Camelia, the Perl 6 bug

IRC log for #metacpan, 2011-10-28

| Channels | #metacpan index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:05 castaway joined #metacpan
01:35 klapperl joined #metacpan
01:42 castaway joined #metacpan
02:43 jsut joined #metacpan
02:46 theory joined #metacpan
03:24 am0c joined #metacpan
03:26 metacpan joined #metacpan
03:26 metacpan [metacpan-web] rwstauner pushed 3 new commits to master: http://git.io/e8txiQ
03:26 metacpan [metacpan-web/master] Highlight fully qualified package names - Randy Stauner
03:26 metacpan [metacpan-web/master] Syntax highlight here docs - Randy Stauner
03:26 metacpan [metacpan-web/master] Add 'feature' "keywords" - Randy Stauner
03:26 metacpan left #metacpan
03:26 dipsy [ Comparing de04d09...a3817af for CPAN-API's metacpan-web - GitHub ]
03:26 rwstauner i'm also sending those and our previous perl brush changes upstream to alexgorbatchev
03:53 oalders and puled
03:53 oalders rwstauner++
03:53 oalders er, pulled
03:53 rwstauner i think i have more coming :-)
03:53 oalders you're a machine :)
03:54 rwstauner :-)
04:13 metacpan joined #metacpan
04:13 metacpan [metacpan-web] rwstauner pushed 1 new commit to master: http://git.io/tyXlFQ
04:13 metacpan [metacpan-web/master] Highlight fat commas, quote-like ops, and \&funcs - Randy Stauner
04:13 metacpan left #metacpan
04:13 dipsy [ Commit 7c91db091c0fb6ee170b1c7192bb75e20141b652 to CPAN-API/metacpan-web - GitHub ]
04:13 rwstauner there's still more that could be done (only perl can parse perl, after all), but that's a few nice additions
04:14 rwstauner and enough of me getting side tracked on that
04:14 oalders nice work.  :) pulled!
04:19 rwstauner holy cow... just submitted a 19-commit pull req to send all our changes upstream: https://github.com/alexgorbatc​hev/SyntaxHighlighter/pull/52
04:19 dipsy [ #52: Improvements to the perl brush by rwstauner for alexgorbatchev/SyntaxHighlighter - Pull Request - GitHub ]
04:19 oalders that'll keep him busy :)
06:11 omega can I link to another authors page from POD somehow?
06:15 dpetrov joined #metacpan
06:45 alnewkirk joined #metacpan
07:19 dpetrov https://metacpan.org/search?q=​DBIx%3A%3AClass%3A%3AResultSet
07:19 dipsy [ Search for "DBIx::Class::ResultSet" - metacpan.org ]
07:20 dpetrov shouldn't the Dbix::Class::ResultSet
07:20 dpetrov be on a first place?
07:37 Ranguard dpetrov: https://github.com/CPAN-AP​I/metacpan-web/issues/207 and https://github.com/CPAN-AP​I/metacpan-web/issues/266
07:37 dipsy [ #207: Searching for X doesn't find X first (where X = DBIx::Class::ResultSet) - Issues - CPAN-API/metacpan-web - GitHub ] [ #266: Net::Amazon::S3 not found - Issues - CPAN-API/metacpan-web - GitHub ]
07:40 dpetrov Ranguard: sorry didn't check the issue list
07:40 Ranguard np :)
07:42 dpetrov also... does it make sense to show ++ modules higher?
07:42 Ranguard I think the idea was to have that eventually
07:43 dpetrov okey ;)
07:43 Ranguard but ++ is against a distribution, not a module, so it's not actually quite so obvious if that is correct
07:43 Ranguard btw, I know they're always looking for people to help out - if you have time :)
07:44 dpetrov sure
07:45 dpetrov if only I knew how can I help :)
07:45 Ranguard https://github.com/CPAN-API/metacpan-web/issu​es?labels=Volunteer+needed&sort=created&a​mp;direction=desc&state=open&page=1 - for a start :)
07:45 dipsy [ Issues - CPAN-API/metacpan-web - GitHub ]
07:46 dpetrov quite a list :)
07:46 Ranguard yea, and that's just the non-core stuff :)
07:48 dpetrov :)
07:49 dpetrov openid/twitter authentication shouldn't be hard at all
07:51 dpetrov oh twitter exists already
07:58 clintongormley joined #metacpan
07:59 jsut_ joined #metacpan
08:26 dpetrov joined #metacpan
08:32 clintongormley Ranguard: acme: re your comments about Net::Amazon::S3 etc, is this different from how it was? or always been like this?
08:32 clintongormley i ask because mo has a new branch called 'precision' which should give much better results
08:40 Ranguard clintongormley: my comment was just to put more info into the ticket to help someone testing, acme's comment was that it has changed since he reported it, probably best to see what 'precision' does and add these cases to the tests
08:42 clintongormley yep - mo said that he was seeing much better results with that branch
08:49 clintongormley mo: when are you thinking about putting precision live?
08:56 clintongormley Ranguard: just tried out the precision branch locally, and the results for Net::Amazon, Net::Amazon::S3 and DBIx::Class::ResultSet are spot on
08:57 Ranguard sweet :)
09:21 * Ranguard updates ticket so people are aware
09:32 mo I guess we can merge that branch
09:42 daxim joined #metacpan
10:21 am0c joined #metacpan
10:43 metacpan joined #metacpan
10:43 metacpan [metacpan-web] monken pushed 4 new commits to master: http://git.io/FD1J4g
10:43 metacpan [metacpan-web/master] improve precision - Moritz Onken
10:43 metacpan [metacpan-web/master] cleanup - Moritz Onken
10:43 metacpan [metacpan-web/master] give scripts a negative boost, cleans up seearch for 'dbix' - Moritz Onken
10:43 metacpan left #metacpan
10:43 dipsy [ Comparing 7c91db0...2d7f958 for CPAN-API's metacpan-web - GitHub ]
10:48 metacpan joined #metacpan
10:48 metacpan [metacpan-web] monken pushed 1 new commit to master: http://git.io/2tkivA
10:48 metacpan [metacpan-web/master] default to 'and' operator for author search, prevents author results to show up for e.g. 'dbix class' - Moritz Onken
10:48 metacpan left #metacpan
10:48 dipsy [ Commit d28060bb2fdd284d25b573fb24986f433515bfe2 to CPAN-API/metacpan-web - GitHub ]
10:51 dpetrov mo++
10:51 Ranguard mo++
10:52 mo please let me know if some results got worse
11:01 gabimuc|work joined #metacpan
11:23 avar My biggest problem with mcpan search results is that I tend to type in lower case and space seperated random sub-words of package names
11:23 avar e.g.: https://metacpan.org/search?q=apache+cookie http://search.cpan.org/search?q​uery=apache+cookie&mode=all
11:23 dipsy [ Search for "apache cookie" - metacpan.org ] [ The CPAN Search Site - search.cpan.org ]
11:23 avar sco finds "apache::cookie" as the first hit
11:24 avar But this is a more relevant example: https://metacpan.org/search?q=poe+irc http://search.cpan.org/searc​h?query=poe+irc&mode=all
11:24 dipsy [ Search for "poe irc" - metacpan.org ] [ The CPAN Search Site - search.cpan.org ]
11:24 avar POE::Component::IRC is the most likely think you want when you search for poe irc I would think
11:25 mo avar: could you open a ticket with those examples?
11:25 avar sure
11:26 mo thanks!
11:30 avar https://github.com/CPAN-AP​I/metacpan-web/issues/370
11:30 dipsy [ #370: Perform better on sub-words of full module name like search.cpan.org - Issues - CPAN-API/metacpan-web - GitHub ]
11:30 avar Have you looked into trying to get download data from some popular cpan mirror and using it for the scoring?
11:31 avar Or actually, does a field in the document you're searching through have the number of modules that have this module as a dependency? I think that would find both of these examples I'm looking for if you came up with the right custom_score query
11:31 mo working on that :)
11:32 mo actually a pagerank implementation for cpan modules
11:34 avar nice
12:33 oalders omega: i think you'd just link to the full author url L<https://metacpan.org/author/PAUSEID>
13:13 omega oalders: ok, thanks
13:25 virtualsue joined #metacpan
13:41 rwstauner joined #metacpan
13:49 rwstauner has the autocomplete gotten faster?
13:50 rwstauner i used to have to wait for it, right now it's almost beating me
13:50 rwstauner maybe that has to do with the new precision stuff?
13:50 rwstauner mo++
14:11 vti joined #metacpan
15:20 clintongormley1 joined #metacpan
15:24 clintongormley joined #metacpan
15:27 am0c joined #metacpan
15:33 theory joined #metacpan
15:39 mstratman joined #metacpan
15:42 clintongormley1 joined #metacpan
15:45 dpetrov https://metacpan.org/module/Catalyst
15:45 dipsy [ Catalyst - The Elegant MVC Web Application Framework - metacpan.org ]
15:45 dpetrov why there is no list with the available methods?
15:48 rwstauner dpetrov: not sure what you mean
15:50 doy dpetrov: because you clicked on "toggle table of contents" maybe?
15:51 dpetrov :)
15:51 dpetrov rwstauner: what doy said
15:51 dpetrov but I didn't clicked on it
15:51 rwstauner doy++
15:51 dpetrov (or at least I didn't mention)
15:51 dpetrov doy++ # thanks
15:51 rwstauner always go to doy, first :-)
15:51 daxim that's a crappy interface design if I've ever seen one
15:52 daxim (and I'm nethack addict)
15:53 clintongormley joined #metacpan
15:54 clintongormley joined #metacpan
15:59 clintongormley1 joined #metacpan
16:00 clintongormley2 joined #metacpan
16:07 clintongormley joined #metacpan
16:13 clintongormley1 joined #metacpan
16:29 clintongormley joined #metacpan
16:32 rwstauner are we using Template::Toolkit beneath Template::Alloy?
16:33 joel er, isn't Template::Alloy a reimplementation of the TT2 syntax?
16:33 rwstauner oh is it?
16:34 rwstauner i thought Alloy was just a wrapper for the others
16:34 joel well, it's more than that, it implements various Template syntaxes
16:34 joel but it doesn't wrap them, it reimplements them in terms of a common system (alloy)
16:35 joel https://metacpan.org/module/Template::Alloy::TT # explains how they differ
16:35 dipsy [ Template::Alloy::TT - Template::Toolkit role - metacpan.org ]
16:35 joel "Alloy uses the same base template syntax and configuration items as TT2, but the internals of Alloy were written from scratch"
16:36 rwstauner i just copied that same text
16:36 rwstauner thank you very much
16:38 joel np
16:41 gabimuc|work joined #metacpan
16:48 smpb joined #metacpan
16:50 jsut joined #metacpan
16:55 clintongormley joined #metacpan
17:02 rwstauner for those of you not following along (which excludes joel), just spoke with t0m in #catalyst about encodings...
17:03 rwstauner If I add $data =  Encode::decode("utf8", $data) in the API.pm it fixes things like the author line here: https://metacpan.org/sourc​e/INGY/YAML-0.77/META.yml
17:03 dipsy [ META.yml - metacpan.org ]
17:04 rwstauner but i have no idea if that's an appropriate change
17:04 rwstauner i would think the user agent would do that
17:04 clintongormley joined #metacpan
17:05 rwstauner as far as i can tell this change would fix https://github.com/CPAN-AP​I/metacpan-web/issues/202
17:05 dipsy [ #202: Load META.yml in UTF-8 for display - Issues - CPAN-API/metacpan-web - GitHub ]
17:05 mo rwstauner: I think it should be decoded
17:05 rwstauner but i don't know if it would be breaking something else
17:05 mo but I'm not sure what implications this has for other pages
17:05 mo because pod seems to be rendered correctly
17:06 mo because pod::simple::xhtml seems to do the decoding
17:06 doy right, because pod has encoding declarations
17:06 mo it might double encode the output
17:06 doy so it can tell what encoding it's supposed to use
17:06 doy things like META.yml don't
17:06 doy so you need to explicitly tell it
17:06 mo right
17:07 doy assuming that META.yml is always utf8 seems reasonable to me
17:07 rwstauner i think i'm angering the api
17:07 doy i don't know how reasonable that assumption would be for files beyond that
17:08 rwstauner https://metacpan.org/source/RWSTAUNER​/Dist-Zilla-Plugin-Run-0.011/Changes
17:08 dipsy [ Changes - metacpan.org ]
17:08 rwstauner first paragraph there has the same issue
17:08 doy perl source also has encoding declarations, so blindly applying utf8 decoding to that would probably also be broken
17:08 mo rwstauner: also, this seems to affect raw responses only, json data seems to decoded ok as well
17:09 mo doy: ascii is compatible with utf8, so I'd say it doesn't really matter since it doesn't make it any worse
17:09 doy mo: but it's not ascii, it's latin1
17:09 mo ah right
17:09 doy json is probably fine because JSON.pm outputs utf8 by default
17:09 rwstauner i was trying to figure out the difference b/t my plain string and the api response, and Devel::Peek suggested one was utf8 encoded and one wasn't
17:10 mo don't rely on the utf8 flag
17:11 mo rwstauner: could you make a branch?
17:11 mo then we can test the hell out of it :)
17:11 rwstauner i really have no idea what i'm doing with this stuff, just trying to figure out the differences
17:11 rwstauner ok
17:11 mo the encoding from the api server seems to be fine: http://api.metacpan.org/sou​rce/INGY/YAML-0.77/META.yml
17:11 rwstauner module source is also wrong, currently
17:11 mo same problem I guess
17:11 rwstauner well, that's what i've been trying to figure out
17:11 doy for module source, you probably want to scan for 'use utf8' or so
17:11 rwstauner if i dump the response to STDERR it looks fine to me
17:12 rwstauner but somewhere in the template it screws with it
17:12 doy rwstauner: your terminal locale is likely utf8
17:12 rwstauner doy: it is
17:12 rwstauner but Devel::Peek showed a difference b/t two strings
17:13 rwstauner so i was wondering if perhaps the response from the api *is* utf8, and it's just not getting the flag so the templater doesn't treat it right
17:13 rwstauner but again, i have no idea what i'm doing
17:13 doy that's possible
17:14 doy the trouble is that you can't just assume that everything coming from the api server is utf8
17:14 rwstauner right
17:14 doy because the api server is just serving up the raw files
17:14 rwstauner the question is, can we tell?
17:14 mo let's just assume utf8, it works for 99% of all files
17:14 doy i think .yml and .json would be safe to treat as utf8
17:14 rwstauner mo++
17:14 doy for .pm, you can scan for 'use utf8'
17:15 rwstauner doy: i'm actually looking at the pod
17:15 doy for .pod, you can scan for '=encoding utf8', although the renderer already seems to do this
17:16 rwstauner how about .pm with pod in it?  :-)
17:16 doy really, anything you'd otherwise be rendering as plain text should probably assume utf8
17:16 doy because you have to assume something
17:16 rwstauner i agree
17:16 doy rwstauner: you still need =encoding utf8 i believe
17:17 doy .pm and .pod should respect the encoding declarations
17:17 doy (and Makefile.PL, and things like that)
17:17 doy scripts, etc
17:17 mo make sure you do something like eval { $raw = Encode::decode_utf8($pod, Encode::FB_CROAK) };
17:19 mo I'd rather start ut8_decoding all files and add exceptions as we come across bad encodings
17:19 mo you can have utf8 in your source without "use utf8"
17:19 rwstauner me too
17:19 doy if you do, you're doing something broken
17:19 rwstauner true, it was working for me
17:20 rwstauner i thought 'use uft8' was just for identifiers
17:20 mo iirc it sets the utf8 flag on strings
17:20 doy right
17:20 doy so technically you could have strings with utf8 in them without utf8, they just won't be marked as utf8
17:20 rwstauner oh, right, Devel::Peek showed me that
17:20 doy which is probably going to break things
17:21 rwstauner doy: depends on your terminal settings ;-)
17:21 doy unless you explicitly handle that, at which point, why aren't you doing 'use utf8
17:21 doy '
17:21 doy rwstauner: just because it happens to work fine in some settings doesn't mean it isn't broken
17:21 doy (:
17:21 rwstauner totally agree
17:21 theory joined #metacpan
17:22 mo but we are talking about the tiny little fraction of files that actually use latin1 chars > 128 which might be rendered wrong, right?
17:23 mo and files with a completely different encoding
17:23 rwstauner i'd guess that would be a much smaller set of breakage than misrepresting the files that actually are utf8
17:23 mo right
17:24 rwstauner ok, i'll look at it a little more and push a branch later
17:25 rwstauner thanks. mo++ doy++
17:26 doy well, i'm just in general against "oh, we'll fix your encodings for you, don't worry", because it encourages people to continue to do broken things with encodings, which is a big enough problem as it is(:
17:26 doy but yeah, shrug
18:13 frew joined #metacpan
18:25 mauke left #metacpan
18:30 dpetrov joined #metacpan
18:51 jsut_ joined #metacpan
19:45 virtualsue joined #metacpan
19:50 theory joined #metacpan
19:52 theory joined #metacpan
19:53 dpetrov seems like DBIx::Class::InflateColumn::Object::Enum  is not indexed
20:15 theory joined #metacpan
22:30 miyagawa_ joined #metacpan
23:41 theory joined #metacpan

| Channels | #metacpan index | Today | | Search | Google Search | Plain-Text | summary