ~5200 so far, presumable because of the /. 09:20
Given that my previous record was ~2400
That's cool.
nothingmuch יו×Ø×Øשי! 09:23
huraah!
geoffb :-)
nothingmuch (sorry, didn't notice hebrew was on ;-)
geoffb heh
nothingmuch transliterated that is "yorrshi"
sounds happy
maybe it means something in an asian language ;-) 09:24
geoffb Here's an interesting question: What percentage of strings of N phonemes actually make a word used in some language? 09:25
OK, maybe just interesting to me
nothingmuch it's interesting
i bet there's statistics
geoffb That would be pretty cool to see. 09:26
Well, I have no good excuse to be awake at this hour
so I think I will try to not be
& 09:27
(Oh, I forgot: congrats to everyone on 6.2.10!) 09:28
& # no, really
scook0 spoot
I can get foo.bar
or I can get listops
but never both in the same build
nothingmuch scook0: =( 09:31
would having access to a moderately fast machine help?
and why isn't feather helping, btw?
scook0 nothingmuch: I never bothered to get a feather account
nothingmuch hmm
anyway, i can give you access to a dual processor OSX box 09:32
autrijus waves weakly and stretches
nothingmuch hi autrijus!
autrijus wikipedia admins++ # en.wikipedia.org/wiki/User_talk:Gt1234
scook0 (back when feather started up, I wasn't interested in making code-changes, so I didn't need another box)
slashdot?
nothingmuch heh =) 09:33
autrijus although that vandalism stayed up for 6 hours :-/
(someone changed my wikipedia photo into goatse... much fun. *not*.)
nothingmuch www.google.com/search?q=gt1234
you'll notice that many of the hits are similar
"your post has been removed as a violation" ....
i think someone loves you, autrijus =) 09:34
autrijus riight =)
I can do with less of that kind of love.
scook0 autrijus: I can't seem to get listops and `foo.bar` in the same build
peppering my code with `try` doesn't seem to help either 09:35
what can I use to trace the parser?
autrijus scook0: where's your latest patch?
scook0: you can use trace.
trace :: String -> a -> a
handy stuff. a refreshing desert in the oasis of referntial transparency.
scook0 I would prefer a refreshing dessert, to be honest 09:36
scook0 ducks
autrijus *smile* 09:37
I need to get some rest... glad that autrijus.org is back to life 09:38
nothingmuch the desert is actually quite refreshing in a clean-your-mind-as-you-accept-death-from-dehydration kind of way
autrijus: if you like you're more than welcome to backup crucial files to my machine, btw
i have a raid-0 but no real backup plan
(though I do plan to do something for mail soon) 09:39
autrijus nothingmuch: thanks for the offer :)
scook0 thinks nothingmuch is addicted to helping people 09:40
autrijus good addiction, that
nothingmuch scook0: i think that sys admin level help should be shared
for example, I don't want to pay for a backup MX
i have a friend who does it for me
to get these benefits I have to provide them too 09:41
it's just socialist thought =)
(in the anarchistic, back to the roots sense)
if i can help at no cost to me, i go ahead and do it
because it makes people be nicer to me
it makes me feel good
and it makes people be nicer to other people, in a cascade 09:42
autrijus funny, as I often explain anarchist attitude as "in the socialisitc, back to the roots sense"
scook0 :)
nothingmuch =)
i think they're so synergetic it's hard to tell them apart
they cover two aspects of how a good society should ideally support itself
autrijus ...provided that noone is able to kill another 09:43
...and that communication is inexpensive.
nothingmuch well, that's not part of an ideal society
i'm not pretentious
but so far lambdacamels have done me only good =)
autrijus neither am I... on IRC you can't kill one another and communication is instaneous :) 09:44
nothingmuch =)
rafl prepares new official 6.2.10packages for Debian. 09:47
autrijus rafl++
nothingmuch css-- for the 40 millionth time 09:54
vasc autrijus, i assume i am talking with the pugsmeister himself?
kgftr|konobi css++
nothingmuch vasc: the one an only.... AUTRIJUS!
kgftr|konobi: how come?
autrijus vasc: aye, although a somewhat sleepy one. 09:55
vasc autrijus, oh?
no sleep this night yet?
erm, the night has come and gone :) 09:56
autrijus just woke up; didn't get quite enough sleep after the long stretch to 6.2.10
vasc i see
sorry to pester you, but i wanted to ask a really off-topic question
autrijus vasc: oh, you are part of freeciv team. thank-you for making a significant portion of my life disappear into hedonistic fun
only next to nethack
vasc autrijus, lol
autrijus offtopic question is fine 09:57
vasc i noticed you offer darcs and subversion access, i was wondering how you managed that?
autrijus vasc: currently, very crudely; I use svk to merge the svn.openfoundry.org one to svn.perl.org 09:58
and at the same time do a "darcs record"
so the darcs part is read only
nothingmuch vasc: the problem with sending darcs patches is that you get them back as conflicts 09:59
autrijus: you could try tailor.py - it's progressed since you started this
autrijus nothingmuch: yes, I know, and I figured out a way to creatively use tagging to track metadata after discussion with david roundy
but it's... not very high priority and a bit complex.
nothingmuch is tempted to use tables for layout
svnbot6 r7524 | rafl++ | Debian: 10:17
r7524 | rafl++ | * Update debian/smoker.yml to upload smoke results automatically.
r7524 | rafl++ | * Removed code from debian/rules that allows to disable smoke uploading using
r7524 | rafl++ | environment hooks. PUGS_BUILD_OPTS="upload_smoke=0" does the same.
r7523 | Darren_Duncan++ | Deleted the older ext/Locale-KeyedText, which ext/Rosetta-Incubator supersedes
rafl gaal: Why is Class::Rebless' Changlog empty? 10:27
nothingmuch because he's lazy? 10:29
;-)
that's why most of my changelogs are empty
or nonexistent 10:30
nothingmuch thinks he'll make Changelog a hexdump of /dev/random just to up his kwalitee
rafl Well, so the Debian package won't have an upstream changelog. So what.. 10:31
vasc autrijus, thanks a lot, and good luck for your project, not that i think you need luck though :)
theorbtwo nm: At least make your changelog a dump of the relevant portion of the svn log... 10:34
autrijus vasc: more luck couldn't hurt... I always make sure to carry two or more blessed luckstones around 10:35
rafl: you aware we bundle C::Rebless for PIL2JS?
rafl autrijus: Well, Actually.. no. 10:37
autrijus oh ok. there is a nonindexed copy in PIL2JS/lib
so we don't strictly need to depend on CR
rafl Really. Why do we ship it with PIL2JS? 10:38
It wouldn't be a dependency anyway. I'd only suggest spidermonkey-bin and class-rebless. It's not needed for pugs to work.
autrijus rafl: because dduncan reported that the warning is confusing and it's best to have pil2js prelude uniformly generated (or not) 10:39
and gaal consented the bundling as C::R is small anyway
rafl Why not put it in PREREQ_PM?
autrijus for the next release, maybe. 10:40
but it was late and we didn't have any dependency for CPAN modules before for the default configuration
rafl OK, so I won't add the suggestion to the debian package.
autrijus k. 10:41
rafl We don't bundle spidermonkey yet, do we? ;-) 10:43
Btw, we have a to YAML. 10:45
Juerd nothingmuch: In fact, speed shouldn't be considered relevant even (re dvorak) 10:46
Yes, it's faster
autrijus rafl: mm, we bundle a MiniYAML 10:47
Juerd But a better goal is to type as fast as before with less movement
Speed will then automatically improve.
And as much as your body thinks is natural
rafl autrijus: Is it really used by all tools? I don't think so. (I can't think of an other reason to put it in my depends line) 10:48
autrijus rafl: yeah, smoke needs yaml writing 10:55
bbiab... 10:59
pasteling "scook0" at 220.253.96.162 pasted "Parser patch Mk III" (147 lines, 5.9K) at sial.org/pbot/13598 11:03
scook0 nothingmuch: if you're willing to test, there's a new patch up
gaal I think we should just put perl 6.0 in PREREQ_PM and be done with all this messy implementation stuff. 11:10
What? We wanted it to be self-hosting, no? Might as well be self-dependent.
</silly> 11:11
rafl: you can Recommend: Bundle::Pugs::SmokeKit
nothingmuch scook0: yes
but i'm off to lunch, so i'll report when I'm back 11:12
gaal that's yaml, petal, test-tap-* and stuff
scook0 nothingmuch: k
gaal as for a changelog... point. I actually have a pending bugfix to C:R (one line patch), so I should contemplate... 0.04! 11:13
rafl gaal: I would have to package that then. I don't like the idea of packaging bundles. 11:23
gaal: Don't to that much releases. That work for the packagers (me!) :-) 11:25
gaal: You were using Debian, right?
gaal rafl: using debian when? i do use debian 11:27
but not for building pugs, as my debian box isn't ramful enough 11:28
rafl gaal: Great. So you can test both the Debian and the C:R packages. :-)
gaal :-) yes indeedy
are they in unstable already? 11:29
...or d'ya want to let me dpkg them manually?
i mean dpkg -i
rafl No, not in unstable yet.
feather.perl6.nl/~rafl/debian
gaal sec
rafl pugs 6.2.10 will be there soon. 11:30
gaal i'm off to a meeting soon... 11:31
rafl The packages will wait for you. :-) 11:32
Nooooo! ./usr/share/perl6/5.8/FA/t/DFA.t 11:36
nothingmuch ouch 11:42
scook0: no go 11:45
rafl I can't see why it gets installed there. Anyone?
scook0 nothingmuch: curses
was it something really stoopid?
(woo, nautilus crashes again!) 11:47
rep why do you use nautilus anyway 11:49
scook0 because nautilus comes with gnome
and gnome comes with ubuntu
and I can't be bothered setting anything else up
rep i have gnome, and yet i don't use nautilus!
(nor anything else) 11:50
scook0 I don't mind nautilus...except when it crashes
rafl: there seems to be an extra copy of t/ inside lib/FA/ 11:52
svnbot6 r7525 | rafl++ | * Removed extra copy of t inside ext/FA-DFA/lib/FA. scook0++ 11:55
rafl For some reason perl5 contains Makefile.old stuff after doing a realclean. Why?
The makefile hooks look right.
src/gen_prelude* and src/Main* also doesn't seem to be cleaned properly. 11:58
svnbot6 r7526 | rafl++ | Debian: 12:07
r7526 | rafl++ | * changelog: Added changes for 6.2.10.
r7526 | rafl++ | * pugs-modules.lintian-overrides: Added overrides for new modules.
r7526 | rafl++ | * rules: clean up the source tree properly.
r7526 | rafl++ | * rules: don't install ./usr/share/perl6/5.8/FA/t/
r7527 | rafl++ | Debian:
r7527 | rafl++ | * rules: removed the smoke uploading code. make smoke handles it.
rafl Does the smoke uploading using smoker.yml really work? I see smoke_upload => 0 when building. 12:13
nothingmuch scook0: sorry 12:19
the error is exactly the same as last time 12:20
unexpected "'", 01-tap.t, line 7 column 5
scook0 weird
because mine seems to work
nothingmuch odd
oh shit
i get it
when I redownloaded the patch I accidentally applied the wrong file 12:21
=(
scook0 phew
nothingmuch is an idiot
scook0 it happens
anyway, I'm going to bed soon 12:22
nothingmuch ciao then
scook0 so if you do run the test, either paste the results (and I'll backlog), or else email me
unfortunately, I'll probably be super-busy until at least Thursday :( 12:23
ciao
svnbot6 r7528 | rafl++ | * debian/rules: Removed special handling for ext/FA-DFA/. The tests won't be 12:45
r7528 | rafl++ | installed accidently anymore.
gaal rafl: any prefered tests other than just dpkg -i ? 13:11
rafl gaal: I don't know other tests for you module. For pugs (currently smoking) it should be tested if the backends work (needed modules/precpm prelude, etc. are found) and if the haskell module is usable. I don't know that much about hs. 13:13
xerox rafl: what does smoking means in that context? 13:14
rafl xerox: It run's make smoke. 13:15
xerox What does it do, if I could ask? 13:16
rafl It's quite similar to make test (runs the test suite) but outputs a nice html file which will be uploaded to smoke.pugscode.org afterwards. 13:17
xerox Thanks.
clkao 8
gaal rafl: Class::Rebless is OK 13:18
the tap-* modules want petal, but you know THAT needs URI
and MKDoc::XML
but that's it :)
yaml presumably is already packaged for debian 13:19
rafl gaal: What needs URI? 13:22
gaal: Because I also maintain MKDoc::XML and Petal (I adopted them recently) I know that both depend on everything that's specified in PREREQ_PM. URI is not listed there. 13:24
gaal rafl: hmm, i thought petal did, checking if it's perhaps something else 13:37
d'oh, Test-TAP-HTMLMatrix does. 13:38
rafl Depends: perl (>= 5.6.0-16), libtest-tap-model-perl, libpetal-perl, liburi-perl 13:39
rafl doesn't forget dependencies!
gaal well, okay then ;) 13:40
rafl just found out that he maintains he didn't knew about.. 13:48
eric256_ snow! 14:09
gaal eric256_: where? 14:10
eric256_ Denver Colorado...our first snow since i moved here ;) 14:11
rafl In the fridge?
eric256_ out of 12 people in the office only 2 are here...cowards
gaal four years ago, almost to the day, i saw the first snow in a lonely place in mongolia
eric256_ of course our power is coming and going so i might get to go home...or just because no one else bothered to come. lol
i also understand for the first time in my life *why* trees loose there leaves.. the ones that havn't lost there leaves are just getting beat on. 14:12
rafl gaal: Uploading the packages. 14:13
qwr move somewhere more north and soon you'll discover that there is nothing exciting about the snow ;)
eric256_ hehe 14:14
rafl gaal: done. 14:16
gaal k 14:17
ingy_ autrijus: done (pugs.kwiki.org/) 14:18
autrijus egg selent
ingy++
ingy_ I've been offline for 5 days 14:19
xinming_Beijing ?eval 1+2;
evalbot_7528 3
ingy_ well actually I've been pretty offline for 6 months, eh
xinming_Beijing ?eval {eval "1 + 2"}.say;
evalbot_7528 <SubBlock(<anon>)> bool::true
xinming_Beijing hmm, shouldn't it be 3?
rafl ingy_: You can also kick nwp to finally upload kwiki to Debian. 14:20
ingy_ on the upside I just started inventing a new database ;)
xinming_Beijing eval "1 + 2" returns 3, and it might be equal to 3.say
?eval 3.say
xerox ?eval 1+1
evalbot_7528 3 bool::true
2
xerox What is '.say'?
ingy_ xerox: a printing method 14:21
xinming_Beijing xerox: "string".say is equal to `print "string\n"`
xern: understand? :-)
rafl And equal to say "string".
xerox Sure.
xinming_Beijing hmm, certainly, "string".say can also be written in "string\n".print form.
xerox ?eval say "thanks"
evalbot_7528 thanks bool::true
xerox xinming_Beijing: gotcha. 14:22
xinming_Beijing xerox: hmm, It's a OO like thing in perl 6, which named `indirect notation`\...
xerox ?eval say "eurt::loob <->"
evalbot_7528 eurt::loob <-> bool::true
xinming_Beijing hmm, maybe, I might be wrong for the name.
autrijus ingy_: what, you extending NIH to the database arena? 14:23
ingy_ autrijus: yeah kinda
autrijus ingy++ # the knight who say NIH
xinming_Beijing Ok, Can anyone here answer my question? what will `{ eval "3+4" }.say` print? hmm, I might go and write the test...
ingy_ it's not an sql db
although someone will write a dbi wrapper I'm sure 14:24
gaal xinming_Beijing: not 100% sure, but { something }.say is asking for say on a closure
xinming_Beijing In fact, "{ eval '3 + 4' }".say will do the right thing, But I am not sure about the closure without double quota...
ingy_ autrijus: I've learned to not be ashamed and to truly embrace my NIH syndrome
autrijus ingy_: yes, use your uniqueness to its advantage 14:25
gaal {7} is a closure too
which is why <SubBlock(<anon>)> looks about right.
xerox ?eval {}.say
evalbot_7528 <SubBlock(<anon>)> bool::true
rafl gaal: So, do they work? :-) 14:26
gaal checking :)
eric256_ hmm i signed off...and yet i'm still here 14:27
rafl Some have circular deps. You should install them together.
xinming_Beijing hmm, I think, say will accept the string and numeric context, And closure should be "eval"ed first... just MHO...
gaal rafl: look good so far 14:28
xinming_Beijing bool context will automatically "upgrade" to numeric context.
:-
rafl gaal: Also libghc6-pugs-dev? I could not test it so far. 14:29
gaal p5embedd works 14:31
rafl gaal: haskell eval should not. hs-plugins isn't uploaded yet. 14:32
gaal parrot is not embedded?
rafl No, no parrot in Debian. Working on it, though.
gaal ahh
ok
rafl It's kind of much work, so I focused on pugs. 14:33
pasteling "gaal" at 192.117.178.34 pasted "libghc6-pugs-dev install fail" (10 lines, 547B) at sial.org/pbot/13602 14:34
gaal rafl: sure
rafl Hm, I think I broke libghc6-pugs-dev. :-( 14:36
Bah, I need to adjust the paths in installed-pkg-config. 14:39
xinming_Beijing rafl: hmm, Is there any reason why GHC in debian still 6.2? even in the sid version... 14:40
rafl xinming_Beijing: 6.4-4.1 is in sid. 14:41
xinming_Beijing: 6.4.1-1 will go in as soon as 6.4-4.1 is in sarge.
xinming_Beijing: With ghc 6.2 we couldn't even compile pugs. 14:42
xinming_Beijing: See packages.debian.org/ghc6
What run's ./Setup configure when building pugs? 14:48
xinming_Beijing rafl: :-S, I ever tried to compile pugs in sid, But failed. hmm, and IIRC, the reason of fail is because of 6.2 :'(
rafl xinming_Beijing: sid has 6.4, even on all architectures. See the link I've posted. 14:49
xinming_Beijing rafl: I saw.. I remember wrong. :-)
rafl: hmm, util/build_pugs.pl 14:52
rafl: maybe.
rafl xinming_Beijing: Right.
xinming_Beijing after you perl Makefile.PL, you will find `$(PERL) util/build_pugs.pl _+SETUP --prefix=$(DESTDIR) ...` 14:53
in Makefile
rafl Right. ./Setup configure is called with the args between _+SETUP and _-SETUP. 15:03
offby1 to what number are the Pugs release numbers converging? 15:21
2 * pi ?
iblech offby1: Yep. 15:22
offby1 forgive me, but I don't see how 6.2.10 fits into that sequence ... 15:25
iblech offby1: Only the first two parts of the version number (6.2) converge to 2pi. The last part is simply incremented (6.2.8, 6.2.9, 6.2.10 --> 6.28.0, 6.28.1, etc.) 15:26
coral heh
offby1 ah, that explains it! 15:28
offby1 thought he'd forgotten what pi was
eric256_ tries to decide if he is hungry enough to walk and or drive in the snow. lol 15:32
offby1 decides for him: yes 15:34
gaal autrijus: just noticed a frivolous wkifification of *] and SQL::Routine[* in the release announcement on your journal. 15:39
offby1 do the various examples in the pugs source all work, or are some of them suffering bitrot? i.e., if I want to learn more about programming in perl6, are they a good place to look? 15:44
iblech offby1: They are definitely worth reading. Also look at the modules in ext/ and the testsuite in t/ 15:46
integral if they've suffered bit rot, learn perl6 and fix them =)]
iblech ?eval "Also note that the sourcecode to the evalbot is in examples/network/evalbot/ (with the necessary module in ext/Net-IRC/)"
evalbot_7528 'Also note that the sourcecode to the evalbot is in examples/network/evalbot/ (with the necessary module in ext/Net-IRC/)'
iblech (as an example of code which definitely works) 15:47
rafl Someone should continue to work on XML-SAX. I want to write the same for Jabber chats. :-) 15:48
offby1 integral: I wouldn't mind, but until I'm pretty good with perl6, I have a chicken-and-egg problem: can't learn perl6 without good examples; can't make the examples good until I know perl6 15:49
iblech offby1: Luckily, #perl6 can help you :) 15:50
rafl: ooh :)
integral offby1: iirc most of the programs in examples/ itself do work, but some of the subdirs are more dodgy. 15:51
offby1 integral: OK, thanks.
gaal fr 15:54
er. 15:55
rafl: have a new package for me to test?
rafl gaal: building. It runs the tests. My hardware isn't that fast. 15:56
gaal np :) 15:57
btw (probably related) 15:58
dpkg -P libghc6-pugs-dev
ghc-pkg: cannot find package Pugs
rafl gaal: Yes, because postinst could not register it with ghc-pkg. 16:00
gaal k
eric256_ offby1 - if you find stuff that doesn't work let us/me know. I'll be happy to help you learn/repair examples 16:01
;)
16:01 eric256_ is now known as eric256
offby1 eric256: is your email address in some file in the examples directory, or are you reachable only by IRC? 16:02
eric256 IRC is probably best since then ther are lots of people to help 16:04
perlbot nopaste 16:05
perlbot Paste your code here and #<channel> will be able to view it: sial.org/pbot/<channel>
pasteling "eric256" at 66.102.136.66 pasted "nmake error" (17 lines, 924B) at sial.org/pbot/13604
eric256 i'm gettnig that error "Can't locate YAML.pm" when trying to nmake 16:06
gaal eep 16:07
eric256 pugs still seems to run and work but i got that error....is it safe to ignore?
gaal it is if you don't need the JS backend
i didn't think it depended on yaml :(
iblech eric256: Fixed in r7529 16:08
gaal: Right, it was just for debugging ("use YAML; die Dump(...)")... :(
svnbot6 r7529 | iblech++ | * Usual svn props.
r7529 | iblech++ | * PIL2JS: PIL::App: Removed some debugging leftover, eric256++ for noticing.
gaal ah
iblech Luckily make didn't abort after the failed runjs.pl call 16:10
putter, fglock: FYI, running make smoke-perl5 here currently 16:14
eric256_ bahhh...stupid power....and irc thinks i'm still there. lol
iblech 18:08 < iblech> eric256: Fixed in r7529 16:15
eric256_ thanks
trees havn't lost there leaves, so they are breaking and taking out power lines ;) 16:16
16:30 typester is now known as sleepster
rafl e/wi 44 16:32
eric256_ much better. 16:38
16:39 eric256_ is now known as eric256
fglock__ putter: ping 16:50
svnbot6 r7530 | fglock++ | * perl5/Code - Currying:
r7530 | fglock++ | sub x ( $a, $b ) { say "$a + $b" };
r7530 | fglock++ | $z = &x.prebind( 'a'=>1 );
r7530 | fglock++ | $z( b=>2 );
r7530 | fglock++ | * BUG: Code.pm doesn't implement named arguments properly, so this prints Pairs instead of Values
eric256 any idea why a failing open is raising an error refering to a location in Prelude.pm ? 16:51
iblech fglock__: oooh ++ (but note that prebind is spelled assuming in P6)
eric256: Probably because &open is implemented in the Prelude
fglock__ iblech: I'll fix it - thanks 16:52
eric256 string length is still length right? 16:53
PerlJam eric256: bytes for length in bytes, chars for length in chars, codes for length in codepoints, etc. 16:54
eric256 ahhh
thanks
PerlJam I think we have bytes, chars, codes, langs but I'm not sure about the last 2 really as I never use them.
eric256 ?eval $*PROGRAM_NAME 16:55
evalbot_7530 \undef
svnbot6 r7531 | fglock++ | * perl5/Code - s/prebind/assuming/ iblech++
eric256 ?eval $*PROGRAM_NAME.perl
evalbot_7530 '\\undef'
iblech eric256: $*PROGRAM_NAME is (as well as %?CONFIG etc.) hidden in safemode 16:56
eric256 odd ... motd.p6 uses splitpath($*PROGRAM_NAME)[1];
bit warn .perl on it produces 16:57
*** \\MkMatch {matchOk = True, matchFrom = 0, matchTo = 0, matchStr = "", matchS
ubPos = [], matchSubNamed = {}}
which isn't what that program is expceting, its trying to get the directory...any ideas?
rafl evalbot_7531: You're serializing a regex match. That's fine.
PerlJam Isn't the .perl method for generating a representation as perl code?
iblech eric256: ~$foo will get you the string
eric256 ahh... problem was perl didn't seem to ever be ~ it automaticaly 16:58
fglock__ iblech: it works with junctions too :)
rafl Or "$foo"
iblech PerlJam: It is. Extending t/builtins/perl.t.
fglock__ iblech: sub x ( $a, $b ) { say "$a + $b" }; $z = &x.assuming( 'a'=> 1|3 ); $z(b=>2)
iblech fglock__: oooh! Does vanilla Pugs support this too? 16:59
fglock__ iblech: I implemented it in perl5/Code.pm - but it is quite simple <15 lines 17:00
eric256 my $progdir = splitpath($*PROGRAM_NAME)[1] || '.';
svnbot6 r7532 | iblech++ | t/builtins/perl.t: Tests for .perling rx objects. PerlJam++ eric256++
iblech fglock__: looking 17:01
eric256 doesn't work at all because its not stringifing the match
fglock__ iblech: please note that there is a bug in named args in p5, so the results are a bit weird
iblech eric256: Hm? "a" ~~ rx:P5{b} || "foo" seems to work 17:02
fglock__: nod
eric256 could it be something with the File::Spec module? the line i posted leaves '$progdir' eq '' 17:04
so i added a line below it $prodir = '.' unless ~$progdir;
which isn't exactly pretty
BTW: how would i clone a hash? 17:05
iblech Ah, I think I know why: Although the match is "" (empty string), it did successfully match
so || chooses the LHS 17:06
eric256 hehe "" but true
iblech not sure whether this is a bug
right, exactly
[clone a hash] %hash.clone (but probably not yet implemented)
fglock__ iblech: yes - pugs supports it!
eric256 while it might not be a bug, it definitly is not DWIM
maybe splitpath should stringifiy before returning?
iblech eric256: I think so. 17:07
eric256 should i just leave the workaround in for now? or change File::Spec.pm?
iblech eric256: I'd fix File::Spec 17:08
fglock__ ?eval sub x ( $a, $b ) { $a + $b }; my $z = &x.assuming( 'a'=>1|2 ); $z(2|3)
evalbot_7532 3
iblech ?eval sub x ( $a, $b ) { $a + $b }; my $z = &x.assuming( 'a'=>1|2 ); $z(2|3).perl 17:09
evalbot_7532 '((3 | 4) | (4 | 5))'
PerlJam eric256: What's wrong with %a = %b; for cloning a hash? 17:10
iblech PerlJam: %a = %b doesn't clone deeply (%b<X> = [...]; %a = %b; %a<X> = ...; %b<X> changed as well) 17:11
eric256 doesn't seem to work...well it works, but its yelling odd number of elements
actualy they are doing. my %copy = hash %x;
?eval my %x = (x => 1, y =>2); my %copy = hash %x;
evalbot_7532 Error: cannot cast from VUndef to Handle (VHandle)
iblech ?eval my %x = (x => 1, y =>2); my %copy = %h 17:12
evalbot_7532 Error: Undeclared variable: "%h"
iblech ?eval my %x = (x => 1, y =>2); my %copy = %x
evalbot_7532 {('x' => 1), ('y' => 2)}
eric256 scratches his head
iblech The extra hash() should not be necessary
eric256 k...dunno why they had it in the first place ;)
not even sure they realy need a copy in this spot
PerlJam iblech: so .clone() is a recursive copy? i.e., it call .clone() on all the elements of the data structure until it hits the leaves? 17:14
iblech PerlJam: Yep. 17:15
geoffb morning, all . . . 17:16
svnbot6 r7533 | eric256++ | Fixed typos in motd examples, and strigified output of File::Spec::Win32::splitpath()
iblech Evening, geoffb :)
geoffb How have you been, iblech? 17:17
iblech fglock__: Implemented assuming in PIL2JS :)
fglock__ iblech: cool!
iblech geoffb: Quite good, hurray for 6.2.10 :)
geoffb good, and: Oh YEAH! 17:18
And cool that PIL2JS is getting assuming, btw
eric256 realy isn't karma whoring, just forgot that SVN doesn't go up the tree from where you issue the command. ;)
iblech geoffb: 4 lines :) (but there're probably bugs WRT slurpy hashes/arrays in there) 17:19
svnbot6 r7534 | eric256++ | Strigified output of File::Spec::Win32::splitpath()
r7535 | eric256++ | Strigified output of File::Spec::Unix::splitpath()
geoffb iblech, cool! nodnod
svnbot6 r7536 | iblech++ | PIL2JS: &assuming!
r7536 | iblech++ | * PIL2JS.pm: Added lib6/ to the sub-pugs's @INC, so I don't have to type "make"
r7536 | iblech++ | all the time.
r7536 | iblech++ | * Prelude::JS::Code: Implemented (a simple form of) &assuming.
fglock__ iblech: same problem in p5, I used a *@_ signature 17:20
iblech fglock__: Yep, exactly.
autrijus oooh assuming! 17:21
geoffb autrijus, thanks for posting good comments to /. while I was slacking with the family yesterday . . . 17:23
autrijus geoffb: well, I couldn't stay up long enough to track the discussion... although it went pretty well
the release note got +4 informative so I'm happy
geoffb autrijus++ 17:24
autrijus I'm less happy about the goatse thing on my wikipedia page.
but well, c'est la vie.
geoffb autrijus, no kidding
There's always that fraction of a percent of people.
autrijus it's my fault of redirecting autrijus.org to wikipedia anyway.
but why it could've stayed up for 6 hours without somebody fixing it is beyond me.
geoffb On a planet this big, one in a million means 6,000 people . . . .
autrijus nodnod
elmex elmex++ 17:25
perlbot What kind of idiot karmas himself? Your kind of idiot!
geoffb Well, despite the microwave/internet generation expectation, things aren't always instant, sadly
geoffb wants the chicken cooker from Fifth Element 17:26
elmex wants the chicken from Fifth Element :)
autrijus wants the Fifth Element :)
geoffb heh
elmex ok, rooted ;)
autrijus pwnerz 17:27
elmex * does evaluate it's operand as if it was written 'literally' ?
autrijus elmex: hrm?
elmex well, the '*' operator in perl
6
(i'm following the pair semantics discussion on perl6 mailinglist) 17:28
iblech elmex: my @a = ($foo, $bar, $baz); foo(*@a) is same as foo($foo, $bar, $baz)
autrijus okay. time to 1)work on eurooscon talks, 2)get pneuma transcribed to -BPIL2, and 3)write up some documentation for $work
mmm lightning talk submission deadline for euroscon is today.
elmex iblech: yes, but what does '*' do?
geoffb elmex, it's saying, expand this container into several arguments, rather than passing a ref to the container 17:29
elmex geoffb: ah, thanks
geoffb np
elmex now i understand very much perl6 very much better
Khisanth so that is why it's a * 17:30
eric256 could someone try running cribbage_scoring.p6 in examples....its giving me a weird error, i think its a parse error of some sort
autrijus because it resembles a "splat" when you stomp into a container and make it scatter around 17:31
Khisanth yeah :)
who came up with that?
autrijus larry I think, but it may be imported from ruby 17:32
wolverian and ** is "stop on it twice"
er, stomp
eric256 ** = stomp on it with both feet ;)
autrijus eric256: look at line 6
eric256 lol. thanks
eric256 learns to *read* before asking stupid questions
wolverian oh, I'm starting a programmer's editor project in perl6 17:33
if anyone is interested feel free to mail me at [email@hidden.address]
(well, not promising actual progress, but hey, it's fun) 17:34
Khisanth yet another editor?
PerlJam wolverian: Just emulate autrijus and interested people will show up
geoffb Interested, yes. Can contribute any tuits whatsoever, sadly no. If you make any progress, let us know, it would be fun to try ou
wolverian Khisanth, its name is currently 'none'
eric256 is there a p6 editor already? hmmm i need a p6 highlighter for Context ;)
wolverian not yae :)
now, have to go, later :)
Khisanth "yeah" would be better
it will create some confusing conversations! 17:35
PerlJam yaep
:)
Khisanth p?
PerlJam yet another editor project 17:36
17:44 rep is now known as reppie, reppie is now known as rep
eric256 ?eval my @x = (1..5); @x.map:{$_++}; 17:47
evalbot_7536 Error: Can't modify constant item: VInt 1
eric256 ?eval my @x = (1..5); @x.map:{$_ +1};
evalbot_7536 (2, 3, 4, 5, 6)
eric256 ?eval my @x = (1..5); @x = @x.map:{$_ +1};
evalbot_7536 [2, 3, 4, 5, 6]
iblech eric256: IIRC a pugsbug causes that your first ?eval doesn't work (it does work in PIL2JS, FYI), see t/builtins/lists/mutating_listops.t 17:50
PerlJam btw, that reminds me ... at least in my version of pugs, this generates an error: my %h = map { $_ => 1 } @stuff; now I know why and I know of at least one way to fix it, but this is a relatively common idiom I think. Surely little things like this are bound to cause friction in the p5 crowd. 17:52
eric256 yea i was trying to get inf_lazy_list.p6 to work...but it is broken beyond my abilities or knowledge ;)
geoffb (from backlog:) what is NIH in ingy_ context? 17:53
eric256 monty python skit 17:54
PerlJam geoffb: Not Invented Here.
:-)
eric256 or that. ;)
geoffb Knew of those expansions, just didn't know what ingy_ and autrijus were on about . . . you never know with those two
iblech PerlJam: notabug -- { ... => ... } is taken as hashref constructor. You've to disambiguate: {; ... => ... } or { ... => ...; } or { (... => ...) } 17:56
PerlJam iblech: I understand. It still rankles a little bit. 17:57
eric256 would @stuff.map:{ $_ => 1} work? 17:58
nope. 17:59
my %x = map { $x, 1 } @stuff; works....but doesn't seem right
iblech eric256: It should -- :{...} always introduces a slurpy block (which isa Code, never a Hash) 18:00
eric256 doesn't though
PerlJam Well, that's a pugsbug then 18:01
Khisanth so my %h = map :{ $_ => 1 } @stuff; would be ok?
eric256 ?eval my @s = (1..5); my %x = @s.map:{$_ => 1};
evalbot_7536 Error: cannot cast from VList [VList [VStr "perl6",VInt 1]] to Pugs.AST.Internals.VCode (VCode)
PerlJam Khisanth: should be.
Khisanth PerlJam: well then the : seems pretty minor
thomason newbie q: has anybody encountered 'Can't exec "./Setup": No such file or directory at util/build_pugs.pl line 49'? 18:02
eric256 lunch &
PerlJam Khisanth: almost all of the ways to tell perl the block is a code block rather than a hash ref seem pretty minor if you ask me. Except perhaps %h = map sub { $_ => 1 } @stuff;
iblech thomason: Hm, shouldn't happen. Could you nopaste the output of "perl Makefile.PL; make"? 18:04
thomason aye
Khisanth will just wait for "Learning Perl 5th edition" :)
thomason nopaste?
hrm, doesn't work here
iblech perlbot nopaste
perlbot Paste your code here and #<channel> will be able to view it: sial.org/pbot/<channel>
thomason ah
pasteling "thomason" at 169.230.87.20 pasted "build failure" (157 lines, 10.2K) at sial.org/pbot/13608 18:06
autrijus Khisanth: surely it needs to be "6th edition" 18:08
iblech thomason: Ah. Your GHC installation seems to lack the Distribution.Simple module, I think
autrijus thomason: your ghc has no cabal.
you sure it is 6.4+?
thomason ah 18:09
Glasgow Haskell Compiler, Version 6.4.1, for Haskell 98, compiled by GHC version 6.4
stock gentoo build
geoffb thomason, what distro?
Someone had said that some distros package cabal separately
thomason gentoo unstable
Khisanth autrijus: but 4th edition is suppose to the the last one for p5!
thomason ah, yup, cabal is a separate package
thanks
geoffb Khisanth, when is 4th due?
np
Khisanth isn't 4th already out? :) 18:11
geoffb Speaking of which, I wonder when the next perl build is going to flow through to etch . . . last I checked it was blocking stuff
autrijus mm bbiab.
Khisanth www.amazon.com/exec/obidos/tg/detai...0?v=glance :)
geoffb Oh, duh, I thought you'd said Programming Perl
dammit, it is just moronic for a package to be available in debian oldstable, stable, and unstable, but not in testing 18:14
stevan autrijus: ping 18:16
howdy geoffb, everyone
geoffb Grrr, I love Debian's pro-portability attitude and all, but holding up something huge like perl based on failed arm builds rankles
Khisanth I guess Programming Perl 4th editio nwill cover perl6 :)
geoffb howdy, stevan
stevan geoffb: nice article/blog :)
geoffb stevan, why thank you! :-)
stevan you are now the official Pugs scribe :) 18:17
geoffb The /. post has made it far and away the most popular thing I've ever posted
stevan, gee, no pressure or anything. :-)
fglock__ hi stevan
stevan geoffb: no, none at all ;)
hey fglock__
I saw the currying stuff in the backlog,.. very nice 18:18
fglock__ :)
stevan has been reading too much LISP lately,.. have to fight the urge to switch .......
geoffb yes, fglock__++ and iblech++ for currying
stevan well (Haskell Curry)++ and (Alonzo Church)++ too 18:19
svnbot6 r7537 | iblech++ | t/subroutines/currying.t: Fixed the test: 18:20
r7537 | iblech++ | sub foo (+$a) {...}
r7537 | iblech++ | foo 42; # should fail
r7537 | iblech++ | sub foo (?$a) {...}
r7537 | iblech++ | foo 42; # ok
r7537 | iblech++ | foo a => 42; # ok too
Aankhen`` goes to sleep.
iblech Night, Aankhen`` :) 18:21
Aankhen`` G'night.
iblech And hi stevan :)
stevan hey iblech, long time no see
iblech searches the lastest JS smoke for a not yet implemented feature he could implement 18:23
geoffb :-)
leo__ autrijus: ping 18:24
geoffb leo__, he's around, but not instantly, as (last I saw) he was doing more last-minute talk writing 18:25
Juerd Please, if someone disagrees with your proposal, do more than repeat the proposal.
geoffb I'm going to assume you're referring to something on p6l, Juerd? 18:26
Juerd I've read "no, in this proposal, ..." too many times now. These posts aren't trying to add to it or further specify, they are meant for change. 18:27
Yes
The pair thread
I like the underlying idea very much, but discussion is killed by repetition, and I hate that.
But that's not something to discuss on the list
geoffb sigh . . . and that's why I rarely read mailing lists any more 18:29
Juerd Very few people are good at discussing things online 18:30
I wish I were part of those few, but I can try 18:31
svnbot6 r7538 | stevan++ | Perl6-MetaModel - just removing drop shadows, they are so early 90s
r7539 | vladtz++ | Script to annotate a tests.yml with older versions.
r7539 | vladtz++ | This will make it easy to find places where the sources are in flux.
autrijus leo__: pong
leo__ meeting
autrijus oh yes, meeting!
Juerd Speaking of meeting - who will be at the extra amsterdam.pm meeting the 20th? 18:32
geoffb stevan, what do you use to create your diagrams?
iblech stevan: According to the file extension, graffle 18:33
Juerd: re (p6l post from you) "Which leaves us with no basic datastructure that can hold both positional and named arguments." -- right, this is why we need luqui's tuple proposal, see svn.openfoundry.org/pugs/docs/notes/theory.pod 18:34
Juerd iblech: I think that then I don't understand the tuple proposal :) 18:35
iblech iblech: Basically, a tuple is a data structure which can hold positional args, named args, slurpy blocks etc.
And *$tuple will DWIM. Examples:
my $tuple = (a => 1, (b => 2)); 18:36
Juerd I see
iblech foo *$tuple; # named arg "a", positional pair (b => 2)
Juerd It's scary to have special syntax on the RHS of assignment, based on the type of the LHS.
I'd prefer this to be in the constructor, which is of course already a call, and thus suited.
Hm, what is a tuple anyway? Is it a type? 18:37
Created by syntax?
iblech You mean like my $tuple = Tuple.new(a => 1, (b => 2))? I'd like this too
Yep.
But tuple are not special until splatted -- so you can, for example, store tuples in arrays etc. 18:38
Juerd It's vague to me how tuples are declared, defined... 18:39
cognominal autrijus, I /msged obra as a question for #parrotsketch but he said to ask here "can we use pge from pugs with Perl6 syntax?"
iblech Under luqui's current proposal, my $tuple = (42,) creates a tuple, $tuple.isa(Tuple) then.
stevan geoffb: it is a Mac OSX app called OmniGraffle
Juerd iblech: Because of the comma?
autrijus cognominal: you mean {} callback in pge?
geoffb stevan, yeah, with iblech's hint, I googled. 18:40
Juerd iblech: i.e. the comma in scalar context is a tuple constructor?
iblech Juerd: Yep. But, as you said above, this conflicts with , as being the list/array constructor
cognominal autrijus: I don't know what you meant, I meant just using perl6 regexen
stevan autrijus: when you have a moment I would like to discuss eigenclasses with you
autrijus cognominal: can you speak in code? :)
geoffb cognominal, if you
Juerd iblech: I have to think about this :)
autrijus cognominal: you can use /...p6rule.../ in pugs already
Juerd iblech: Thanks for further explaining
geoffb cognominal, nevermind, autrijus just said what I was going to 18:41
autrijus stevan: okay, wake for the parrotsketch to finish (in half a hour)
cognominal autrijus: but does it use pge?
autrijus cognominal: sure does.
cognominal oki, thx
stevan autrijus: ok, sounds good
iblech Juerd: No problem :) 18:42
Juerd: I'll reply to your p6l post, so the rest of p6l is informed as well
PerlJam Are Tuples officially part of p6 now? 18:48
iblech PerlJam: At least @Larry hasn't posted a "tuples are official P6 now" message to p6l 18:49
autrijus ...and even if they did, until they make into the svn repo for Synopses, that is still not canonical 18:53
(by definition, canonical = part of canon)
geoffb Is the P6::Bible still suffering from severe lag WRT p6l, or is that fixed now? 18:55
PerlJam autrijus: what's the topic of parrotsketch today? 18:56
autrijus ingy_: make some of us comaintainer of P6::Bible?
PerlJam: www.parrotcode.org/misc/parrotsketc...h.20051010
PerlJam ah, excellent. that might just save me from even more IRC-induced madness 18:57
Or it might encourage me to join the channel and do slightly more than lurk ;)
iblech Hm. How can I write chr(12) ~ "3" in a double-quoted string using backslash escapes? "\123" does, obviously, not work 19:02
I seem to recall that "\d[12]3" should work, but unsure... 19:03
Khisanth \0123? 19:04
iblech This includes the 3 again...
I want "\12" ~ "3", but without the need for ~ and two strings... 19:05
geoffb ?eval "{\012}3"
evalbot_7539 '{}3'
geoffb Well, that's interesting
?eval "{ \012 }3"
evalbot_7539 '{ }3'
geoffb ?eval "{ 1 + 2 } 3" 19:06
evalbot_7539 '3 3'
geoffb ?eval "{ 1 + 2 }3"
evalbot_7539 '33'
Khisanth aren't you trying to get "<form feed char>3" ?
19:06 wilx` is now known as wilx
iblech Khisanth: I am. (But the 12 and 3 are only examples) 19:06
geoffb I'm thinking there's a bug with parsing "{ ... }" 19:07
Or building the string from it, at least
Khisanth ?eval "\0123"
evalbot_7539 '{'
Khisanth ...
where the heck did the { come from? 19:08
iblech ?eval chr(0123)
evalbot_7539 Error: unexpected ( or "c" expecting program 0100 is not octal in perl6 any more, use 0o100 instead.
iblech ?eval chr(0o123)
evalbot_7539 'S'
iblech ?eval chr(0d123)
evalbot_7539 '{'
Khisanth iblech: hmm well I am expecting p5 behavior there
geoffb ?eval "{ 0d12 }3" 19:09
evalbot_7539 '123'
geoffb ?eval "{ chr(12) }3"
evalbot_7539 '3'
geoffb ?eval "{ qq{\012} }3" 19:10
evalbot_7539 '3'
iblech geoffb: Thought so too, but no! {...} in double-quoted strings starts a closure which will later be interpolated. \012 in normal expression context evaluates to a reference to 12. (But references to scalars shouldn't automatically deref, but that's another topic.)
geoffb TIMTOWTDI, and they all suck
iblech, sure, but "{ ... }" should not result in the { } being in the output string 19:11
iblech geoffb: Oh, right!
geoffb: Write a test please? :)
geoffb iblech, not right now. I'm having severe problems concentrating on $work as it is. 19:12
-Ofun it is most definitely NOT
iblech geoffb: Sure, I'll write it then
geoffb thx muchly
iblech np
svnbot6 r7540 | iblech++ | New t/syntax/parsing_braces_in_strings.t: "{\01;42}" should eq "42", not "{;42}". 19:19
autrijus okay, got two topics for Euro Oscon lightning talk 19:20
1. "Visual Basic rocks!"
2. "Larry was a mariner"
geoffb 3. "Autrijus is punchy!" 19:21
autrijus :)
dduncan go for it! 19:22
while you're at it, demonstrate how fun it is to implement perl 6 in visual basic
autrijus dduncan: good, I'll do that.
the 2. would be a full adaptation/parody of "Earendil was a mariner"... easily my fav poem
and allison has agreed to do it with me :) 19:23
dduncan and fun is had by all
autrijus verily. 19:24
dduncan aye
geoffb :-)
svnbot6 r7541 | iblech++ | Pugs.Pretty: Don't output evil bytes in .perlifications. Instead, use 19:26
r7541 | iblech++ | escapes (\t, \n, etc.) or {...}: "\123".perl eq '"{chr 123}"'.
r7542 | iblech++ | PIL2JS: Prelude::JS::Operators: Implemented +| +& +^ +< +>. 19:32
stevan autrijus: ping 19:49
autrijus stevan: pong. I need another 3 min or so to wax poetry on the flattery of VB 19:51
after that I'll be free
(so you can type ahead) 19:52
stevan autrijus: ok,.. since it is such a good cause :)
autrijus of course!
Juerd geoffb: The lag is huge.
geoffb: But there are also many things that still need decisions 19:53
geoffb: Since some of these yet unmade decisions are very important, it'd probably be a waste of time updating the bible as much as needed, before they are made.
svnbot6 r7543 | iblech++ | * Makefile.PL: Unbreak "make haddock" (mkdir dist/doc/html can't work if !-d dist/doc). 19:55
r7543 | iblech++ | * PIL2JS: PIL2JS.pm: Use File::Temp for greater security (we don't want Pugs to
r7543 | iblech++ | be listed on securitytracker).
geoffb Juerd, important as in "Will change a whole bunch of existing things" or as in "Will add a lot"? 19:57
Because if the former, updating now would not be a waste . . .
Juerd will change few but omnipresent things :)
geoffb er,
Juerd latter :)
geoffb "if the latter"
bah
hmmm
Juerd As time and projects progress, the more subtle issues need to be solved
Often, it turns out that a change has impact in several areas 19:58
Still I think the design is very solid, as the basic ground rules have changed very little over time.
autrijus stevan: done.
Juerd Most of the big changes are syntactic.
eric256 thought TSa bowed out of the mailing list? ;)
stevan autrijus: excellent 19:59
so, eigenclasses
geoffb Fair enough. But I'm feeling partial deadlock in the P6 aether, and that's disturbing the force . . . to mix several metaphors
Juerd eric256: I thought so too, but in the same message he continued discussion
svnbot6 r7544 | iblech++ | * Pugs.AST.Internals: Pairs are stringified as "$key\t$value" now, per
r7544 | iblech++ | www.nntp.perl.org/group/perl.perl6....age/23224:
r7544 | iblech++ | Also that a pair ($x => $y) stringifies to "$x\t$y"
r7544 | iblech++ | * t/data_types/pair.t: unTODOed several tests.
stevan autrijus: in order to have inheritable class methods, we basically need to create an eigenclass for each class
Juerd eric256: So perhaps he was using a different metaphor, and we interpreted his words in a way that suited us best ;)
eric256 the discussions on p6l seem to be gonig in circles, i can't follow them to save my life. ;) 20:00
stevan because they need to be in a parallel hierarchy to the normal classes
eric256 Juerd...could definitly be, i hadn't thought of that. languages barries often do crop up in odd places ;)
autrijus yes.
stevan all attempts to create them "on demand" lead to much silliness in terms of trying to keep the inhieritence in order
Juerd geoffb: We need a clear definition of the dark side in order to avoid it, and that has been lacking :)
geoffb eric256, the circular discussions are part of the partial deadlock I referred to 20:01
stevan however, this gets a little messy when you introduce subclasses of Class
Juerd eric256: Some things go in circles, but those things are discussed mostly by the same people again and again.
stevan for reasons we discussed before
eric256 need like czar of the list...this is done, this is the answer. shut up and move on. ;)
stevan however, I think that behavior can be specified in a sane way 20:02
Juerd eric256: There are still important new arguments for several sides of the diamond, so I want to encourage continuing trying to figure things out.
stevan: You must be careful not to leak too much of the metamodel back into the thing it is a meta above (around)
eric256 yea it just gets soo confused. then people go back to earlier proposals... well if this works then great. from the outside it looks an awful lot like a townhall meeting
autrijus stevan: so you are still thinking ondemand creation? 20:03
stevan so the questions I have are 1) does this all sound good (yes, I need validation :) and 2) should all object instances get an eigenclass? or should those be on demand
Juerd stevan: Sometimes it can be good to draw a line. Do we *need* a certain feature, or is it just nice to have. If the latter, confusion is a good reason to keep it out.
eric256 lots of posturing and side arguments that have little to do with the actual topic ;) if it works then i'll just watch blessfully ;)
autrijus stevan: I think all class objects want an eigenclass no matter they have class meth or not -- larry said as much in A12
stevan Juerd: I have adopted a "everything and the kitchen sink" policy on the meta-model, however, this does not mean we expose it all 20:04
autrijus stevan: for objects I'm less sure. I _think_ we can get by of creating them on demand if we can figure out a sane semantics; but then, it should behave as-if they are created a priori, no?
Juerd eric256: I agree that more decisions need to be made. Perhaps Larry should give up some control if he doesn't have the time to do this himself.
stevan autrijus: they don't inherit,.. so on demand is simple
eric256 even if they are the wrong decisions they can always be changed in a month....but them i'm a progress over planning kind of guy most the time 20:05
stevan however, this just means I need to explicitly code it to only create for a class, and for a instance be on demand
Juerd eric256: I have to add to that, in all honesty, that I say this partially because I think some of the recent decisions that _have_ been made have disturbed me greatly.
eric256: It's just much easier to work with bricks than with playdough.
stevan autrijus: explicit eigenclass creation fits nicely into Class::CREATE 20:06
since that is where other magic happens too
Juerd (re progress over planning)
autrijus stevan: right, and I think it's not a big problem, as only the Class Class has eager eigencreation by default
stevan: so it's contained in one place 20:07
as you said
eric256 Juerd: yea. it seems like until (done) { decide; test_on_pugs; re evalutate decesion based on experience rather than theory; }
stevan excellent :)
I will add that in this week then
how goes the haskellification of the mm?
Juerd eric256: Perhaps a feature(idea) freeze would be in order 20:08
autrijus stevan: release and flu are both past, so that is the next thing on my plate tomorrow
Juerd Work out the current issues first, even if we have ideas that might solve multiple problems at once (because really, most ideas don't, even though they are thought to do) 20:09
autrijus stevan: though demagicalizing pairs is important too (but that can fit in with container haskellification)
Juerd I have some ideas that I don't release to p6l, because the list is too distracted already
But sometimes I feel forced to do so, because a certain item has come up
autrijus ...I confess I only read p6l thru weekly summary.
(or on specific threads.) 20:10
but a daily scan proved to be very costly.
eric256 Juerd: or at least a summary of all positions inovled...its doubly complex because through a thread peopls ideas evolve...makes it near impossible sometimes
Juerd I have one very controversial idea that I fear I might have to release soon :)
geoffb autrijus, I second that
stevan I third that
autrijus nothingmuch is not here, otherwise he'd forth that
Juerd eric256: Yes
eric256 lol
geoffb heh 20:11
Juerd Which brings us to: the summaries
geoffb So it does
Juerd I think the summaries too often just explain what is discussed rather than the different sides
autrijus yeah. perl.com no longer sponsors them?
Juerd They are abstracts rather than summaries, sometimes.
geoffb Juerd, too true, especially of late 20:12
Juerd To really know what the different possibilities are, you still need to read entire threads
I can understand this for some of the highly theoretical threads
eric256 i think it realy just comes down to needing someone to make decesion. i mean there will always be people that disagree with how things are being done. the current discussions wont end until someone just gives up. ;)
autrijus eric256: exactly.
normally $Larry, or damian, plays such an arbiter 20:13
Juerd Does that mean we don't have normality at this moment? :)
autrijus it is implied, yes.
eric256 keeps forgeting its snowing till he looks over his shoulder. hehe 20:14
Juerd (z z 9 pl. z a :))
autrijus stevan: so, eigeness. is there something on the user level (i.e. in t/) that you'd like to add still?
stevan: because I'd like to quickcheck, or at least hunit, it first
stevan autrijus: something to add? 20:15
autrijus stevan: as in, universe-side behaviour changes for the eigeness model
instead of metaverse-side
or is 36 complete? 20:16
stevan autrijus: oh no,.. it is horribly broken at the moment
it keeps creating eigenclasses for each method
autrijus whew. good :)
indeed
Juerd Why is evalbot reconnecting all the time?
autrijus stevan: though... that is not in the test itself 20:17
geoffb Juerd, rebuilds after each commit,
stevan I had to back up and do a Mini-MetaModel version, to get me head around all details
iblech Juerd: To make the revnum easily visible
geoffb and the revnum is part of the nixk
er nick
autrijus stevan: ok, that sounds sane.
Juerd geoffb, iblech: ah
geoffb I suppose it could just change nick rather than reconnecting
autrijus stevan: I'll start modeling pneuma again tomorrow and let you know if I run into anything bad.
Juerd Is there no way to keep the socket?
As irssi does?
(See irssi's source code for /upgrade)
stevan autrijus: 36 needs more tests 20:18
Juerd Restart with a newer pugs, but keep the socket around, then do /nick)
autrijus stevan: *nod*
Juerd That's less distracting on IRC and keeps attention from irc opers away from feather :)
stevan I will probably port the ones from the Mini_MetaModel_w_Eigenclasses
Juerd Who maintains the bot?
stevan and add some more too
iblech Juerd: The problem is that there doesn't exist any IPC currently in Pugs
Juerd: luqui *IIRC*
autrijus stevan: hm, or do you think I can start modeling MiniMetaModel_w_eigenclasses first? 20:19
Juerd iblech: Not needed... All you really need is a low level exec and a way to regain a handle based on fd to make it happen
iblech: Or... A very simple irc bouncer :)
(Why do I think of the hard solutions before I think of the very easy ones?)
stevan just remembered another issue 20:20
eric256 couldn't the IRC bot just call an external script to run snippets?
iblech Juerd: Yep, but %*SIG<USR1> = { my $rev = get_cur_rev(); $bot.nick($rev) } would probably be easiest :)
Juerd iblech: Surely it has to actually RUN on that revision?
stevan autrijus: the eigenclasses tend to make $?CLASS and $?PACKAGE a little weird,.. however, i think that can be easily overcome
Juerd iblech: Or is ?eval not in synch with it? :)
iblech eric256: It does. evalbot.p6 (IRC) calls evalhelper.p5 (resource limits), which calls pugs 20:21
stevan autrijus: I would stick with pneuma, etc... the eigenclass change is fairly minimal
autrijus stevan: okay, good
eric256 lol. so no need for the actual bot to be on the newest version at all
Juerd Ohh
Then yes, a /nick would suffice.
autrijus stevan: I think the "allomorphism" mentioned in A12 already would calse some ambiguity with $?CLASS
iblech eric256: Yep, the revnum is purely for informational purposes. Even if the bot did not restart, ?eval would still run the latest Pugs 20:22
autrijus stevan: but as long as we make sure that each eClass for iClass reflect back to the iClass and not the cClass for the iClass, it should be fine
stevan :D
iClass connected to the eClass the eClass connected to the cClass ...
eric256 is the interface portion of IO pretty nailed down and your jsut working on the implementation? 20:23
stevan autrijus: once the eigenclasses are in, I will be concentrating on the MOP document, I expect that will lead to many p6l posts
autrijus stevan: good. I think we should codename 6.28.0 "stevan" 20:24
pdcawley_ What's local's name now?
stevan is not sure that is a good idea :/
autrijus stevan: probably not, but entertaining to think about 20:25
pdcawley_: temp
pdcawley_ Good oh.
stevan codenames all his projects "autrijus"
pdcawley_ contemplates 'my method ARRAY::attributes {...}'
stevan lexical singleton methods? 20:26
pdcawley_ Well, not even singleton methods reallly.
Just making sure that role helper methods never escape the role's scope.
I suppose my multi(Scalar $thing:) {...} would do the job as well, but I kind of like the method version. 20:27
or my multi method attributes(Scalar $thing) {...} or whatever it's supposed to be. 20:28
stevan slowly backs away from pdcawley_
pdcawley_ What?
stevan just teasing
autrijus pdcawley_: "my multi" is canonical; declaring a method for a class can be done with 20:29
stevan the method approach is interesting,.. however, I think it would likely be just sugar for the multi approach
autrijus class Foo is reopened
but that can't be done lexically
so stick with multis.
pdcawley_ points out that the Smalltalk/ruby hacker in him would just bung the helper methods on the classes in question and assume that he'll deal with any problems when they arise.
autrijus pdcawley_: but lexically? 20:30
pdcawley_ Nah, wouldn't bother with the lexically.
autrijus okay, dynamically, sure!
stevan points to the LISP hacker in him which would just lexically redfined the method :)
pdcawley_ The 'hang on, don't let anything have a wider scope than is absolutely necessary' part of me would prefer the lexical version.
stevan: Exactly. 20:31
autrijus I still can't see anything wrong with "my multi".
stevan (+ + lexical redefinintion of method)
pdcawley_ doesn't like multis that don't dispatch on multiple arguments.
stevan autrijus: lexical redef is sexier
pdcawley_ And... what Stevan said.
stevan declares Perl 6 the New LISP
of course that is continginent on nothingmuch's Exceptuations being approved 20:32
which are very much like LISP conditions
iblech putter, fglock__: FYI, new PIL-Run smoke online.
autrijus my method attributes (Scalar $thing:) { ... } 20:33
so you'd like that?
pdcawley_ wonders if he should be feeling guilty about TSa's departure.
autrijus hm.
stevan pdcawley_: from what I hear, he left becuase of the summaries
iblech Need to sleep now, night all and happy -Ofunning! :)
stevan ;)
theorbtwo TSa departed?
autrijus iblech++ # woot!
theorbtwo Oh, what's the subscribe address for the summeries?
pdcawley_ autrijus: Not really, I want to lexically declare a method on another class.
autrijus pdcawley_: from what I hear, he is not really left :)
pdcawley_: "Scalar" is that class 20:34
my method blah (OtherClass $self:) { ... }
pdcawley_ Then surely my method Scalar::attributes($self:) {...} makes more sense.
stevan should it be my or local? 20:35
pdcawley_ But your syntax has a pleasant weirdness to it...
autrijus I contend I agree.
stevan or temp or whatever local is now
autrijus stevan: you can use temp too, but they mean different things.
stevan ah 20:36
pdcawley_ stevan: my I think. temp scoping would mean it could escape to method called on other objects that don't do the current role.
fglock__ iblech++ 20:37
wolverian hmm, I'm still confused towards item/list context versus type context
eric256 why don't they call it local instead of temp or my? ;)
autrijus wolverian: item/list contexts each carry a type 20:38
only void is typeless
?eval want
evalbot_7544 "Scalar (Any), LValue"
autrijus ?eval ~want
evalbot_7544 "Scalar (Any), LValue"
autrijus ?eval list want
evalbot_7544 "Scalar (Any), LValue"
autrijus hrmph.
autrijus smacks evalbot
wolverian autrijus, ohhh. that makes more sense. is want()'s return specced?
autrijus wolverian: yes, S06. 20:39
wolverian thanks :)
oh. hm. how do I inspect what _type_ is wanted? 20:40
(or is that possible?)
or is that carried by the context information returned by want(), as shown by evalbot?
hmm. pugs's want() does not carry type information. right. 20:44
but it should.
things are clearer now. sorry for being so slow. :) 20:45
autrijus er it does 20:46
wolverian oh
except it's always Any 20:47
autrijus mm, weird.
pdcawley_ giggles at a post by luqui in which he takes the monte carlo approach to spelling by using 'affect' and 'effect' as nouns meaning the same thing. Presumably the theory being that at least one of 'em must be correct.
wolverian oh, neat; S06 says gather is lazy. 20:48
I didn't realise that's official. :)
eric256 hey sometimes its hard to rememerb which is wich, and 50% wrong is better than 100% wrong ;) 20:49
pdcawley_ decides not to make a crack about it in the summary. 20:51
Juerd wolverian: gather shouldn't be lazy by default! 20:59
wolverian: only if what's inside is pure, it can be
But this has been mentioned before
cognominal trying to install hs-plugins... with 'sh configure', I get sh ./configure 21:06
configure: error: cannot find install-sh or install.sh in . ./.. ./../..
what am I doing wrong? I run mandriva cooker.
do I miss some magic invocation before 'sh configure' ? 21:07
autrijus cognominal: ask #haskell dons? 21:10
cognominal oki 21:11
autrijus it's dons not don :) 21:13
afaik he is in sydney, not sicily
eric256 will use lib be builtin? or will i t be a module? 21:17
autrijus eric256: I think a module for sure, just like in pugs now 21:21
eric256 just curious because some examples don't work because they can't find the lib module 21:22
autrijus hm, you need to run with -Iblib6/lib or install pugs 21:24
cognominal hum, I am still missing a step I get "ghc-pkg: cannot find package plugins" when running Makefile.PL 21:28
autrijus cognominal: "make register"
cognominal I tried that too 21:29
/usr/bin/perl5.8.7 -Iinc util/ghc_setup.pl register --prefix=
*** Running: ./Setup register
No description file found, please create a cabal-formatted description file with the name <pkgname>.cabal
I see a Pugs.cabal.in though 21:31
autrijus oh hm, did you run Makefile.PL? 21:32
and "make"?
I mean "register" in the hs-plugins dir
not pugs
wolverian Juerd, oh, well, S06 said that. 21:33
eric256 i had been staying away from make install....is it safe to use now? 21:35
also...is there going to be an environment variable to change the lib in perl6? seems like a pretty common need. 21:36
geoffb make install seems to be working decently as of this release, eric256 21:37
It's not perfect, but it's also not deadly, or entire-disk-spamming
eric256 okay ..../me crosses his fingers 21:38
geoffb [OT] What's the fastest Perl 5 data structure serializer that produces textual serializations? 21:42
dduncan something in XS probably
when you say 'textual', do you mean 'human readable'?
autrijus geoffb: svdump?
aka Devel::Peek 21:43
geoffb dduncan, yes, and diff-friendly
autrijus, hmmm, OK. Does it have a thaw?
Or is it freeze-only?
autrijus I don't think it has a thaw 21:44
dduncan if the output is valid perl code, it can thaw itself
autrijus Data::Dumper::Streamer?
I'm not sure you'll have to bench
mugwump Data::Dumper is pretty fast compared to, eg, YAML
geoffb OK, just figured there was a pretty good chance someone here might have already done so on one of their other projects. :-)
eric256 hmmm are there any conditions that evaluate the block before the conditional? 21:45
geoffb mugwump, I thought that was because Perl 5 still has the lamest YAML implementation
mugwump no comment
eric256 so that $new = $lottery_numbers.pick() until $new == none(@your_numbers); would DWIM
svnbot6 r7545 | stevan++ | Perl6::MetaModel - 21:48
r7545 | stevan++ | * fixed the current eigenclasses so they don't keep creating
r7545 | stevan++ | eigenclasses and only create one per object (on demand)
r7545 | stevan++ | - added some tests for this
eric256 anyone?
cognominal autrijus, why don't you give ftp://ftp.cse.unsw.edu.au/pub/users/dons/hs-plugins/hs-plugins-0.9.10.tar.gz as the place to get hs-plugins? 21:52
autrijus cognominal: I don't know, maybe it was not released then? if it works for you, please go ahead and change things 21:55
rafl gaal: ping 21:56
gaal: Fresh packages for you. 21:57
eric256 in examples/nested_loops/functional.pl the example used a ==> which doesn't work yet, but it is easily rewritten to workd 21:59
svnbot6 r7546 | eric256++ | Changed so that the same number can't be picked twice. 22:00
eric256 should i just drop a comment in with the old version? i think i'm not being true to his goal by fixing it. lol
autrijus drop a comment in with the old version. 22:01
eric256 k. just thought working examples are better than *right* examples...
rafl How about merging smoker.yml and config.yml? 22:04
autrijus worksforme.
gaal: sanity-p?
eric256 are methods multi methods by default? 22:07
autrijus no I don't think so
wolverian ..and that sucks
eric256 wolverian...you mean about multi methods? 22:08
wolverian yes 22:09
I should say that I _think_ it sucks. I'm by no means certain.
eric256 i dunno. its not too bad...though you should be able to say that all methods in the class you are creating are going to be multi
class x { use multi::method; method x () { }; method x(Int x) { } }; 22:10
probably will be easy to add as a macro....are macros lexicaly bound?...food for though
thought even
autrijus macros can be lexically bound yes. 22:11
svnbot6 r7547 | eric256++ | Fixed up some changes in syntax so that the examples all work. Tweaked functional.pm slightly for now to avoid errors because of the lack of ==>
eric256 cool so a module like the one i just said would be easily made with macros...right?
autrijus eric256++ eric256++
eric256: yes.
eric256 sweet...do macros exist in pugs yet? 22:12
autrijus yes, the source-filter form
see t/macros
eric256 cool. thanks
submethod trace( : Str $msg) { 22:13
does that mean something different than 22:14
submethod trace( Str $msg) {
autrijus should be the same thing. 22:16
(I think.)
eric256 hehe. first one errors.
?eval submethod trace(: Str $msg) {}
evalbot_7547 Error: unexpected "S" expecting "?", "*", "+", "++" or parameter name
autrijus oh. hm, leading colon in signatures.
?eval sub x (:Str $x) 22:17
evalbot_7547 Error: unexpected end of input expecting bare trait, trait or block
autrijus ?eval sub x (:Str $x) {}
evalbot_7547 undef
autrijus ?eval submethod x (:Str $x) {}
evalbot_7547 undef
autrijus ?eval submethod x (: Str $x) {}
evalbot_7547 Error: unexpected "S" expecting "?", "*", "+", "++" or parameter name
eric256 spacing is making it mad?
autrijus apparnetly 22:18
eric256 hehe. alright. i'll just drop the : for now. unless that causes something weird?
autrijus no, just drop it I think 22:19
eric256 done
svnbot6 r7548 | eric256++ | Fixed function signature. 22:23
geoffb
.oO( Spending an hour or two refactoring is very calming . . . )
 22:28
eric256 any reason $x ~~ s:perl5:g/\s*//; would hang? 22:32
autrijus er, yes. 22:33
eric256 hmmm. well its stupid...but hanging seems a bit overkill
autrijus very true
eric256 lol
geoffb PCRE-ism?
autrijus no, bad handling in Pugs.Prim.Match line 115 22:34
eric256 hehe. part of a QOTW...and thats the solution given...suprised they didn't realise what broke it..
autrijus I guess zero width match needs special care 22:35
my brain is kind of fried from flu to do this now. I'll add a note.
eric256: turn this into a test please?
let it fail and segfault.
er, hm, it hangs instead 22:36
(it segfaults here)
geoffb
.oO( My $work desire for Perl 6: An HTML::Mason derivative that has decently fast component calls. Grrr. )
eric256 never segfaults here 22:38
lemme know what you want me to do.. 22:39
autrijus I'm thinking
still turn this into a test.
I'll fix now
eric256 thanks. and sorry for finding it ;) 22:40
autrijus no, it was very important :)
I wonder why it was not uncovered earlier.
eric256 no one was stupid enough to code it that way
lol
geoffb perhaps we need more stupidity check tests.
Parsing tests for empty strings, empty blocks, extra punctuation that "shouldn't" make a difference, and so on 22:41
autrijus that reads QuickCheck, but yes.
geoffb OK, sure, but also stuff like eric256 found: stuff that could cause bad behavior, but that we should gracefully detect. 22:42
Tests that check if we have a recursion limit, for instance
autrijus *nod* 22:43
okay, fixed, I think.
eric256: test in t/rules/s_perl5.t? 22:44
eric256 how do i phrase it so it doesn't hang the test if it is still broken? 22:47
autrijus you don't, as I've fixed the bug.
eric256 ohh. okay. ;) 22:48
rafl pugs 6.2.10 is being uploaded to Debian now. Should be in by tomorrow evening. 22:50
autrijus rafl++ rafl++ rafl++
rafl It was slower than I wanted, though. Made a mistake in putting the haskell lib into /usr/lib/haskell-packages instead of $prefix/lib. :-) 22:51
autrijus if it works it works :)
have you tried compiling test.hs?
import Pugs
main = pugsMain
with
ghc --make test.hs
geoffb rafl++ # MMmmmm, packages . . . .
cognominal hi, Ymmv :) 22:52
autrijus greetings, variadic one :)
svnbot6 r7549 | eric256++ | Fixed more typos in some examples so they now work.
r7550 | eric256++ | Added test for s:perl5:g/\s*//;
meppl difficult to pronounce "ymmc" 22:53
Khisanth Your _ _ variadic?
geoffb Your Mileage May Vary
Ymmv is lunatic, so YMMV with me :)
autrijus good, I think you'll fit in here :) 22:54
geoffb He's crazy, hand him the committer bit . . .
rafl autrijus: Some unresolved symbols. Some to altdata. -package altdata fixes them. This remains: /usr/lib/haskell-packages/ghc6/lib/Pugs-6.2.10/libHSPugs-6.2.10.a(Haskell.o): In function `s2kg0_dflt': 22:55
: undefined reference to `SystemziEvalziHaskell_evalzu_closure'
/usr/lib/haskell-packages/ghc6/lib/Pugs-6.2.10/libHSPugs-6.2.10.a(Haskell.o): In function `__stginit_PugsziEmbedziHaskell_': 22:56
: undefined reference to `__stginit_SystemziEval_'
/usr/lib/haskell-packages/ghc6/lib/Pugs-6.2.10/libHSPugs-6.2.10.a(Haskell.o):(.rodata+0x3c): undefined reference to `SystemziEvalziHaskell_evalzu_closure'
PerlJam patiently waits for Pm to check in his shift-reduce parse.
autrijus -package plugins?
PerlJam er, parser
autrijus PerlJam: I tried it again in the sketch.
rafl autrijus: Fixes it. The debian package isn't compliled with hs-plugins, though. It's still not in the archive. :-( 22:57
PerlJam yeah, I saw. He's not keen on releasing not-quite-finished work is he?
autrijus k, so no worries.
svnbot6 r7551 | autrijus++ | * saner detection for zero-width matches on s:g to kill
r7551 | autrijus++ | the infinite loop as reported by eric256++.
r7552 | autrijus++ | * finalize Parrot embedding properly and let it flush
r7552 | autrijus++ | correctly, by calling Parrot_exit at exit. leo++
eric256 ?eval my $x = "hello "; $x ~~ s:perl5:g/\s*//;
evalbot_7551 6
eric256 ?eval my $x = "hello "; $x ~~ s:perl5:g/\s+//;
evalbot_7551 1 22:58
autrijus PerlJam: yeah... but I learned that he probably really has code behind it
eric256 lol. hmmm
autrijus whenever I was not releasing something due to not-quite-finished, it means I don't have anything at all and all the throwaway prototypes has been thrown away :)
PerlJam heh
autrijus (sad but true) 22:59
eric256 there we go....tested 95% of examples and they all work now. ;)
PerlJam Pm is probably similar. In the time i've known him, there are things that take a long time to complete because he's thinking about them. When he's finally thought enough to actually do something, the bulk of the code materializes within a day or two.
eric256 a couple long running ones or ones that needs perl6 rules didn't get tested in case someone else wants to look at those. ;)
autrijus PerlJam: nod. but I try to chalk out Visio pictures and random irc/journal blabbering and proof-of-concept .hs code meanwhile :) 23:00
thinking aloud vs thinking quietly, I think.
elmex hmm.. i just read a ruby article... where a proposal was done: s/eigenclass/idiotclass/g
PerlJam yep.
elmex - meaning that naming it in anyway is ... weird 23:01
eric256 later
autrijus er, no, 23:02
idioclass
not idiotclass
"idio", from german root, "self-standing"
elmex oh
hm
svnbot6 r7553 | autrijus++ | * also perl5_free the p5 interpreter at global finalizer
r7553 | autrijus++ | time. Tested and works with both p5+parrot embedding. woot1
elmex well, isn't it just an implementation detail ?
ingy_ autrijus: ... 23:03
cpan-upload -mailto [email@hidden.address] -user INGY -password ******** Perl6-Bible-0.20.tar.gz
Finished!
autrijus but it has... weird connotations
ingy_: thanks!
ingy_ autrijus: building the bible is all automated for me
PerlJam autrijus: and an obvious misspelling that isn't too flattering.
autrijus elmex: well, the ability to tack on arbitary methods on random objects without changing it class
PerlJam: indeed
elmex: is more than an implementation detail.
you can't quite do that in perl5.
rafl Is there any other Distribution than Debian that comes with pugs, btw?
elmex autrijus: well, self/smalltalk/other languages have that ability without any special name 23:04
autrijus elmex: right, I agree
ingy_ Unfortunately I got dragged into meetings, but just ping me here when an update is needed
autrijus but I think the eigen name is fun.
ingy_ it only takes 5 minutes
autrijus ingy_: cool!
ingy++
ingy_ (+ 5 more for paranoia)
autrijus ingy_: it's all automated, so what about a web form? :D 23:05
elmex k :)
ingy_ sorry I can't seem to contribute more...
autrijus ingy_: so I can channel it to svn.perl.org watcher
ingy_: so whenever there is a checkin there is a new bible
(checkin to perl6/AES/ that is)
ingy_ autrijus: how about whenever there is a checkin, it pings me here? 23:06
autrijus ingy_: ok, deal
ingy_ It's *mostly* automated ;)
I like to sanity check it
autrijus I figured you were exaggerating a bit :)
sanity checks are good.
23:06 ingy_ is now known as ingy
ingy My automatic module review and upload facility has human verification checks... 23:08
offby1 That sounds like Zippy:
autrijus right, because ingy is also fully automated
offby1 Yow! My AUTOMATIC MODULE REVIEW and UPLOAD FACILITY has HUMAN VERIFICATION CHECKS!
autrijus yow.
mugwump and ITS UNSTOPPABLE :) 23:09
ingy I'm RELIGIOUS!! I love a man with a HAIRPIECE!! Equip me with
MISSILES!!
ingy channels purl 23:10
autrijus for further enlightenment, consult ./pugs examples/network/http-client.p6 23:11
;)
rafl Hi nomeata 23:14
nomeata++; # for uploading the pugs package to Debian for me
cognominal where is the Perl6::Bible control system where I can read the diffs?
s/control/version control/ 23:15
autrijus cognominal: svn.perl.org/perl6/doc/trunk/ 23:18
cognominal thx 23:19
autrijus np
leo__ started just 'make' after a 'svn up' of a few revs - the previous one was 'make unoptimized' - does that work? 23:33
never mind - I ^C-ed it anyway - taking too long ;-) 23:37
ingy autrijus: do you msg anymore?
autrijus ingy: you need to nickserv your nick register 23:38
to msg anybody
23:38 sleepster is now known as typester
autrijus it's silly, so try msg me in magnet or ircnet 23:38
leo__: it works, but "unoptimised" is faster 23:39
leo__ yep - thx
meppl gute nacht 23:58
geoffb night, meppl
meppl ;) 23:59

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!