pugs.blogs.com | pugscode.org | pugs.kwiki.org | paste: sial.org/pbot/perl6 | <stevan> Moose... it's the new Camel ":P | .pmc == PPI source filters!
Set by Alias_ on 16 March 2006.
00:21 ajs_home joined 00:22 neoesque joined 00:24 hikozaemon joined 00:34 coumbes joined 00:48 unobe_ joined 00:53 SubStack joined 01:18 unobe__ joined, unobe__ is now known as unobe 01:19 jserv-- joined 01:28 rdice joined 01:34 weinig is now known as weinig|away 01:40 sri_ joined 02:03 justatheory joined 02:43 SamB joined 02:58 viirya joined 03:13 avar joined
svnbot6 r10423 | scw++ | Pugs::Grammar::MiniPerl6 test updated for cooperation with ratchet later 03:18
03:23 avar joined 03:57 macroron joined 04:05 Kattana joined 04:40 takesako is now known as audreytak 04:43 audreytak is now known as takesako 04:45 penk joined 05:31 jshriver joined
jshriver hello 05:31
What's new in perl 6, or what's diff in it an the 5 series
05:33 xern joined 05:53 iblechbot joined 05:54 FurnaceBoy left 05:55 Quell joined 06:41 xern joined 06:42 iblechbot joined 07:10 avarab joined 07:11 SM_ax joined 07:13 neoesque joined 07:17 Aankhen`` joined 07:30 drrho joined 07:41 ko1_away is now known as ko1 07:48 baest_ joined 07:58 baest_ joined 07:59 baest_ is now known as baest 08:04 marmic joined 08:05 ko1 is now known as ko1_away 08:08 baest joined 08:20 SubStack joined 08:24 SM_ax joined 08:31 elmex joined 08:47 baest joined
theorbtwo pingsabout. 09:38
Bah, I hate it when I'm not sure if something should be here or on the ML. 09:39
Conrad says:
The basic idea is this: since the pugs svn tree is mirrored at perl.org, we
could have a subdirectory (for example "Perl_6_ Wiki") in the docs branch of
the source tree that would have a wiki-like hierarchy of html pages (or pod
pages for generating html) below it. The downside is that it would take
somewhat more effort to make changes and additions.
</blockquote>
I tend to think that a wiki Done Right could be a web front-end to a SVN repository -- that is, each page is a file in the svn repo, and modifying the page via the web form is just another way to make an svn checkin. 09:41
integral imho what I'd really like is a auto-updating job running on feather that produces an index of .../doc, and renders the pod to HTML so people can read it on the web confortably.
theorbtwo (And we allow people to access the svn repo directly, without going through the web fe if they want.)
I wonder if openfoundry could handle the load of checking if a page is up to date very-very-often.
integral umm, the svnbot already does checking :) 09:42
and the auto-reload-the-evalbot-thing too
theorbtwo Hm, true enough.
...so apparently it can.
integral we just need to capture someone with tuits 09:43
theorbtwo Conrad has a major point that if modifying the docs isn't easy, it won't get done all that much. That's rather the point of a wiki.
Anybody know if openfoundry allows giving wider rights on a subdir then is normal for a repository? 09:44
integral viewing needs to be easy too. the current: view raw source in webbrowser direct from svn, or check out and use perldoc isn't easy enough
theorbtwo I quite agree.
integral always itches for the red pen when reading stuff 09:45
theorbtwo Speaking of, I'd love the editing interface to (optionally?) put every paragraph in a different textarea.
When editing longish wikis, it's generally horribly difficult to find the bit of text you wanted to change.
integral I believe that's a mediawiki feature 09:46
theorbtwo Oh!
Wikipedia didn't do it the last time I looked.
integral it puts wee '[edit]' tags next to each heading on the page, so you just click that to edit a specific section
theorbtwo The sections can still be quite large, though. 09:47
integral by 'section', I also mean subsections, subsubsections, subsubsubsections, etc :)
integral & off to do some mathing 09:48
theorbtwo Have fun. 09:49
Hm, there's an openfoundry wiki available to us. 09:50
integral we have a pugs kwiki: pugs.kwiki.org (see the topic!) too 09:51
theorbtwo Oh, I forgot about that one.
I tend to think the problem is more too many wikis and not enough documentation of any of them then it is not the right wiki. 09:52
(Not that having the one true wiki run a good wikisystem wouldn't be good, but it's a minor point compared to there being a one true wiki system.)
Hm, the openfoundry frontend doesn't seem to allow access control on a subdirectory level. 09:56
Hmmm... given audrey's "be loose with commit bits" policy, perhaps we should just make a page on the web fe that allows you to "invite yourself", and then fills out the form on openfoundry. 09:57
That'd give anybody who asks on that page full commit rights, though. 09:58
Or, in the alternative, make creating a user on the wiki site create a user on the openfoundry side with the same userid, but manage the password ourself (by setting the user's email to one a script handles), and change their email address and send a "forgot password" request when they should graduate to a real user. 10:01
That sounds like a lot of work, though.
In the next alternative, mess with the svn repo in ways that can't be done via openfoundry's frontend.
10:17 szbalint joined 10:38 elmex joined 10:39 elmex_ joined 11:01 neoesque joined 11:46 weinig|away is now known as weinig 11:51 Limbic_Region joined 11:56 bjoern_ joined 12:21 jsiracusa joined 12:24 lisppaste3 joined 12:40 Gothmog_ joined 12:47 turrepurre joined 13:05 hexmode joined 13:25 cnhackTNT joined 13:26 cnhackTNT left 13:38 Limbic_Region joined 13:40 macroron joined 13:42 jserv-- joined 13:45 neoesque joined 13:55 Alias_ joined
Alias_ seen audreyt? 13:55
jabbot Alias_: audreyt was seen 9 hours 15 minutes 49 seconds ago
Alias_ seen TimToady?
jabbot Alias_: TimToady was seen 7 days 7 hours 25 minutes 58 seconds ago
14:10 bjoern_ joined 14:21 justatheory joined 14:39 cdpruden joined 15:03 FurnaceBoy joined 15:07 OldeFortran joined 15:33 weinig is now known as weinig|, weinig| is now known as weinig|away
particle_ ?eval say { 'a' => 'b', 'c' => 2 }; 15:34
evalbot_10383 OUTPUT[ab c2 ] Bool::True
15:51 rdice joined 15:53 mj41_ joined
Limbic_Region Juerd ping 16:32
anyone a mingw user?
Alias_ yes
Limbic_Region Alias_ do you use it to build Parrot, and if yes, do you mind me /msg'ing you some questions 16:33
Alias_ nope
Limbic_Region oh, so no Parrot for you or you use Cygwin or something else?
particle_ l_r: build failing?
Limbic_Region particle_ - not even getting started 16:34
particle_ ah.
Limbic_Region basically the backslashes in the path are not getting intepreted as path delimiters
so it is looking for Perl at C:perlbinperl.exe
Alias_ No parrot for me 16:35
particle_ can you wander over to #parrot on irc.perl.org?
Limbic_Region no - that's what the Juerd ping was for
particle_ i can answer questions here, too
Alias_ We use mingw in the Vanilla Perl builds
Limbic_Region I can only IRC from work via CGI::IRC
he only has it currently set up for #perl6
particle_ oic
Limbic_Region is guessing it is something obvious 16:36
for instance, I am unable to create C:\Windows\System\<conf>.ini because of security lockdown on the box
but now that I got Pugs building again I wanted to link with parrot :-( 16:37
particle_ during configure, parrot takes platform hints.
on your platform, that's /config/init/hints/mswin32.pm
check the elsif ($is_mingw) section, especially the slash => '\\' 16:38
perhaps changing that will help?
Limbic_Region checking
particle_ after configure is complete, there's lib/Parrot/Config.pm where you can check the settings, too (but you can't change them there.)
Limbic_Region but since it executes under a shell - I am thinking it wants \ to become / not \\ 16:39
particle_ so change the hint to '/'
re-config
re-build
Limbic_Region trying now
Juerd Limbic_Region: pong 16:41
particle_ anyone else see matt sergeant in the new york times today?
www.nytimes.com/2006/05/21/business...ref=slogin
Limbic_Region Juerd - do you think it would be possible to set up feather's CGI::IRC to allow #perl6 on freenode AND #parrot on irc.perl.org? 16:42
particle_ no go, but I am still playing
Juerd I'll have a look if this system can handle that easily 16:43
If not, I can just copy the tree over
Limbic_Region having set up CGI::IRC in the past - it means that you will be able to go to #perl6 or #parrot on freenode or irc.perl.org
they don't (or at least not last time I used it) have a way to lock in a channel with a network 16:44
particle_ - I am not sure mswin32.pm is getting used - the STDERR debugging statements I put in aren't coming out 16:45
msys.pm makes more sense - looking there 16:46
particle_ you can use perl Configure.pl --verbose-step=hints to tell you your platform 16:47
Limbic_Region was just doing that now
parrot has changed a lot in the 1.5 years since I used it last 16:48
particle_ change is good :)
Juerd Limbic_Region: Done 16:52
Limbic_Region thanks Juerd 16:53
particle_ headed over there now as chaning the delimiter in the hints didn't change the path of Perl in the Makefile
16:58 chris2 joined 17:05 Alias_ left 17:23 ruoso joined 17:29 weinig|away is now known as weinig 17:35 Juerd2 joined 17:41 juerd2 joined
Juerd Wow, CGI:IRC is incredibly stupidly written. 17:42
17:43 juerd2 joined
Limbic_Region Juerd - that was my impression too from several years ago 17:44
I would have hoped that it would have been improved since then
Juerd Well, it's a non-modular mess.
Logic is spread all over the place
So adding a feature like this isn't as simple as it should be.
17:46 larsen joined 17:47 juerd2 joined
Juerd Okay, done now. 17:47
Hard coded it :) 17:48
Limbic_Region I see - thanks 17:50
17:53 justatheory joined, weinig is now known as weinig|coffee 18:38 ruoso joined 18:46 FurnaceBoy_ joined 19:15 avar joined 19:17 ramzi joined, rdice joined 19:29 trym joined 19:41 mj41_ joined 19:48 sri__ joined 19:53 weinig|coffee is now known as weinig 20:12 fglock joined
fglock Juerd: if you donōæ½xB4t mind hacking cgi-irc, would you add "charset=utf-8" to the headers? "Content-type: text/html; charset=utf-8" I think 20:16
Juerd No, that's a promise that it can't keep. 20:21
Solve the IRC problem (change the protocol to support communicating character encodings), wait till the major servers and clients use the improvements, and then come back to me :) 20:23
fglock Juerd: maybe add "AddDefaultCharset UTF-8" to apache?
Juerd fglock: No. That's nonsense.
fglock Juerd: ok
szbalint irssi + recode support is great
Juerd Encodings aren't something you just add and hope things work out well.
They require conscious decisions.
Like: incoming IRC messages have to be decoded, and later, encoded to utf-8, if you want utf-8 output 20:24
Otherwise you get invalid characters.
But you cannot decode things properly.
IRC hasn't standardized on UTF-8, either.
Many people use latin-1, latin-15, or even non-latin encodings. 20:25
Promising to the browser that the data will be utf-8 is bad, if you already know you can't keep the promise.
I'm all for utf-8-ification. But not blindly. It has to be done with care.
And IRC prohibits any sane upgrade path.
Sorry.
szbalint Personally I've been using ISO-8859-2 for years. But I'm on four networks. 2 uses UTF-8, one is mixed and one is old ISO-8859-2. 20:26
What Juerd says is very true, unfortunately.
Juerd By the way, I wrote a Perl unicode tutorial yesterday, and merged reviewer's comments today.
juerd.nl/perlunitut.html
If you prefer pod, then use .pod 20:27
20:42 qu1j0t3 joined
fglock I wonder if there could be a "use utf8" checkbox during login, so it could be optional - does this make sense? 20:43
20:43 FurnaceBoy joined
ruoso Juerd++ great doc 20:48
Juerd Thank you
fglock: If you want to do this as the end user, just configure your browser? 20:49
bjoern_ it would be good to have something similar for module authors
Juerd fglock: Look for View > Character set > UTF-8
bjoern_: Most module authors don't have to do anything special.
bjoern_ like, what should module authors put into the docs relative to unicode
I'm not sure I agree :)
Juerd Only if your module communicates with non-perl, it has to handle this stuff.
With non-same-process I mean
And then, exactly the same rules apply as described in this tutorial 20:50
ruoso Juerd, is there a way to disable auto-upgrade?
Juerd Decode all incoming data
Encode all outgoing data
20:50 b_jonas joined
Juerd ruoso: No, and some things depend on this, so you shouldn't. But audreyt wrote a module to help debug, encoding::warnings. 20:51
ruoso Juerd, nice...
Juerd bjoern_: The trouble is that modules are supposed to be written for generic application.
ruoso auto-upgrade causes hard-to-track bugs in some cases...
Juerd bjoern_: Thus, a string should be just a string. However, if you start encoding strings automatically, you loose compatibility with binary strings, and can no longer handle stuff like images, music, movies, etcetera, properly. 20:52
ruoso audreyt++
Juerd ruoso: If you are disciplined at decoding incoming information, upgrading is no longer a problem. 20:53
A consistent coding style helps recognizing such problems.
Well written modular code, where input, processing, and output are properly separated, usually doesn't suffer from implicit upgrading problems, if it decodes and encodes explicitly. 20:54
20:54 fglock joined
ruoso Juerd, the problem is when someone with bad style uses a module that expect a good-style practice 20:54
Juerd ruoso: Yes, but the module can't fix it. 20:55
ruoso Juerd, use encoding::warnings 'FATAL' is nice...
fglock Juerd: when I change to utf8 it logs me out - but in the end it works
Juerd fglock: That's because the browser reloads the page :) 20:56
Or, well, tries to.
ruoso it must, because the request was in one enconding... and will start to be in another...
the connection must be reset... 20:57
so the new encoding is used in the transfer
Juerd It could try to be intelligent, and check if any bytes have changed in between encoding differences :)
ruoso heh
Juerd If the request contained pure ASCII, and it it did in this case, the browser could be made smart enough to realise it doesn't have to re-request if both the old and the new encoding are ASCII compatible.
s.it it.it is.
ruoso the problem is that it doesn't know what's coming... 20:58
but, thinking better..
it should be able to split that
no matter the connection encoding...
the browser could encode/decode....
hmm... actually not for utf8 20:59
forget what I said
fglock back to class &
20:59 fglock left 21:01 kanru joined 21:28 DaGo joined 21:33 justatheory joined 21:58 Limbic_Region joined 22:04 bpalmer joined 22:13 jsiracusa joined 22:22 rdice joined 22:26 apple-gunkies joined 22:30 ruoso_ joined 22:31 ruoso_ is now known as ruoso 22:39 SubStack joined 22:41 apple-gunkies joined 22:54 froh-doh joined 22:57 gunya joined
clkao in perl6 rules, when you have inline code { return "foo" }, doesi t became whatever is being captured? 23:47

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!