pugscode.org <<Overview Journal Logs>> | r1773/winxp: 587/4256 smoke/win2k (occasional): xrl.us/fqum | pugs.kwiki.org
Set by theorbtwo on 11 April 2005.
mugwump I've written up a little "add an operator HOWTO" at pugs.kwiki.org/?PugsOperators 00:00
stevan mugwump: very nice 00:08
markstos mugwump: the screenshots were a nice touch. 00:14
stevan autrijus: found a bug for you 00:41
pugs -e 'my %h = ("a" => 1, "b" => 2); say %h{"a"} <=> %h{"b"} ?? "true" :: "false"'
errs with : No compatible subroutine found: &h
and that is only the begining of the fun
<, <= and <=> do not work/parse properly with lists and hashes 00:42
see t/pugsbugs/spaceship_and_datatypes.t for more details
I found this while adding more tests to t/builtins/list/sort.t 00:44
jabbot pugs - 1926 - adding tests to sort.t and a new pugsbug 00:50
mugwump orders a copy of TaPL & ATiTaPL 01:29
nothingmuch bwahahahahaa! 01:33
HTMLMatrix now does everything that testgraph.pl does, except for the stuff that it doesn't 01:34
nothingmuch.woobling.org/mpeg_matrix.html 02:13
nothingmuch.woobling.org/mock.html 02:15
mugwump boggles to see he is 300 commits behind on the pugs tree 02:45
crysflame i like watching it 02:47
machack666 what is the expected behavior with hyper operators when you are dealing with lists of different sizes? does it extend the shorter list, similar to the scalar promotion, or does it truncate the output list? 03:04
example: "say ((1..6) >>+<< (1,0))" yields the list (1,3). 03:06
but the argument could be made for the (1,0) list to be expanded in several ways: (1,0,1,0,1,0) -- ie, repetition. 03:07
or for a given operator, there is (may be) an identity element: 0, in this case. So it would expand to (1,0,0,0,0,0).
Thoughts, anyone? 03:08
mugwump scanned p6l? you may have hit a corner case no-one's talked about yet 03:11
However, I would expect that code to generate lots of undef warnings 03:13
machack666 I may forward the question there...
mugwump and return (2,2,3,4,5,6)
(as undef numifies to 0)
machack666 how about the case (1..6) >>*<< (2) ? (2,0,0,0,0,0) ? 03:14
for multiplication, you could want the unidentified elements untouched 03:15
mugwump sure, because AIUI you can use a variant on the hyperoperator to say only one side is hyper,
ie
(1..6) >>* 2 # or something like that
machack666 I guess I'm trying to resolve the auto-promotion of a scalar to a list in the hyper-op case 03:16
presumably if you've done something like (1..6) >>*<< 3 it means one thing, but if you do (1..6) >>*<< (3) -- or some calculated value in parens, say -- you get a different value than expected. 03:17
you would either have to force scalar context on the rhs: (1..6) >>*<< +($a-$b) 03:18
or automatically promote the shorter list to an appropriate length
anyway, p6l is probably the appropriate place to discuss. Sorry for the off-topic 03:19
mugwump true, I suppose it's a list operator, not an array operator
nothingmuch ping? 05:38
kungfuftr: ping 05:45
gaal morning 05:46
nothingmuch hola gaal
did you see the table output?
i wonder why we aren't just uising "<a title="tooltip"/> 05:47
newlines?
gaal ah! a title="wonderful"!
dunno
though - as i said before, it woudl be great to ahve this cutandpastable 05:48
(i know, can't do that w/o js)
nothingmuch =(
gaal hey! you solved the msieproblem!
nothingmuch++
actually, it's probably not at all evil to use the title attrib, *and* add js that turns it into a selectable window. automatic graceful fallback! 05:49
so just need to test this for newlines. 05:50
speaking of which, your html doesn'
t have them :)
nothingmuch i'll put it up soon
i know
160kb less for mpeg audio frame tests
gaal w00t
nothingmuch and mod_gzip gets 2x more 05:51
gaal w0000t
(though i find that surprising)
nothingmuch anyway, bbiab
gaal (it shgoud have compressed the old one much better)
nothingmuch i'll clean the code
gaal i have to go to $work now
nothingmuch it compress 460kb to 12kb
gaal see you later!
nothingmuch and 300kb to 8kb 05:52
ciao
gaal hasta luego
06:37 castaway_ is now known as castaway
ingy hola 07:03
uploading new Perl6::Bible
(say your prayers) 07:04
castaway ,)
ingy <---- sleepy time
nighty
castaway night ingy
nothingmuch morning 07:05
castaway mornin nm 07:08
Entwicklung perl6 Bible ? 07:11
castaway on cpan, a module containing the Synopses and Apoca-whatsits, basically
Entwicklung oh ok, i thought it was a new perl6 book 07:13
autrijus greetings lambdacamels. this is but a short message saying I'll be off to $work for today and tomorrow, and try to find some time get PA02 outline done in between. I'll keep an eye on the irc log though. :) 07:20
&
theorbtwo Where's the darcs repo for testgraph & friends. 08:01
Juerd gcom: pong 08:03
castaway 15:50 < nothingmuch> darcs get nothingmuch.woobling.org/Test-TAP-Model
theorbtwo Thanks.
castaway (iek. /lastlog nothingmuch was a bit of an overkill search)
nothingmuch theorbtwo: testgraph is WIP
i have to clean stuff up 08:04
and i'm messing with CSS which is very slow progress
castaway WIP?
nothingmuch work in progress
castaway oh, work in progress
theorbtwo nods. 08:06
Looks to be a complete rewrite. 08:07
nothingmuch i hate hate hate css 08:08
theorbtwo: not exactly
it's the good pieces taken out 08:09
theorbtwo nm: Everybody hates CSS.
nothingmuch but lots of refactoring
theorbtwo It's just better then the alternative.
nothingmuch it's so overly complicated
except that if it went that far it should have been even more complicated
theorbtwo Oh, I don't think it's overly complicated as much as not quite complicated enough in two places.
1: Positioning should take an xpath of what to be rel. to. 08:10
nothingmuch it's a very broad spec, with too little depth
theorbtwo 2: You should be able to say "class foo should format like class bar, except...".
nothingmuch my grudge has always been with block layout
theorbtwo Those two things would make it much more usable.
nothingmuch it's 100% voodoo
theorbtwo Yeah, that's what 1 is meant to fix.
You end up with all sorts of tricks to say "put it /here/", where /here/ is where some other element is. 08:11
nothingmuch non breaking behavior
in the little details, i mean
margin, padding, border widths
these are the things that have always frustrated me
to tears
theorbtwo Hm, I don't have many problems with that, when I'm not being overly perfectionistic. 08:12
nothingmuch say
can you have a row whose columns are spaced as though independant from the other columns?
theorbtwo Use two sepperate tables.
nothingmuch that is if i have the columns in one row, and two in another, the columns of the second row have a colspan of 1.5 08:13
i'm trying to work around that
=)
theorbtwo Oh.
Not quite independent, then.
nothingmuch it's giving me trouble that border-spacing can only sort of solve
and i end up with yucky yucky fixes
theorbtwo Double the colspans everywhere, so that 1.5 becomes a 3, and a 1 becomes a 2.
(So they're all integral.) 08:14
castaway fetches tea and looks pointedly at theorbtwo.
theorbtwo Since the code is being generated by script, it's much less important that it be easy-to-write. 08:15
nothingmuch i'm trying to reduce size
testgraph.pl output is around 1mb last i checked
i can pull it down to about 400k 08:16
theorbtwo That'd be impressive.
nothingmuch (since it's xhtml that works)
and then another 25% by removing useless spaces
(foo | xmllint --noblanks - > foo.xml)
this makes it compress down to around 20k
but it still takes a long while to render
anyway, i think i'll give up trying to tweak this little diddly 08:17
it's just really annoying that the little test case blocks have some spacing around them
theorbtwo If there's no spacing around them, how to you tell where one block stops and the next begins?
nothingmuch there's a border 08:18
table.m td { padding: 0; margin: 2px; }
table.m { margin: 0; margin-bottom: 2px; border-spacing: 2px 0; empty-cells: hide }
that's the best i can do
the little cells are spaced between them
but they act as one unit
almost 08:19
except that i have to pull weird tricks to control the margins of: the bottom most table (margin:0)
the leftmost colum of ever table (margin-left: 0) and the ditto for the rightmosst
theorbtwo OK, I see what you're doing.
nothingmuch the last two are easy with CSS
theorbtwo More or less.
nothingmuch the table margin thing must be controlled within the templating and i hate doing stuff like that 08:20
it's just aweful to maintain
theorbtwo shrugs.
nothingmuch i think i'll give up for now, leave that commented in the css
and someone with more mojo might look at it later 08:21
theorbtwo nods.
Re making it smaller, I take it that you're using custom &foo;s to compress out the common bits? 08:22
nothingmuch custom what? 08:23
castaway entities
nothingmuch ah, no
classes are one char where possible
there's less elements
overall the structure is more concise
that is all
theorbtwo Ah.
castaway html obfu?
theorbtwo Nifty.
nothingmuch castaway: actually more readable, IMHO ;-)
theorbtwo I wasn't optimizing for size or readability; I was optimizing for getting it out the door. 08:24
nothingmuch nothingmuch.woobling.org/testgraph.petal
castaway generally one does
nothingmuch view source 08:25
otherwise it will probably try to be very witty and display it
which sort of works anywho
nothingmuch.woobling.org/mock.html
theorbtwo Oh, no wonder yours is so much simpiler; I was avoiding a title on purpose. 08:27
...and you don't have links.
nothingmuch i do have links 08:28
not for mock though
since it's a p5 module
if there is the possibility for a link, one is generated
and why were you avoiding a title?
i started with your <div>s, but they are not valid
theorbtwo Most browsers don't handle them with multiple lines.
nothingmuch hmm
i think maybe we should try to get some hybrid solution for them 08:29
theorbtwo It wasn't until I tried viewing it on IE that I realized that multiple lines would be good in this context.
nothingmuch title normally
but javascript + div if supported?
theorbtwo Ew, js.
theorbtwo shrugs. 08:30
nothingmuch oh my 08:31
theorbtwo Really, I'm more or less of a "whatever works" theory.
But CSS seems to work with less work then JS.
...and work better then a title.
nothingmuch but it's flakey
browsers should die
that's one place where I wouldn't mind monopoly
css does
on safari it looks pretty aweful
not that it's easy to get it right 08:32
castaway looked better on Opera than firefox ,)
theorbtwo Hm, I haven't tested on safari at all, but it looked perfect on konq last I tried.
nothingmuch well, anway
theorbtwo Eh, not after a tiny workaround.
castaway :)
castaway attempts to get some stuff done, in the hope that she can fit in some test writing after. 08:33
theorbtwo actually gets some tea.
shapr gimme some tea! T!
shapr boings
shapr waves pom-poms
castaway passes shapr some tea. 08:34
mornin bouncy
shapr slurps noisily
Good morning! How ah yah?
theorbtwo Hey!
castaway not to bad.. pondering writing some tests
theorbtwo I thought you said you left me tea in the thermos! 08:35
castaway I did ?
I did !
theorbtwo About 3mL.
shapr I didn't drink it!
castaway oops
sowwy
make new? 08:36
nothingmuch ok, now to refactor Model so that HTMLMatrix is cleaner 08:37
kungfuftr *sob* 09:06
no more automated builds of pugs for me, my build machine just can't cope any more
nothingmuch: looking good 09:07
nothingmuch kungfuftr: you can cut down prolly 09:08
consolidated smoking should be here real soon now
sleep 1200 between each smoke, that should be good enough
kungfuftr nothingmuch: it takes 50 minutes to build and test puhs on my box
=0)
but it's now hitting swap
=0) 09:09
nothingmuch ah
that's aweful
castaway eep :) 09:10
kungfuftr needs to replace his server, it's a bit ancient 09:12
gollem: OI
nothingmuch: have you got Test::TAP on freepan? 09:13
nothingmuch kungfuftr: no, no releng work yet
patches welcome for module::build madness
copyrights 09:14
etc
nothingmuch is documenting and testing
kungfuftr nothingmuch: iirc you can just use freepan as an archive and then tag for release (for releng) 09:17
nothingmuch kungfuftr: the code is up
publicly
if that's what you want 09:18
kungfuftr =0)
nothingmuch darcs get nothingmuch.woobling.org/Test-TAP-Model
and soon nothingmuch.woobling.org/Test-TAP-V...HTMLMatrix
kungfuftr =0)
nothingmuch (doc blitz now)
kungfuftr tests?!
nothingmuch tests?
! 09:19
kungfuftr: ?
kungfuftr hides 09:20
nothingmuch what did you mean?
kungfuftr & work # damn javascript 09:21
nothingmuch dwheeler wrote Test::More for javascript 09:22
=)
kungfuftr fckeditor-- # confusing the shit outta me
scw Hi, there is a parse error but I cannot find out how it happens 09:23
castaway is one where?
scw %h<a> < %h<b> being parsed as %h{'a'}{'%h<b'} 09:24
castaway nice..
scw seems like the qw literal eats trailing spaces but my last rivision didn't fix it 09:25
castaway add a test for it? (I think someone had that recently)
sure you're using the latest?
scw svn HEAD
nothingmuch wjat
what's a good way to distributre an html and a css with a p5 module 09:26
kungfuftr nothingmuch: "htdocs" folder? 09:29
nothingmuch and where will that be installed?
kungfuftr not sure
nothingmuch know of a module that does that? 09:30
kungfuftr look at maypole to see how it handles it
castaway Makefile.PL has ways of installing things other than module stuff 09:31
nothingmuch i don't want to step on anyone's feet
castaway: the how is more simpler than the decisions themselves
theorbtwo I'd say just install them next to the .pm file.
castaway lots of things have extra templates etc.. 09:32
nothingmuch is there a cheap way to find that?
i can think of catfile(dirname($INC{__PACKGE__}), "template.html")
castaway most seem to stick them in the lib dir
eg /usr/lib/perl5/site_perl/5.8.4/i686-linux-thread-multi/Prima/examples 09:33
theorbtwo That'd do it.
castaway shouldnt need to
nothingmuch search.cpan.org/~stro/FindBin-Real-1.04/Real.pm
i'd like that per module =/
castaway looks for the other way
theorbtwo $INC{__PACKAGE__} should always be correct... I think. 09:34
nothingmuch ingy: ping # you probably know a lot about this with all your crazy modules 09:35
Juerd scw: Shouldn't <> inside <> nest properly anyway? 09:36
scw: i.e. < %h<b> needing another > to even be valid syntax
castaway his use Module::Install
castaway cant find any sane ones 09:39
Juerd theorbtwo: That's $INC{+__PACKAGE__}. The + is needed to avoid unquoted string interpretation. 09:40
And isn't it filename instead of package? 09:41
Foo/Bar.pm instead of Foo::Bar
theorbtwo Hm, that it is. 09:44
castaway hmm, less */Makefile.PL is entertaining :) 09:45
theorbtwo perl -e 'sub findininc {my $file=shift; foreach (@INC) {return "$_/$file" if -e "$_/$file"}}; print findininc("strict.pm")' 09:46
gaal_ what's the simplest way to print a typeset verstion of the p6 bible? 10:10
e.g. how do you print plenty of pod (and make it look good)? 10:11
Juerd There's some pod-to-pdf project 10:13
They also publish ready to use pdf's of perldocs 10:14
But I have no idea where to find it, so you'll have to google for it :)
kungfuftr pod2latex and pod2pdf are available 10:18
gaal_ thanks
kungfuftr gaal_: search.cpan.org/~stas/Pod-HtmlPsPdf-0.04/ 10:19
gaal_ stas++
nothingmuch wait(lunch->pid); $? == E_DISAPPOINTING 10:33
darobin hi 10:46
is there a Pugs FAQ before I bore everyone to death with compilation problems?
nothingmuch darobin: i doubt it will cover it
jjjjjl hi can someone help me, I have a popup_menu that users select a particular option, once the selects the option and submits the form, How do call a sub routine eqauling the selction made ?
Juerd jjjjjl: And all this in Perl 6? :) 10:47
nothingmuch jjjjjl: you are probably using perl5 (#perl) and even then we need to know what you wrote it with
darobin nothingmuch: ok, it's just failing while compiling Lexer with RTLD_NEXT undefined (on OSX)
castaway dispatch table, jjjjjl
Juerd castaway: If they even knew what that is, would they have asked that question? :) 10:48
darobin I can't possibly be the only one trying to build Pugs on OSX :)
nothingmuch darcs get nothingmuch.woobling.org/Test-TAP-V...HTMLMatrix
castaway just grins
nothingmuch darobin: oone second
what's your GHC version?
castaway Juerd: its about as understandable as the question :)
darobin 6.4
nothingmuch tries to figure out if RTLD_NEXT is from some external lib 10:49
fink ghc?
Juerd castaway: Really? I first understood the question after reading the answer :)
darobin nothingmuch: no, grabbed it off the ghc page
nothingmuch that's what i'm using too
and it worked for me
so far
darobin weird
nothingmuch one minute
castaway Well I could be completely wrong. 10:50
rgs hi darobin.
darobin hi! 10:51
nothingmuch nm on pugs and the libraries it uses (otool -L) do not list 'RTLD_NEXT'
darobin: what's your gcc_select? 10:52
darobin it's an ifdef, maybe if I set it to 0 it'll work
2.6
nothingmuch ghc uses gcc as a backend iirc, also on osx
try 3.3
darobin okay, thanks 10:53
nothingmuch hmm
has to do with dynamic loading
are you trying to do anything with hs-plugins?
i haven't tried that, and it could be the reason 10:54
can you 'ld -ldl'?
darobin aha, I think I've found something
jjjjjl i did it in perl 5
nothingmuch jjjjjl: so #perl is the channel to ask on
castaway jjjjjl: This is not a perl5 channel
nothingmuch jjjjjl: but you need to give them more info anyway
darobin tries to see if his idea fixed it 10:55
nothingmuch darobin: what was your idea?
darobin I found an old .h that didn't have RTLD_NEXT and I'm thinking that it was picking up this one instead of the other which has it 10:56
nothingmuch ah
darobin it's the kind of obvious thing you can only think of after having asked on IRC...
nothingmuch darobin++ 10:57
castaway meetong o'clock, later folks 10:58
nothingmuch ciao castaway
theorbtwo gives castaway a quick kiss and sends her on her way. 10:59
nothingmuch theorbtwo: please download HTMLMatrix and run example.pl 11:00
oh wait, i have to update Test-TAP-Model
darobin hurrah, it builds, testing... 11:02
nothingmuch: thanks!
nothingmuch darobin: nothingmuch.woobling.org/pugs_test_status <-- results head revision, to save you some time, if you want
darobin nothingmuch: thanks, I'll run them to see if it matches 11:03
nothingmuch i'm working on refactoring and releasing the tools which are used to create that stuff 11:04
those were the darcs repos i posted
darobin now needs to find an excuse to play with Pugs 11:08
nothingmuch if you pull from both repos it should work now 11:17
i forgot to say: darcs push was done a while ago
rgs apparently #perl6 is the 63rd biggest channel on freenode 11:25
castaway returns (quick meetings)++ 11:29
Hmm.. C-x o does "other-window", is there a way of getting to a specific window, when you have several? 12:11
autark-jp C-x b 12:12
castaway thats buffer, not window 12:13
autark-jp ah
castaway I want to go to it, not bring it to me .)
nothingmuch theorbtwo: i have to do some other stuff now 12:15
do you want to patch *TAP* for rel?
copyright etc?
if you do it i'll release it now
castaway suddenly notices she asked in the wrong channel ,)
nothingmuch if you feel like bothing with Module::Build et al, you can just go ahead and release, period 12:16
stevan nothingmuch: ping 12:40
nothingmuch pong
darcs get nothingmuch.woobling.org/Test-Tap-{...TMLMatrix} 12:41
stevan nothingmuch: yes I saw
nothingmuch comments? 12:42
stevan nothingmuch: very impressed :)
I was thinking, with the "where to put the HTML" question 12:43
you might want to just put it as a __DATA__ section in a package
nothingmuch that complicates things with Petal
stevan and let perl figure out where it is
nothingmuch and i also want it to be accessible
stevan oh
oh well.. it was a thought :)
nothingmuch it's a hackable template
so i'd like to encourage hacking it 12:44
i might resort to __DATA__
stevan well if it is hackable,.. do you really want it in the perl libs dir?
nothingmuch but for now i'd like to try and find a more open solution
stevan perl -MTest::TAP::Visualize::Template -e get_template > ~/Desktop/my_template.petal 12:45
something like that
so people can make their own copy easily
nothingmuch hmm
good point
stevan ideally I would not want to mess with the perl INC dirs to change the tempalte 12:46
nothingmuch i like that
stevan and be able to pull in my own from anywhere
castaway hmm,m have it like CPAN et al, it looks for a local $HOME/.XX version, else defaults to the one in lib ?
nothingmuch you don't need to
stevan :)
nothingmuch you can do that
stevan cool
nothingmuch just give it a Petal objec
t
and it will send it (page => $self) 12:47
stevan castaway: good idea too
~/.test-tap-home
nothingmuch eek, i'm not writing a config system
stevan ok,.. version 0.02 we will do that :)
nothingmuch patches welcome either way
castaway may look... (somewhen :) 12:48
rob_ hmm, just wondering what you guys would think of a distributed module service for perl6, much like zeroinstall, but taken a step further 12:49
stevan scw: did r1932 fix the space eating problem? I was confused by your comments while reading the backlog
castaway distributed? didnt someone want to try that? 12:50
lots of "dont want to run modules from other peoples servers" complaints, iirc
stevan castaway: I agree,.. that makes me nervous
rob_ castaway: well but they dont have a problem with running from cpan
stevan rob_: but we download it local first
castaway rob_ you dont run from cpan, you download, install, run from locally
stevan rob_: it is less about malicious module authors 12:51
rob_ so why not simply using cpan as a dispatcher, and only dispatch to "signed", confirmed, whatever modules
stevan and more about man-in-middle stuff
castaway that way you know what you have.. running from somewhere else at runtime, someone could change it, and you wouldnt know
go read the perlmonks thread on it.. probably by gmpassos
stevan rob_: what happens if your server crashes during my mission critical program? 12:52
rob_ stevan: you know zeroinstall? its much like that what i would like to see happen with perl modules
castaway doesnt know it
stevan rob_: not familiar,... got links?
rob_: this? zero-install.sourceforge.net/ 12:53
rob_ stevan: zero-install.sourceforge.net/
right :)
i think thats a _very_ neat project and would really like to see it beeing used in some real world scenario 12:54
stevan rob_: it is very cool
zero-install.sourceforge.net/security.html
it says: No downloaded software is ever run as root, or as any user except the one who asked to run it. 12:55
getting perl libs to cooperate like that might be hard
rob_ hmm.
stevan you would almost need to run them inside some kind of "jail"
nothingmuch i disagree
i think there should be support for it
but that it should not be "the" way 12:56
stevan rob_: I like the idea of distributed automatic installations
nothingmuch i think i have about 2000 lines of patches code in my @INC
s/ed\b//;
castaway hmm, its interesting, yes..
nothingmuch sometimes i need fine grained control of where
or what
i think a facility to install it
castaway As an optional WTDI, it could work 12:57
stevan yeah,.. what nothingmuch said (which is much better than what I was typing)
nothingmuch like 'export PERL5OPT="-MCPAN::ZeroInstall"'
stevan rob_: give ingy and mugwump a ping and talk to them about FreePAN
nothingmuch but i don't want mission critical software falling to bits because of a slightly incompatible fix to a piece of software
castaway but as I said, may also be useful reading the perlmonks thread
right, so you'd need to be able to specify exactly which versions of what to use/get 12:58
rob_ well, that should be done in the programm like 'use Foo:bar >=1.04' 12:59
or similar
castaway Also we do a bunch of stuff that gets installed on machines without internet connections, so still being able to bundle modules is a must
nothingmuch also, lots of times CPAN.pm and especially CPANPLUS just don't work:
http proxy mess 13:00
rob_ castaway: sure that should not be the _standard_ way to go, but as an option it would be really nice
nothingmuch firewall rules
mirror down
permissions bad
castaway OTOH an easy way to install such a zeroinstall *server* here, would be nice to send out updates with :)
yup, rob_, if you can get people over the security aspects
nothingmuch so when it happens it needs to be very robust
transactionally designed 13:01
castaway nm, it only has to fetch once, after that it stays 'installed'
stevan nothingmuch: transactional ++
nothingmuch allowing deterministic manual intervention
on the first go it might break
and if it breaks, you most definately don't want it to be stuck broken
castaway sure
nothingmuch so detection of breakage should be very pedantic
and should be revertable
castaway yesindeedy
nothingmuch and so on and so forth
stevan support for the "if is aint broke, dont fix it crowd"
nothingmuch it could be useful, but only if it's done absolutely right 13:02
stevan I know myself there are some servers where I dont want to update
castaway thinks rob_ should go off and write a spec :)
nothingmuch btw, rob_ eq dargonchild? 13:04
stevan nothingmuch: I dont think so (he is at work, rob_ seems to be on a dialin)
rob_ damn, my ssh tunnel just broke. 13:26
nothingmuch hates when that happens
rob_ nothingmuch: no, rob_ != dargonchild
mj any volunteers? WinXP File::Spec abs2rel: 13:36
pugs -e "use File::Spec; say 'begin'; say abs2rel( '\a' , '\b' ); say 'end';" # begin <Inf loop>
perl -e "use File::Spec::Functions qw(abs2rel); print abs2rel( '\a' , '\b' );" # ..\a
stevan mj: I havent had time to look at it, but feel free if you want to try 13:54
and let me know if you need commit access, and I can set you up
mj i have one as rootmj 13:58
some porting and then i can try 13:59
ninereasons eep. hangman is broken :( that will never do.
mj perl -e "use File::Spec::Functions; print catfile( '' ,'a' );" # \a 14:12
pugs -e "use File::Spec; print catfile( '' , 'a' );" # \\\a
kungfuftr ie-- # caching git
apache++ # get ie to behave properly
ninereasons is the start of the array index invariable in perl6 ? S28draft sounds to me like $[ is going away and probably not being replaced ? 14:24
masak ninereasons: is there a need to keep it variable?
PerlJam ninereasons: yes. arrays always start at 0
castaway hooray :) 14:25
ninereasons I don't think so; but it's been drummed into me that the end of the array in perl is $#a not necessarily @a-1
^masak
hooray, PerlJam
castaway well unlearn it :) (plus those terms change anyway)
wolverian @a.last (.end?) 14:26
ninereasons gladly, castaway :)
@a.end
wolverian are there .head and .tail, by the way?
(the functional me speaking..) 14:27
castaway .end
would be nice :)
PerlJam wolverian: I don't think so.
Could be though.
wolverian PerlJam: is it a made decision, or just something that hasn't come up?
PerlJam Wouldn't take much to implement
ninereasons .end is specced i believe, castaway
PerlJam wolverian: I don't think I've seen anyone mention it, but that doesn't mean it hasn't come up
ninereasons: yeah, last I checked we has .begin and .end 14:28
s/has/had/
But my head has it those were subject to change
wolverian method List::head { .[0] } method List::tail { .[1...] } # works? :)
PerlJam (well, more subject to change than anything else I mean ;-)
mj # load text from fiel ... better solutions?
for (=$tut_fh) -> $line {
$file_t ~= $line;
}
wolverian mj: slurp() 14:29
if it's implemented, anyway. :)
mj great, slurp rules 14:31
wolverian slurps mj
xern anyone used Regexp::Parser before? 14:36
gugod orz ōæ½xAAōæ½xBAōæ½xBFōæ½xF9ōæ½xBB~ōæ½xA5ĪŖk 14:37
outch
sorry, wrong channel
xern i'm planning to write a conversion tool for p5regexp to p6regexp
Orz....
which will be embedded in p5_to_p6.pl
PerlJam xern: good luck!
castaway (that looked like p6 operators tho)
hmm, is that part of the YAPE explain thingy?
xern PerlJam: thanks 14:38
castaway: what looks like p6 op? 14:39
castaway the stuff that gugod just pasted :)
xern i see
stevan ninereasons: I suspect hangman is broken because the AUTHORs file changes 14:40
ninereasons do you think so? I still haven't had a chance to look at it, stevan; 14:41
xern castaway: it's a ascii glyph mimicking someone kneeling on the floor, commonly used in zh-tw community now
stevan looking now
ninereasons it appeared to me though as though maybe pick_committer wasn't working
stevan ninereasons: nope I was wrong 14:42
ninereasons rats. i liked that possibility better. 14:43
stevan any(@committers).pick is broken
it returns all of them :)
castaway oops :) 14:44
ninereasons that's what it looked like to me
from the output
How about if I change it to rand, temporarily?
stevan ninereasons: let me see if I can isolate the pick issue
ninereasons better yet.
stevan hmm t/junction/misc_junctions.t works 14:45
must be .pick()
ninereasons so it's not any, it's pick
jinx
stevan examples/junctions/lottery_numbers.p6 works 14:46
so it looks like any().pick is not working
but my $j = any(); $j.pick does
stevan writes up a test case
ninereasons a naive substitution of $j.pick doesn't work. 14:50
stevan pugs -e 'say any(1 .. 100).pick()'
that works
hmmm
ninereasons strange.
stevan (the plot thickens
ninereasons likes perl6.vim - very pretty. 14:52
mj and slurp equivalent for save ? 14:53
scw Juerd: I don't know, in fact. I just modifyed existing code # 17:36
mj open, pritn FH $text; close;
scw stevan: Err, only in string interpolate # 20:49 14:54
Ouch, I forgot I'm in +0800..
stevan very strange 14:59
pugs -e 'my @a = ("one", "two", "three"); sub foo (@a) { my $t = any(@a).pick(); say ref($t); say $t }; foo(@a)'
returns:
Array
one two three
s/returns/prints/
but yet : pugs -e 'my @a = ("one", "two", "three"); my $t = any(@a).pick(); say ref($t); say $t' 15:00
does the right thting
castaway boggles
stevan just changing line 111 to this: 15:01
my $current_committer = any(@committers).pick(); #pick_committer(@committers);
makes hangman work
ninereasons shouldn't it work as is? 15:02
what's the difference between returning the product of any, and returning any()'s product? 15:03
stevan ninereasons: it seems to have something to do with being executed inside a sub
(really odd)
ninereasons yes, stevan 15:04
stevan ninereasons: even stranger 15:07
pugs -e 'sub foo2 { any(1 .. 10).pick() }; say foo2()'
works
ninereasons integer list? 15:08
stevan no,.. one int 15:09
mj how to grab STDOUT and STERR with pugs?
ninereasons I mean, the argument to pick
er.. any
stevan ninereasons: ? i dint understand?
ninereasons nvmind - I'm grasping at straws
stevan ninereasons: no I understand now 15:10
no the int fails as well
ninereasons yeah, i saw here too
stevan it seems to be something related to array args and such 15:11
ninereasons my @a = (1..3); sub foo (@a) { my $t = any(1..3).pick(); say ref($t); say $t }; foo(@a) 15:12
stevan ninereasons: see t/pugsbugs/junction_any_pick.t
ninereasons Int : 3
stevan weird 15:13
add to the test case
please :)(
ninereasons need to clean it up; i put nonsense in it
stevan ninereasons: svn revert is your friend :) 15:14
ninereasons it's the same as your last test case, when you subtract the nonsense, stevan. 15:16
stevan ninereasons: add the passing one though (that worked) 15:17
mj pugs -e "my $out; open $*OUT,'>', \$out;"
pugs: cannot cast from VHandle {handle: <stdout>} to [Char]
stevan mj: not sure that works yet
mj: what exactly are you doing? sounds like something interesting.. 15:18
mj so tutorial porting stopped .... wiki.kn.vutbr.cz/mj/attach/pugs/tut...ad.p6.html
stevan mj: very cool 15:19
ninereasons got it:
sub foo (@a) { my $t = any(@a).pick(); say ref($t); say $t }; foo(1..3)
stevan that works?
ninereasons Int 15:20
2
max weird :)
stevan ninereasons: very nice
please add the passing case to the t/pugsbugs/junctions_any_pick.t and fix hangman to use that :)
(and make sure to add yourself to the hangman authors list
ninereasons some generous person has already added me, stevan 15:23
stevan svn blame must have listed you 15:24
ninereasons: you cant hide form svn :)
ninereasons :) 15:25
ooh. much nicer: 15:27
my @b = (1..3); sub foo (@a) { my $t = any(@a).pick(); say ref($t); say $t }; foo(*@b);
Int
3
there's the bug, then
diff @b *@b 15:28
stevan ninereasons++ 15:29
ninereasons eek. karma scares me 15:30
stevan why?
nothingmuch perlbot karma ninereasons:
perlbot ninereasons: doesn't have any karma
nothingmuch eh?
perlbot karma ninereasons
perlbot Karma for ninereasons: 1
nothingmuch what's so scary?
ninereasons doom and karma are wed in my mind ;) 15:31
nothingmuch ninereasons--
stevan doom the game? or doom as in "impending doom"
nothingmuch better? ;-)
mj41 mj is waiting for io_redirect_to_scalar 15:32
ninereasons better :) 15:33
nothingmuch waits patiently 15:34
mj41: ping
i won't leave you alone till you tell me what that example does 15:35
ninereasons sorry stevan; in conference 15:37
doom as in en.wikipedia.org/wiki/Karma
mj41 I can probably replace open $*OUT,">", \$out; open $*OUT,">", \$err; my $status = eval $new_pl; 15:39
with system('pugs > out.tmp');
autrijus mj41: src/Types/Handle.hs is the class api
but I had not implemented them fully in AST.hs
mj41: you fancy learning some haskell? :)
if not, just wait until this weekend and I'll pick up
mj41 ok, thanks, system is good enought for now 15:40
nothingmuch mj41: you are cruel
mj41 why? I think my English is cruel, not me. 15:41
nothingmuch i just want to know what that example from yesterday was
please please tell me 15:42
mj41 which example?
nothingmuch where you iterate $i and $i+1 at the same time
gaal_ zip. by2 15:48
(nm that was for you) 15:49
ninereasons stevan, did you want me to use that success case to edit hangman?
gaal_ actually, no 15:50
just take two every time
ninereasons stevan, or leave the assignment to $current_committer outside of the sub, as you have it now?
stevan ninereasons: sure
gaal_ nm did you have the boustrophedon example in mind? 15:51
wolverian mj41: open returns the filehandle in p6 15:52
pasteling "mj" at 147.229.221.107 pasted "interate $i ?" (11 lines, 389B) at sial.org/pbot/9296
jez 'Beginning Perl' suggests that it'd be more secure to set certain environment variables before calling DBI->connect() so that the pieces of connection information don't have to be specified directly in the connect() call. My question is simple - how is that more secure? 15:55
wolverian are you sure this is the right channel to ask that? 15:56
Corion [jez]: You don't store the connection data in the script itself, or in the command line (which shows up via top)
... but #perl6 is not the place to ask Perl5 usage questions, especially for (well documented) modules
jez perl6's default mode is perl5 compatibility ;-P
Corion jez: But there is no DBI.pm for Perl6 yet. 15:57
stevan jez: actually it isnt :)
Corion jez: But if you plan on porting it, I suggest you wait until we get objects
jez i wanted to make a suggestion for perl 6. 15:58
autrijus so, objects.
jez if you're going to verhaul the language, how about changing 'last' to 'break'?
autrijus what should I work on first when I get next batch of tuits? :)
jez i don't like the keyword last :-)
or, how about aliasing 'break' to 'last'
stevan jez: suggestions go to perl6-language
autrijus &last ::= &break;
jez yes, do that implicitly 15:59
:-)
autrijus macro last is parsed rx// { return 'break' };
wolverian hmm. isn't the [] around the <a b c> in mj41's paste unnecessary?
autrijus err, the other way around.
wolverian that is, is <a b c> a ref in scalar context?
autrijus jez: see, one good thing of perl6 is you get to define your dialects without imposing it on me :)
jez bleh, i can do that in C, but it annoyingly means i have to put that code in all of my programs 16:00
autrijus wants to call it "esc"
jez why not make it a default perl keyword, break isnt currently used is it?
whats wrong with aliasing it/
autrijus jez: I imagine some people name their subroutines break().
and aliasing it will cause them to break().
jez their problem, that'll be perl 5 code anyway
autrijus pun intended. ;)
jez perl 6 code would prohibit that
autrijus jez: cool. put macros and tests to t/Dialects/ 16:01
suggestions to p6l.
jez are you saying perl 6 adds no new keywords?
autrijus <- not language designer
wolverian jez: suggest this on perl6-language.
jez where is that
wolverian jez: we're not the people to make language decisions.
autrijus jez: I'm saying I'm this implementatino monkey :)
wolverian jez: [email@hidden.address]
(add -subscribe to, uh, subscribe. :) 16:02
ninereasons <autrijus> what should I work on first when I get next batch of tuits? :)
wolverian also dev.perl.org/perl6/lists/
stevan autrijus: I think that making multi-subs dispatch on type would be a good thing
autrijus: see t/subroutines/multi_sub.t
ninereasons how about negative indexes, autrijus?
jez *searches email archive*. ah yes, seems i already suggested it.
though, i suggested it in comp.lang.perl.misc,comp.lang.perl
is that sort of equivalent to the mailing list? 16:03
wolverian no.
(at all.)
jez hmm 16:04
can i see archives of the list?
wanna see if someone else mentioned it
wolverian groups.google.com archives it
although it has interface issues. let me get the real link. 16:05
jez it's a usenet group?
wolverian groups-beta.google.com/group/perl.p....language/
jez it looked like a mailing lisat
wolverian no, it's just kind of mirrored. :)
(on google.)
jez hte group or the list?>
wolverian I mean, it is a mailing lists.
jez o
wolverian s,lists,list,
hmm. is it ok that pugs installs libs to /usr/lib/perl6/5.8/* 16:08
that seems.. odd.
autrijus mm, negative index
which test tests that?
ninereasons there must be something in arrays. 16:09
checking
wolverian (I mean the 5.8 part.)
autrijus wolverian: I agree.
however I don't see that :)
autrijus thinks
wolverian well, it happened here. :)
autrijus wolverian: well, /util/PugsConfig.pm line 138 is all yours 16:10
wolverian agh! it's haskell! 16:11
ninereasons howaboutthat, not a single negative index test in array.t, autrijus
stevan oooo then do mine :) 16:13
I have tests :)
ninereasons t/subroutines/multi_sub.t wins the horserace
jez see i never saw the problem with goto. :-)
has that been removed from perl6?
as long as you only ever use it in place of something like last MYLOOP 16:14
stevan jez; no
jez i don't like the semantics of 'last'
my mind thinks 'make this the last iteration of this loop'
which would be the logical equivalent to next
ninereasons logical corrolary
jez yes 16:15
ninereasons corollary ;)
wolverian I never liked 'last' either. it reads to me like 'go to the last iteration'.
autrijus ninereasons: not a good excuse. write tests! trade tests for feature!
ninereasons autrijus, sir, yes sir!
stevan autrijus++ # implementing the barter system
autrijus ninereasons++ 16:16
jez um, could anyone point me to thread(s) on the perl6-language list that discuss the changing in perl 6 of variable type symbol usage? ($,@,%)
wolverian <a b c>.kv fails.
jez: I'd just read the apocalypse. 16:17
stevan jez: not sure that was ever up for discussion
wolverian also, qw() shouldn't really work, since it's a sub call. mm. I should just see if there are tests for these. :/ I'm not used to this test driven development thing! 16:18
Corion I think the more consistent use of sigils was one of the first things decided
(as it is the most confusing thing for new beginners)
wolverian I seem to remember so as well. 16:19
16:19 _metaperl is now known as metaperl
ninereasons did we learn what should happen when we say, '@a = 1..Inf; say @a[-1]' ? 16:21
autrijus yes; it makes you wait a long while. 16:22
wolverian wait for the universe to end? :)
autrijus at least, I don't see a better implementation.
ninereasons shouldn't it automatically "shorten()" or curry (if that's the right terminology)
?
or die horribly? 16:23
autrijus we may implement die horribly later.
anything else is probably the bad idea.
16:23 Corion_ is now known as Corion
wolverian the .kv tests don't test <a b c>.kv. I could add that, I suppose! 16:23
added a test. 16:27
gaal_ beh, what is that even supposed to mean? 16:28
wolverian my $array = <a b c>; $array.kv
gaal_ and what is it supposed to mean if there's an even amount of elements?
oh wait kv doesn't return pairs; pairs does.
wolverian yes! 16:29
Khisanth what is .kv for arrays? 16:30
index and values?
wolverian Khisanth: yes.
autrijus: I guess I could ask for commit access now that I have this test added.
autrijus wolverian: what, you're not a committer?
wolverian no, I'm not.
autrijus that's a crime. :) your email?
wolverian [email@hidden.address] 16:31
autrijus very nice domain.
wolverian thankyou. :)
autrijus invitation sent; welcome aboard.
wolverian I want to go on the balcony and yell incoherently a bit. I think I'm taking this a bit too seriously, maybe. 16:32
autrijus that's fine. we will wait :)
(yell incoherently)++
mj41 say( ('-' x 60) ~ "\n" ~ $out ~ "\n" ~ ('-' x 60) );
say ('-' x 60) ~ "\n" ~ $out ~ "\n" ~ ('-' x 60); # ee
autrijus mj41: what's wrong with that? 16:33
Corion mj41: Well, under Perl5 you would have got a warning about that
autrijus: I guess it prints "1" ~ "\n" ~ $out
(and so on)
wolverian I'll go buy some cat food and commit once I'm back. bye. 16:34
Corion because the parentheses bind directly onto the say() call
wolverian I thought perl6 has a whitespace rule that makes those lines equivalent.
okay, gone. ->
mj41 so say ~('-' x 60) ~ "\n" ~ $out ~ "\n" ~ ('-' x 60); # DWIM 16:40
putter what is the syntax for "unusual" sub names? eg, sub empty? () {...}. currently this doesnt parse. the docs give examples of such operators, but no subs/methods/etc. foo? <<foo?>> something else? 16:43
ninereasons then, remove the parens, mj41
mj41 :-) thanks 16:47
ninereasons :)
mj41 pugs template toolkit wanted :-) 17:07
autrijus ooh I see array.t 17:13
ninereasons++
"trade tests for feature" mode activated. 17:14
tests passes; it's implemented now as r1942; enjoy.
Limbic_Region one minute turnaround from test presence to feature implementation 17:16
Limbic_Region debates writing several tests while autrijus is in the "trade tests for feature" mode
autrijus noooo! I'm in this "deadline approaching, must write perl5 code for a living" mode 17:19
Juerd scw: Times are rather pointless without timezone.
scw: And really, many geeks don't live in normal time :)
kungfuftr ullo all 17:23
ninereasons negative indices++ ;
Odin- Juerd: What is normal, anyway? 17:24
(UTC, naturally. That's the "standard time". ;)
kungfuftr UTC/GMT 17:28
seen nothingmuch
jabbot kungfuftr: nothingmuch was seen on Wed Apr 13 15:42:45 2005 GMT
kungfuftr bah 17:30
autrijus I want more tests for neg idx on lvalue context 17:31
and on exists, on delete.
any takers? :)
(it's actually a quite complicated problem.)
kungfuftr bah, darc is brokn 17:32
ninereasons I'll add some neg idx tests when i have a minute
Corion r1937 - datenzoo.de/pugs/win2k.html 17:41
autrijus hmm, nm's new code has a final line of summary/subtotals 17:43
stevan horay I am finally registered for YAPC::NA 17:44
stevan has been lazy/pugs-obsessed lately
kungfuftr when's nothingmuch usually about? 17:48
stevan kungfuftr: hard to say,.. some days he never sleeps 17:56
kungfuftr heh 17:57
no petal ports... dammit
ninereasons isn't that one of Ovid's favored projects, kungfuftr ? 17:59
kungfuftr no idea off hand
ninereasons I'm actually a little more sure than i let on. it is one of his favored projects :) 18:00
you might pounce him when he's about, and see what how he's progressing 18:01
kungfuftr ah
=0/
Corion r1944 - datenzoo.de/pugs/win2k.html
tiw hi all 18:16
Corion pugscode.org <<Overview Journal Logs>> | r1773/winxp: 587/4256 smoke/win2k (occasional): xrl.us/fqum (10/4279) | pugs.kwiki.org 18:18
jez heh 18:19
"It's difficult to deal with any subset of Unicode without naming it. These days, if you see [a-z] in a program, it's probably an outright bug."
what exactly does Larry mean by that?
tiw how to dereference a array reference in perl6, i try @$ar and @{$ar} but none of them works
PerlJam jez: using a named character class rather than a-z. e.g., [:alpha:] or [:alnum:] etc. 18:20
gcom jez: a-z assumes pretty much english (or romaji ;) ) only
PerlJam tiw: that action is no longer needed. $ar[0] is the same as @a[0] if $ar = @a; 18:21
jez: sorry I don't happen to know any fancy unicode names to use as examples :)
tiw PerlJam: but @a = $ar, just copy $ar to the first element of @a 18:22
PerlJam tiw: oh, good point. Hmm.
jez gcom: true, but some of us only speak English :-)
and only care about coding for English
theorbtwo It also doesn't catch the proper spelling of, say, that thing you hand a potential employer that isn't quite a CV. 18:23
PerlJam jez: and that's so last century. Perl6 is about the 21st century and beyond.
ninereasons autrijus, did you see my hat trick? 18:24
PerlJam tiw: I think you should ask on p6l. That may be one of those areas that isn't quite ironed out.
gcom jez: isn't the point of view that's being suggested as a "bug" (i do [a-z0-9] all the time btw, lazy habit, but i agree in principle with what i think L is saying
jez: and yeah, usually there are records or inputs that i'm losing by allowing the bug in my code 18:25
tiw PerlJam: i am reading perl6 and parrot, in it @$ra and @{$ra} are used to dereference
PerlJam tiw: Then, I guess it just hasn't been implemented yet. 18:26
castaway even english has occasional accented chars in it
PerlJam castaway: and ligatures and whatnot if you're talking about the queen's english.
castaway liga-whats? ;)
jez cast: if that's the case then how is :alpha: gonna help? 18:27
castaway it contains accented chars
jez btw, i think saying that coding with the assumption that Roman characters are going to be used is last century is pretty stupid. :-) 18:28
sorry, we're all suddenly using foreign characters now?
ninereasons we're all the sudden using unicode
PerlJam jez: no, coding with the assumption that non-latin characters *won't* be used is pretty stupid.
ninereasons perl6 is a language written in unicode
castaway not all of a sudden no, have been for years
jez unicode doesnt have one single standard 18:29
theorbtwo Yes, it does.
It doesn't have one single /encoding/, but we're talking one level above encoding at present.
PerlJam is >this< close to flipping the bozo bit on jez out of shear annoyance :)
jez is a unicode char always 2 bytes? 18:30
theorbtwo No.
PerlJam no
jez but it can be 2 bytes 18:31
PerlJam certainly.
castaway depends on the encoding
theorbtwo Well, it can be usually two bytes.
It'll sometimes be four, even in utf-16.
jez is glad he won't be deal with that in C. 18:32
*dealing
theorbtwo (ucs-2 is never four-byte, technically speaking, but it doesn't encode all of unicode, so don't use it.)
jez the problem i have with Unicode is it's necessarily damn complex, ASCII isnt
PerlJam the point remains though that you aren't in kansas anymore (unless you specifically want to be ;-)
jez you can't really overcome that one
theorbtwo No, you can't.
gcom thinks .chars is a pretty clean way to handle the complexity 18:33
PerlJam jez: unicode and ascii intersect nicely. Choose your level of poison.
jez actually you could if they'd made every unicode char 4 bytes
:-)
theorbtwo But ASCII is simple precisely because it throws out the languages spoken by at least 80% of the world.
...and throws out 5% of the language spoken by the other 20%.
jez mmmhmm, but a large part of that 20% of the world also throws out those languages 18:34
:-)
gcom reverse 80/20 rule, hehe
jez that's the point
gcom call that 20/80
PerlJam random fact: IBM at one time was promoting both EBCDIC and ASCII "standards"
jez why bother with foreign characters when you're not coding for foreign languages?
PerlJam jez: right. If you want to live in that sandbox, you can. 18:35
jez *chuckles* how is that a sandbox?
tiw PerlJam: Can not find test about dereference :-(
castaway because, as mentioned, english has some too.
theorbtwo Because you still have to think what you're going to do when somebody tries to spell rƩsume correctly.
jez i'm talking more than a few accents; greek, chinese, japanese, korean, arabic... etc 18:36
castaway because upi'll get a shock should you ever want to locaise something
theorbtwo If you punt safely, you've already done 90% of the work needed to just support it.
PerlJam jez: Because the rest of the world is interacting with each other and that means that if you've restricted yourself to ASCII, you've decided you aren't playing with the rest of the world. You're in your own little sandbox.
castaway "you'll"
PerlJam jez: There's nothing wrong with that mind you. I expect that there will be many sandboxes that spring up around the world.
jez perl: i'm not interacting with them anyway, i only speak English competently. few people speak more than 2 or 3 languages well
theorbtwo jez, supporting Ć© correctly isn't harder then supporting greek, chinese, japanese, korean, arabic, etc, for many applications.
jez but it's a lot harder than supporting ASCII ;-) 18:37
PerlJam jez: but perl6 supports the global view generally and the local view as a specialization.
Aankhen`` ōæ½xABjezōæ½xBB: You're restricting every one of your users to English too.
theorbtwo Not if you're using perl6, it doesn't.
Aankhen`` And that's not a decision you ought to be taking for them.
theorbtwo Er, it isn't.
jez anak: oh i dunno... this world would be better with a single language. :-)
castaway you're not interacting, your software may be 18:38
jez much less hassle
Aankhen`` Yeesh.
PerlJam Aankhen``: let him make that decision. As soon as he writes some useful software he'll find out that there are people in China or India or Germany or where ever that want to use it.
jez I'd learn Esperanto if everyone else did. :-)
Aankhen`` Try this, jez: type in 'Aank' and press Tab.
gcom or as soon as he wants to check it into cpan
jez Aank: what's your point
people in China will wanna use my *english interface* software? 18:39
why?
gcom thinks there's a little trolling going on, but is leaving room for doubt
Aankhen`` I find it very irritating when people refuse to use nick completion and end up misspelling my nick.
ōæ½xABgcomōæ½xBB: You too? ;-)
jez Aankhen``: it's a habit i've never gotten into
Aankhen`` Then learn to spell nicks right.
theorbtwo Do you mind Aankh?
PerlJam jez: for instance, I guy I know wrote some wiki software. He was very american-centered until it became popular. Now he's got some internationalization because he has users all over the world.
Aankhen`` ōæ½xABtheorbtwoōæ½xBB: Of course not... Aank, Aankh, Aankhen or even Anakin works. :-) 18:40
jez you, OTOH, seem to like specialformatting
Aankhen`` ōæ½xABjezōæ½xBB: That's right.
I just dislike Anakhen or Anaken or... you get the idea.
theorbtwo dislikes "theo". 18:41
jez perljam: does he communicate with said users in non-englis?
PerlJam jez, Aankhen``: both of you know how to use msg
jez English
PerlJam jez: occasionally. But the point is that his *software* has to
jez PerlJam: how do i use msg? *ducks*
Aankhen`` ōæ½xABPerlJamōæ½xBB: Better, I know how to go off and waste time playing games. :-D
wolverian hm, how do I determine whether a test I'm adding is todo or not?
Aankhen`` Later.
PerlJam wolverian: Is it an unimplemented feature? 18:42
wolverian PerlJam: yes.
castaway does it work?
wolverian castaway: no.
jez perl: welllll... wiki software is the kind of thing i might expect that for. however, what about a different example - database logins.
castaway then its todo, obviously
PerlJam it should be a todo then :)
jez perl: i'm gonna support unicode usernames/passwords?
wolverian well, I don't know if it's implemented or not.
castaway I presume you know that todo means "to do" (as in to be done)
wolverian but it definitely doesn't work.
yes, but the first issue is unclear to me.
castaway if it doesnt work, it needs doing ,) 18:43
wolverian heh.
castaway (assuming its supposed to)
jez perl: as this database is for my english website i dont anticipate any non-english needs, and it's not worth my time supporting it
i guess i will when i feel the need
:-)
PerlJam jez: sure.
jez but it will have to come up and bite me
castaway bites jez.
(oops) 18:44
PerlJam jez: right. Like I said before, I expect that there will be many sandboxes springing up around the world until we have full globalization :-)
jez you mean a single world language? heh
i'd like to see 'full globalization' until that mile-high barrier is broken
that or a universal translator device. i've been wondering why one of those hasn't spprung up. we have speech recognition, we have translation, we have speech synthesis 18:45
PerlJam jez: English is the "world language" But I doubt that ethnicities or regions will be giving up their local languages any time soon.
jez just need more research and a universal translator (rudimentry) could be developed
PerlJam jez: you've obviously never studied linguistics. 18:46
machack666 you can solve hard ai next, then. :)
jez i said it'd be rudimentry. so initially it wouldnt carry emphasis, there would be mistakes
PerlJam jez: though you're right. It's just that *only* a rudimentary translator could be developed. :)
machack666 and p = np needs some work, too
jez but it would be ok for clearly expressed sentences
castaway also hasnt tried to tranlate stuff using babelfish much either
jez perl: nah, just look at Star Trek 18:47
PerlJam heh
jez babelfish is poo. there must be better translations around
promt.ru is ok
castaway feel free to write one ;) 18:48
mj41 So stupid, so beautiful. Perl6-pugs powered
wiki.kn.vutbr.cz/mj/attach/pugs/tut...ld.p6.html
wiki.kn.vutbr.cz/mj/attach/pugs/tut-pugs/tut/
need some food
ninereasons iblech, maybe "isnt" would be better than "is" to test your bug discovery? may I change it? 18:50
jez rafb.net/paste/results/2WfdzS70.html
could someone explain how the above increases security? 18:51
or is supposed to?
Corion jez: Maybe ask that in #perl 18:54
PerlJam jez: It doesn't. Anyone who can read the source can see those values.
rgs simon cozens is to be trusted, though 18:55
Corion rgs: Simon Cozens is smoking crack. Good crack some times, but still crack
rgs right.
Corion (but yes, he usually knows what he's talking about :) ) 18:56
jez Perljam: well that's what i thought. 18:58
it's strange to see Simon saying that because it's BS, moving the password to somewhere else in the code increases security?
machack666 DBIx::NamedDSN stores auth info in a place outside of the webroot</plug>
jez people shouldnt see the code in the first place
PerlJam jez: if you get on irc.perl.org, you can query simon himself on #perl
crysflame cozens is very interesting
jez perl: oh that's where he hangs out eh
rgs cozens is lathos on irc
jez he said EFNET #perl
crysflame his nickname is lathos there
jez :-)
castaway You can set those vars in the scripts environment some other way.. readonly login files or something
crysflame EFNET #perl, Rhizo #perl, whatever
PerlJam jez: oh, he's on efnet too I think.
castaway but if a usder can read the script, he can still find them 18:59
integral thinks it does increase security
PerlJam jez: most of the famous (or infamous) perl people hang out on irc.perl.org #perl though
integral using the user's credentials rather than the same ones for all users (plus you get to use krb5 ;-)
PerlJam integral: okay, sure it increases security but only marginally.
castaway slightly
integral Not really, different users can use different credentials, and you get auditing
Corion Storing login/password in the script is bad, because if you post the script to the internet, you got to remember to clean it before posting. 19:00
(as I've experienced often enough)
PerlJam Corion: sorry but, heh. :)
integral doesn't see the point of using a login/password when it's not provided by a user
kungfuftr @ARGV?
castaway plenty of point for that, integral
integral like? 19:01
But it's not buying you anything if it's just there in the script.
Corion kungfuftr: Visible through ps, or top.
castaway like when the app only has/needs one login to the DB, and the users just use the app, and arent supposed to know how to get at the db directly
stevan mj41: very nice!!
integral but then castaway any method you employ is merely obscurity. Using strings(1) on your app isn't hard 19:02
castaway it is if my app is on a server they dont have access to ,)
integral right, then their pc shouldn't be allowed to talk to the DB at all :-) 19:03
jez so like, what is the safe way to store user/pass for db access?
crysflame i can't think of any
machack666 encrypt it, and find way to hide the decryption key. induction strikes again! :D 19:05
castaway somewhere where it cant be read
jez castaway: hurm. howsabout..... beneath my docs directory? 19:07
then the script could look in ~/passwords.foo
castaway no idea what type of app you are talking about
jez CGI
of course your script would have not to be so shit that it would print out passwords.foo 19:08
but it's another layer of security
mj41 stevan: thanks 19:09
stevan mj41: is that in the pugs source tree yet? 19:16
mj41: you might want to try pointing it at the examples/cookbook stuff too
mj41 not yet in pugs, to utils or examples? 19:18
are utils are p5 19:19
s/are/all/
or somebody can create Tutorial-Generator ext/module 19:23
rindolf Hi all! 19:25
stevan mj41: put it in examples/tutorial_gen/ 19:37
hey rindolf 19:38
rindolf stevan: hi
stevan: what's up?
stevan nothing really,.. 6.2.0 hangover basically :)
mj41 ok 19:40
stevan mj41: I would love it if utils/ would get ported to perl6, but realistically Pugs is still to slow in some places 19:41
mj41 yes, I can confirm this 19:42
stevan mj41: :)
mj41: 'make optimized' makes it a little better
mj41 puzzle 19:43
my $a = <a b>; my $b = <a>; my $c = [ <a> ]; my $d = [ <a b> ];
say $a.ref ~ $b.ref ~ $c.ref ~ $d.ref;
stevan that makes sense 19:44
<> is general quoting IIRC
so it is just quoting that string
or rather <> is sorta-polymorphic quoting weird thing 19:45
wolverian the second case is a bit weird, but it's the same in perl5.
gaal i don't understand why $a isa List and not Array though. 19:46
stevan has not quite grasped the distinction between array and list in perl6
gaal are lists var types at all?
PerlJam stevan: the distinction is very very small in perl6 19:47
gaal killed a few trees today and will have plenty of time to read the apocalypses soon :/
wolverian is there a does or isa relationship between List and Array?
stevan wolverian: check,.. we have .isa now 19:48
wolverian right. thanks. :)
okay, there is no isa relationship. 19:49
just wondering if I have to type things as 'List|Array $foo' in my signatures.
castaway doubts it 19:50
wolverian then both should do the same role
(such as Enumerable in Ruby)
gaal what's p6 for print for @Foo::INC ? 19:51
wolverian or maybe pugs is wrong here.
stevan gaal: not sure 19:52
ninereasons kungfufter for some reason it just hit me: when you said 'petal ports' you meant 'as in freebsd' not 'as in 'perl5 => perl6'
Aankhen`` G'night.
wolverian $obj.meta.isa
stevan wolverian: I think we do actually (have List|Array)
ninereasons kungfuftr, i hate leaving loose ends :)
wolverian stevan: I wasn't asking if we have it, but if I need to do that every time I want an enumerable type 19:53
stevan wolverian: see S29 I think I saw an example of that usage in there
kungfuftr moo?
rob_ Unpacking ghc-6.2-i386-unknown-linux.tar.bz2
stevan wolverian: :) I am a little slow today
rob_ ups, sorry
castaway oops, crashed SEE..
gaal thanks, wolverian
wolverian the obvious solution here is to add a role such as Enumerable. I don't know if this has been discussed. 19:54
gaal: you're welcome.
castaway looks for nothingmuch
kungfuftr ninereasons: yar... no bsd ports 19:56
wolverian (the name is bad with the 'does' verb, so it should probably be Index (class List does Index { ... }) or some such.
s/$/)/
mj41 uploaded 20:06
g'night 20:29
from Czech republic
rindolf mj41: night 20:30
gaal: hi.
theorb Allo, all. 21:16
ingy hi theorb 21:19
y "hola" en generales 21:20
<-- makes stuff up
stevan ingy-speak 21:33
metaperl_ stevan, I broke up fp.p6 in examples
I needed to do so for my slides for my talk tonight: www.metaperl.com/talks/p6-fp-slurpy/index.html
stevan metaperl_: ok by me 21:34
metaperl_ ok
stevan silently curses metaperl
metaperl_: did you commit it yet? 21:35
metaperl_ I think the original is still there
there is a directory...
fp.p6 is still there
I added fp.p6
and an fp directory which has fp.p6 broken up 21:36
all under examples
stevan metaperl_: I am not seeing it
I just grabbed r1953
metaperl_ oh maybe I didn't commit it
hold
stevan metaperl_: I would prefer you remove fp.p6
or at least move it in the directory 21:37
no sense in having 2 copies of it actually
metaperl_ ok 1954 is committed but I did not remove your fp.p6 21:39
stevan metaperl_: re: your fp.p5... most of fp.p6 was actually derived from this module search.cpan.org/~stevan/fp-0.02/lib/fp.pm 21:40
metaperl_ yes, many ways to do it - Language::Functional is another
stevan have you filtered out the sw-design moderator notices? I am about to... there doesn't seem to be a way to handle it via the UI 21:41
stevan metaperl_: they have been going to my junk folder for a while now 21:42
metaperl_ ok, I'm going to do that as well
stevan metaperl_: I am moving fp.p6 into examples, and removing examples/fp.p5 since it is in examples/fp/ now 21:43
ok?
metaperl_ ok 21:44
theorb Is the polite thing to do to commit tests as todo if they fail, or only if they represent a feature that shouldn't be done by now? 21:53
kungfuftr has a new really nice way of doing testgraph 21:54
but lest i know not petal 21:55
theorb shrugs -- I thought the old way was just fine myself.
kungfuftr =0) 21:57
theorb Hair standing on end and a pig-nose? 21:58
stevan theorb: todo if its not done yet, not-todo if it should be done (or is a bug) 22:01
thats my general approach
metaperl_: I just added another example which you might find ammusing 22:04
off to eat dinner,... bye all 22:08
theorb Later, stevan!
stevan oh hey,.. theorb did you confirm all your reservations for YAPC::NA?
I talked to John about the hackathon, it is going to start on Wed now (1 day earlier) 22:09
theorb Well, I've RSVP'd for the hackathon... and Jess booked the time off.
But we've not done the rest of the arragnements yet.
stevan ok, well just so you know if you can get their earlier then cool :) 22:10
23:53 metaperl__ is now known as metaperl_