Pugs 6.2.8 released! | pugscode.org | pugs.kwiki.org | paste: sial.org/pbot/perl6 | www.geeksunite.net
Set by autrijus on 13 July 2005.
06:48 Boogie_ is now known as Boogie 06:59 castaway_ is now known as castaway 08:43 _jql is now known as jql 08:56 Maddingue__ is now known as Maddingue
svnbot6 r5884 | iblech++ | * Usual svn props. 09:14
r5884 | iblech++ | * PIL2JS:
r5884 | iblech++ | * Cleaned up assignment rewriting rules (i.e. my $scalar = @array).
r5884 | iblech++ | * my @b = @a; @b[...] = ... changed the original @a, fixed.
r5884 | iblech++ | * my %h = $scalar did only work if $scalar was a Pair; now it works with all
r5884 | iblech++ | scalars.
r5884 | iblech++ | * Prelude::JS::Operators -- Added &abs.
r5884 | iblech++ | * Prelude::JS::Array -- &min and &max (but doesn't work yet correctly).
r5885 | iblech++ | PIL2JS: 10:05
r5885 | iblech++ | * PIL2JS.js -- Fixed bind cycle detection rules (@a[1] := $b; $b := @a[1]).
r5885 | iblech++ | * Prelude::JS::Array -- Fixed &min and &max, all tests pass:
r5885 | iblech++ | m19s28.vlinux.de/iblech/stuff/not_p...max.t.html
r5886 | iblech++ | PIL2JS: Prelude::JS::Array -- Fixed map { [...] }, all tests of
r5886 | iblech++ | flattening_in_map.t now pass:
r5886 | iblech++ | m19s28.vlinux.de/iblech/stuff/not_p...map.t.html
wilx Heh. Looks like iblech is really serious about this :) 10:21
theorbtw1 wonders if he'll be able to do the interesting bits of google maps hacking in p6 any time soon.
10:21 theorbtw1 is now known as theorbtwo
Aankhen`` wilx: I should think so... the JS backend is the most advanced out of all the backends supported by Pugs. 10:26
castaway shouldnt it be a parrot backend/whatsit? 10:27
Aankhen`` castaway: ?
theorbtwo No, parrot is a backend.
castaway Hmm, parrot only runs bytecode, it doesnt turn it into other things? (I guess not, what was I thinking of..) 10:31
theorbtwo It does also turn it into other things... but only as a side-effect, mostly.
It will turn pasm into x86 asm, for example, but only because it does that anyway, in order to get it to run fast on x86es -- packaging it up into an executable after that was easy, so it does it. 10:33
castaway I see 10:34
(I think)
I was actually just thinking one can create parrot bytecode from JS, but thats a different problem
theorbtwo That's a very different problem. 10:35
castaway right 10:36
svnbot6 r5887 | iblech++ | * New test: t/builtins/lists/undef_in_map.t 10:41
r5887 | iblech++ | +map { () } @foo should be 0 (works in normal Pugs, but does not in PIL2JS).
r5887 | iblech++ | * New test: t/builtins/lists/mutating_listops.t
r5887 | iblech++ | map { $_ = ... }, grep { $_ = ... }, for @foo { $_ = ... } should all work.
r5887 | iblech++ | Does not work in normal Pugs, passes 5/5 in PIL2JS :):
r5887 | iblech++ | m19s28.vlinux.de/iblech/stuff/not_p...ops.t.html
r5887 | iblech++ | * PIL2JS: Prelude::JS::ControlFlow, Prelude::JS::Array --
r5887 | iblech++ | &statement_control:<for>, &map, and &grep all support modifying the array
kgftr|konobi castaway++ # JS on parrot
svnbot6 r5887 | iblech++ | they're iterating over now.
castaway eh, what, me? 10:43
(note to self: the longer the svn commit texts are, the more karma one gets..) 10:44
theorbtwo Yep.
You even get karma for the blank lines in there.
castaway nifty, no? 10:45
wilx It works but it is slow. 10:55
Somebody should push the Mozilla guys to make their JS implementation faster :)
castaway mebbe JS needs a bytecode form too 10:57
coral that'd be nice 10:58
theorbtwo No reason the bytecode form of JS can't be pbc.
coral castaway: heh!
castaway pbc? 10:59
theorbtwo Parrot Byte Code.
wilx @karma iblech 11:00
?karma iblech
Hmhm.
castaway karma iblech
jabbot castaway: iblech has karma of 337
theorbtwo top karma
karma highest
jabbot theorbtwo: highest has neutral karma
theorbtwo jabbot, help
jabbot theorbtwo: Tell me more about that.
castaway hmm, get Moz to be a parrot vm?
Odin- That would be interesting. 11:01
theorbtwo ...but at least socialy difficult.
castaway well except that there are "other browsers" too :)
theorbtwo No idea how technically difficult.
Aankhen`` perlbot top 10 karma 11:05
perlbot The top 10 karma entries: C: 57, iblech: 19, Yaakov: 11, gaal: 10, stevan: 9, C/C: 9, japhy: 8, putter: 8, integral: 7, autrijus: 7
Aankhen`` jabbot, top 10 karma
jabbot Aankhen``: Does talking about this bother you?
Aankhen`` Ah well, 'twas worth a shot. ;-)
Hmm... can I create a class that behaves like a hash, sigil and all included? Something that works along the lines of: "my %foo = Foo.new();" 11:28
Aankhen`` wonders if this might be achieved by passing a different representation name to `CREATE`, as mentioned somewhere in A12. 11:31
?eval class Foo { method new (*%_) { return ::?CLASS.bless(%_); } }; Foo.new(:a<b>, :c<d>)
Bleh. 11:32
masak has just read chromatic's "Porting Test::Builder to Perl 6" on perl.com 13:46
would it be too early to build a pugs Test::Base module?
castaway is it ever too early? :) 13:48
masak i have no idea, that's the point :)
i would like to see a Test::Base for pugs
(and autrijus would too, i seem to remember) 13:49
castaway well go for it then :)
masak but i don't know if it's feasible yet
maybe the easiest way to find out is just to try...
i imagine it would be a cross between chromatic's new Test::Builder and ingy's perl5 Test::Base somehow 13:50
or rather a rewrite of the latter using the former :)
integral hmm, is svnbot6 not working? 15:06
svnbot6 r5888 | bsmith++ | Added grouping and ranking of nodes to util/importgraph.pl.
r5889 | bsmith++ | Deleting Pugs.Compile.Pugs2 as it is unused, and little more than a skeleton.
r5889 | bsmith++ | It is also obsolete since the new PIL system is now used for compiling.
integral ah, there is is :-)
15:06 khisanth_ is now known as Khisanth
autrijus I'll sleep now. :) feedback welcome on the type inferencing proposal before I put it to action tomorrow :) 15:08
# www.nntp.perl.org/group/perl.perl6....uage/22519
& 15:09
integral that seems nice and comprehensive 15:16
masak every aspect of perl6 seems like a feature slightly too good to be true/too complicated to implement 15:38
svnbot6 r5890 | bsmith++ | Added export lists to lots of haskell files.
masak but OTOH, here's pugs doing most of it already :)
svnbot6 r5891 | bsmith++ | Moved getLibs from Pugs.Prim.Eval to Pugs.Run. 16:32
r5891 | bsmith++ | Removed some extra references to Pugs.AST.Internals where Pugs.AST is already
r5891 | bsmith++ | imported.
bline Pugs.Prime.Evil 16:34
brentdax bline: That's the Python implementation, right? 16:59
svnbot6 r5892 | bsmith++ | Added parseNatOrRat function to Pugs.Parser.Number.
r5892 | bsmith++ | Added module exports to Pugs.Parser so only it needs to be imported.
Aankhen`` ?eval class Foo { has $.bar; sub baz () { $.bar = "quux"; $.bar } }; Foo.new.baz 18:29
evalbot6 Error: No compatible subroutine found: "&baz"
Aankhen`` ?eval class Foo { has $.bar; sub baz () { $.bar = "quux"; $.bar } }; Foo.new.baz()
evalbot6 Error: No compatible subroutine found: "&baz"
Aankhen`` Oops.
?eval class Foo { has $.bar; method baz () { $.bar = "quux"; $.bar } }; Foo.new.baz()
evalbot6 \'quux'
Aankhen`` Hrm.
Is an attribute without `is rw` still rw for all methods in that class? 18:30
Darren_Duncan it should be 18:32
Aankhen`` OK.
Darren_Duncan I believe is rw is default for attributes
Aankhen`` Ah.
Darren_Duncan whereas the opposite is default for methods 18:33
Aankhen`` Yeh.
Darren_Duncan if you think about it, attributes are meant to hold data different for each object, so making them read-only by default doesn't make sense
Aankhen`` I guess so.
But then are they rw outside the class too? 18:34
autrijus my subtype Duck 18:49
has $.half_life
can doom:()
can quake:(() returns Wolfenstein);
mmm pretty. :)
Aankhen`` Wha... my last merge altered . for no reason... 18:50
autrijus with svk?
Aankhen`` Yes. 18:51
svnbot6 r5893 | Aankhen++ | * HTTP::Headers: fixed small mistake in &:sorted_field_names.
r5894 | Aankhen++ | * HTTP::Message: consistent indentation, more specific declaration of $:headers.
r5895 | Aankhen++ | HTTP::Request::CGI:
r5895 | Aankhen++ | * updated documentation.
r5895 | Aankhen++ | * added $.PARSE_ARGV variable to control whether command-line parameters should be parsed.
r5895 | Aankhen++ | * replaced $:query_string and &query_string with single $.query_string.
r5895 | Aankhen++ | * made declaration for $.query more specific.
Aankhen`` The root directory, rather.
autrijus svk writes merge tickets as props to .; that's fine (I think)
Aankhen`` OK.
As long as I didn't break anything. :-P
That was my first commit from a branch. :-D
autrijus right, which is I think the only time svk does that 18:52
Aankhen`` Works for me. 18:53
Sky seems to be staying up, so I guess that's alright.
autrijus cool 18:54
dudley what does the colon in foo:() mean? 18:56
Aankhen`` dudley: Picks a particular variant of a sub, AFAIK.
Khisanth adverb?
autrijus dudley: it's the long name syntax
it's usually written as &doom:()
come to think about it I should've written that instead. 18:57
but maybe I should've just written
my subtype Duck
method doom () {...}
etc
not sure. anyway the surface syntax is not the point
the point is a way to encode the fact that people often write UNIVERSAL::can() checks in p5 instead of UNIVERSAL::isa() 18:58
and somehow give .can the same static checking status as .isa
(or .does) 18:59
dudley oh, I know, I've just seen it a couple of times before and was curious
autrijus ah. then yes... &infix:<*>:(Int, Int) selects the variant that can instantiate this signature
Aankhen`` Is the long name notation supported in Pugs? 19:00
autrijus not yet 19:01
Aankhen`` OK.
Can I create a variant of a sub by assigning to that variant, like a normal sub assignment? Example: &foo:() = &bar; &foo:(Int,Int) = sub (Int $foo, Int $bar) { ... }; 19:02
svnbot6 r5896 | Aankhen++ | * expanded documentation for HTTP::Query (SYNOPSIS and DESCRIPTION) and HTTP::Request::CGI (SYNOPSIS).
Aankhen`` Gurk.
Aankhen`` bangs forehead on wall.
I really should check out my changes before I check them in... 19:03
autrijus Aankhen``: I suppose the lhs will fail to lookup first before you attempt to assign to it 19:06
Aankhen`` OK. 19:07
autrijus Aankhen``: the canonical way to do that is a "multi" :)
Aankhen`` Yeh, I know.
It's just that I'd rather replace `multi sub foo () { return bar(); }` with a direct assignment, the same way you alias subs normally.
s/assignment/binding/ 19:08
autrijus nod, as I suspect assigning into &foo is as fun as assiging into \$bar. 19:09
svnbot6 r5897 | Aankhen++ | * really expanded DESCRIPTION for HTTP::Query. 19:14
Aankhen`` Rest break. & 19:16
?eval role Foo { ... }; class Bar does Foo { ... }; 19:30
evalbot6 *** ... - not yet implemented at <eval> line 1, column 12-16
Aankhen`` ?eval role Foo { }; class Bar does Foo { ... };
evalbot6 *** ... - not yet implemented at <eval> line 1, column 36-40
Aankhen`` ?eval role Foo { }; class Bar does Foo { }; 19:31
evalbot6 undef
Aankhen`` How is a slurpy List denoted in the long name of a sub? 19:36
svnbot6 r5898 | Aankhen++ | * HTTP::Query: rewrote &param (rw version). 19:43
r5899 | Aankhen++ | * HTTP::Request::CGI: changed $r in &BUILD to $self for consistency. 20:01
r5899 | Aankhen++ | * HTTP::Request::CGI: small formatting changes; more flexible implementation of &BUILD.
r5900 | iblech++ | PIL2JS:
r5900 | iblech++ | * PIL::Parser, PIL::Nodes -- Added basic support for Inf, -Inf, and NaN.
r5900 | iblech++ | * Prelude::JS::Operators -- Implemented &sqrt, &sign, &abs, &log, &log10,
r5900 | iblech++ | &int, and &rand. Tests mostly pass:
r5900 | iblech++ | m19s28.vlinux.de/iblech/stuff/not_p...tins/math/
wilx Heh, interesting, rounding errors in action: 20:15
# Expected: '1.6094379124341005'
# Got: '1.6094379124341003'
log.t.html
svnbot6 r5901 | Aankhen++ | * added more fields to list of request headers in HTTP::Request::CGI. 20:26
Aankhen`` pokes wolverian. 20:27
svnbot6 r5902 | Aankhen++ | HTTP::Request::CGI: 20:36
r5902 | Aankhen++ | * added $.remote_host and $.remote_addr.
r5902 | Aankhen++ | * updated documentation.
Aankhen`` Has `read` been implemented? 20:37
putter iblech: ping? 20:41
Aankhen`` Anyone up for a large porting job? =) 20:45
Well, it's a large distribution, I would only need part of it.
geoffb A) Porting what to what, B) does this involve large sums of money? :-) 20:48
Aankhen`` The answer to B) is no. 20:49
I can tell you A) in private so that I don't hurt people's sensibilities.
geoffb ok . . .
Aankhen`` Hm.
Actually, nevermind.
It's not needed after all.
geoffb :-)
Aankhen`` Though if you're up for it, I'd appreciate ports of Compress::Zlib, Compress::Bzip2 and MIME::QuotedPrint... =) 20:50
geoffb Not that I don't agree, but is `use perl5:...` not acceptable? 20:51
Aankhen`` Considering the fact that I am unable to link Perl 5 into Pugs... no. 20:52
geoffb that's odd . . . oh, are you on Win32?
Aankhen`` Yep.
geoffb sigh 20:53
I keep trying to make peace with Win32, and it keeps giving me reasons to hate it.
svnbot6 r5903 | iblech++ | * t/builtins/strings/string_interpolation.t: unEVAL.
r5903 | iblech++ | * PIL2JS:
r5903 | iblech++ | * Prelude::JS::Str -- Made the third argument to &substr optional and made
r5903 | iblech++ | &substr work with the third argument being negative.
r5903 | iblech++ | * Prelude::JS::Str -- Implemented &lcfirst, &ucfirst, &chars (of course,
r5903 | iblech++ | &chars is only a wrapper for JavaScript's "string".length, so it probably
r5903 | iblech++ | doesn't return the num of chars, but the num of bytes/whatever), &index,
r5903 | iblech++ | &rindex, &chomp.
Aankhen`` You shouldn't even try.
svnbot6 r5903 | iblech++ | * Prelude::JS::Operators -- Implemented &chr and &ord.
r5903 | iblech++ | * t/builtins/strings/ tests mostly pass:
r5903 | iblech++ | m19s28.vlinux.de/iblech/stuff/not_p...s/strings/
geoffb Well, I'm trying to make room in my psyche for other things I need to feel strongly about. 20:54
s/need/would rather/
Aankhen`` Do you use Windows?
geoffb Occasionally. Long time ago I was a windows sysadmin. I came by my hatred honestly. 20:55
Aankhen`` Then ignore Windows.
geoffb Oh, but see, everyone else is using it, so consequently any time I could normally just recommend a simple solution to someone's problem, I always have to deal with the "But I'm using Windows, so that won't work" issue. 20:56
Aankhen`` Ah, you have to interact with it, indirectly. 20:57
geoffb When I started writing my Perl/OpenGL articles, the single largest feedback group was "It doesn't work on Win32!"
And remaining a publishable author is predicated on lack of negative feedback . . . .
nodnod
Aankhen`` Welp, I use Windows everyday. 20:58
I content myself with the knowledge that by this time next year, I won't have to.
(then I can just use it for games)
geoffb Oh? What's changing?
Aankhen`` My family is moving abroad.
Hence a new computer.
geoffb one of your choice, I take it? 20:59
Aankhen`` Oboyoboyoboy I drool everytime I think of hardware-based virtualization.
geoffb heh
Aankhen`` Yeah, we'll each get individual laptops, plus I get a desktop.
Gonna work out pretty expensive.
geoffb Aankhen`` is clearly not short of cash . . .
nod.
Aankhen`` Believe it or not, this comes under "necessities".
geoffb ?!? Family of computer consultants? 21:00
Aankhen`` Each member of my family -- that is my mother, my sister and me -- has something or the other that is important and has to be done on the computer.
Right now, we have one desktop, which basically belongs to all of us, and one laptop, which was gifted to my mother.
The way it works is, my mother's not particularly keen on using the laptop for now, so my sister uses the laptop and I use the desktop. 21:01
However, my mother has work coming up which will require the laptop extensively.
So once again, the flames are going to fly.
geoffb wheee
Aankhen`` Aaanyway. 21:02
I want it to be 2007 right now.
obra why?
Aankhen`` Four-core processors from AMD with hardware-based virtualization. =) 21:03
I just need two quad-core processors, with three monitors and six graphics cards, and appropriate storage & memory. Is that so much to ask? 21:05
geoffb three cards and six monitors?
or are you going to SLI each pair?
svnbot6 r5904 | iblech++ | * Usual svn props.
r5904 | iblech++ | * PIL2JS:
r5904 | iblech++ | * Prelude::JS::Operators -- Implemented &hex, all tests pass:
r5904 | iblech++ | m19s28.vlinux.de/iblech/stuff/not_p...hex.t.html
r5904 | iblech++ | * Now most tests in t/builtins/ pass, except io/, of course. Pondering using
r5904 | iblech++ | JSAN's HTTP.Request or so to fake simple things like file-open, slurp, etc
r5904 | iblech++ | (slurp "foo" would GET cur_server/cur_path/foo then)...
Aankhen`` SLI.
Though actually, I could make do with "just" four. :-P 21:06
Assuming those processors have Pacifica, I could run Gentoo, Windows and OS X simultaneously. One monitor for each, one graphics card for each (two for Windows since it has all the games).
geoffb Dare to dream, Aankhen``, dare to dream. 21:07
Aankhen`` Oh, I do. :-D
geoffb Hmmm, very few Debian testing updates hitting my boxen recently -- I'm guessing somethings getting backed up, and I'm going to come back from OSCON to find that every box needs to install 200 updated packages. 21:09
Aankhen`` Heh.
autrijus I can't believe I spent 3 hours on journaling :-/ 21:10
Aankhen`` Is it up? 21:11
Oho!
autrijus it's 5am now... really need to sleep. journal finally up, yes.
Aankhen`` o_O
Sleep well.
geoffb good night, autrijus, sleep well
autrijus :)
Aankhen`` Mmm, I need to sleep soon too.
21:15 Limbic_Region_ is now known as Limbic_Region
brentdax I am somehow unsurprised that Autrijus is the most active committer in open source. 21:19
geoffb Where did you see this stat, brentdax
?
Aankhen`` Whoa. 21:20
brentdax CIA.
Aankhen`` The cleanest places on Earth are IC manufacturing facilities.
brentdax cia.navi.cx/
Aankhen`` s/manufacturing facilities/fabrication plants/ 21:21
brentdax Hey, every other geek environment is littered with soda cans and Twinkie wrappers, we have to make up for that somewhere.
Aankhen`` Apparently an IC fabrication plant is 10,000 times cleaner than a hospital operating room.
G'night. 21:28
Juerd Greetings from What The Hack, in Liempde, the Netherlands. 21:29
brentdax Greetings, Juerd. 21:30
Juerd It's really great here; lots of really interesting conversations and presentations
brentdax (From a house in Southern California with broken air conditioning. Argh.)
Juerd We're experiencing heavy showers outside 21:31
I'm in a big tent.
It's more a hall than a tent.
brentdax Hmm...if I rename that method with the weird problem, it goes away. Does "resolve" have any special meaning to Pugs?
(or rather, if I rename the method it's calling so they don't share a name.) 21:32
Juerd There is one bit of sad news, though, everyone here I've spoken with about Perl, thinks that Perl is practically dead.
That's not a good start for Perl 6, the name staying the same.
Perl 6 probably has a greater chance of survival with a better name. "Perl" is unfortunately tainted.
Limbic_Region . o O ( Topaz2 ) 21:34
brentdax Unlike the language whose acronym originally expanded to "Pretty Home Pages"?
Limbic_Region Juerd - you at the "What the hack" conference?
nevermind, me reads back a bit and sees that Juerd already said this 21:35
Juerd Limbic_Region: Yes. 21:36
brentdax I think people are thinking of Perl 6 as vaporware, if they know about it at all. A release will change their mind.
Juerd brentdax: No, I've met only a few PHP enthousiasts
brentdax: We're clueful people here.
brentdax Heh.
Limbic_Region wonders if Juerd is in a hammock too - www.cnn.com/2005/WORLD/europe/07/29...index.html
Juerd brentdax: No, people are actually fairly confident that it will eventually happen and be a good language. It's just that Perl now has a VERY bad reputation amongst the hackers here, which will have its effect on Perl 6 too. 21:37
geoffb OOC, what has a good rep?
Wow, pugs REALLY doesn't like heredocs in a big program -- it goes to 100% CPU and seems to hang 21:40
Juerd Ruby and Python have high karma
geoffb Heh.
Juerd Limbic_Region: No, I haven't found an available hammock yet :) 21:41
brentdax Ruby, at least, deserves its karma.
Juerd Python unfortunately has higher
geoffb sigh, no accounting for taste 21:42
Juerd But it really did surprise me to learn how low exactly people think of Perl 21:43
And contrary to what I'm used to on IRC, people actually have good argumentation 21:44
geoffb So what are these good arguments?
Juerd We all know they're right, and that's why Perl 6 is being bilt. But still, the name "Perl" carries bad karma :(
s/bild/build/
Khisanth we do? 21:45
Juerd Yes. We're not building Perl 6 because 5 suffices.
Perl 5 does not suffice
geoffb: Let me try and formulate that later, when I get back from France in a week
geoffb: It's too much for now, over a very slow link
Khisanth agrees that there is always room for improvement
geoffb Juerd, fair enough 21:46
I'll hold you to that. :-)
Juerd Yes, but when there is already a language that does things better, there's not just ROOM for improvement, you're' expected to actually improve. And quickly too.
This is the main problem, as Perl 5 does not allow for quick improvements to keep up with the other dynamic languages 21:47
But still, the most often heard argument is that Perl is ugly and requires too highly educated people, which is very (too) expensive for many companies. 21:48
brentdax Switch.pm would not be such a Herculean effort in Perl 6. (I mean, besides the fact that it'll be built in anyway.)
Khisanth that last argument sounds very silly without expanding 21:49
geoffb Ugly -- I know people who feel that way, I don't agree, but whatever. Requiring too highly educated people so its expensive? Well that's just being completely clueless about the value of smart people.
Juerd I do agree that Perl is ugly. Ugly in the way that lots of interpunction is needed without it actually making it more elegant
brentdax The value of a hacker is...the square, maybe the cube, of their skill level.
Khisanth ugly or not is an opinion :)
Juerd Do note that people here do use Perl a lot. 21:50
brentdax Whereas the price of a hacker rises more or less linearly with skill.
geoffb I cannot help that companies think that it is cheaper to higher unskilled labor, but I find it as silly as Paris fashion.
Juerd They just can't use th^H^Hit at their jobs
Except the lucky few who do system administration
Perl isn't used much for building applications, or even web based applications 21:51
geoffb The last two web-related jobs I had were done in Perl -- HTML::Mason, as a matter of fact; Amazon runs on Mason, etc. What more do these people want?
Juerd How many NEW application projects are built with Perl? 21:52
And how many in Python and PHP?
The sad fact is that "we"'re losing the battle.
geoffb Honestly, I dunno -- but certainly enough to keep CPAN bulging at the seams. And using PHP and Python because they are the new things is like using Java a few years ago. 21:53
Juerd Perl 6 can be a great comeback, but people I spoke with fear that it will be seen as a new version of the language that has a bad name
obra See also rails.
geoffb fair enough
integral the BBC use it, their single-signon web app was new
Juerd While when "marketed" as a *new* language, it would have a great change^H^Hce of surviving
Well, CPAN is nice, but crap.
brentdax Crap how? 21:54
obra evalbot, start a flame war
geoffb ?eval '*dons flame-retardant outfit* FLAME ON!' 21:56
evalbot6 '*dons flame-retardant outfit* FLAME ON!'
Juerd brentdax: The modules themselves are 21:58
breOf course, I'm now generalizing
People don't know well enough that CPAN is an open system
And many people expect quality, because it all looks so official, documentation being available in a central place and all. 21:59
Some even wonder why the Perl modules are so diverse in API and see a relation to PHP in that respect.
brentdax So people don't look in the FAQ, therefore CPAN sucks.
geoffb <clkao>pugs sucks, it misplaces error messages for loops with parsefail contents</clkao> 22:00
Juerd I've tried very hard to explain how it works, and was able to get the message across, but I am quite pessimistic because I have to admit that in general, people think badly about perl :(
geoffb brentdax, sadly, people don't read
Juerd brentdax: Peolp edon't read. That's something we, as the Perl community, have to get used to.
We're no longer in the world with only RTFM'ing hackers. Only references are still used much. :( 22:01
Right here, at the WTH infodesk, we experience the same thing
integral hmm, does perl need to move out of that world though? Doesn't it work very well where it is, and Java and Python very well where they are in the world of non-manual readers?
Juerd We have many signs with lots of different information. There are at least 40 diifferent pages stuck to the walls. 22:02
People no longer read them, because it's way too much, and asking is easier.
In other words: our attempt to answer FAQs on paper, really did fail.
integral: It works very well, but not if we want it to compete (unquote) with the other languages
integral why do we want it to compete? 22:03
Does any of the "quality" of CPAN comes from the fact that there's not a wider audience using it?
Juerd Several reasons
I'm not sure what quality you're referring to 22:06
CPAN has some really brilliant and great modules, of course.
But the majority is not much better than the available PHP copy-and-paste repositories :) 22:07
And stuff like magic ARGV using 2-arg open is really hyped as very insecure here 22:09
And well, of course they're right
brentdax wages a protracted battle against apathy. 23:45
ayrnieu Don't fight it. Just make sure you're apathetic about the right things. 23:50
brentdax My SoC project is not one of the right things, unfortunately.