Perl 6 - the future is here, just unevenly distributed

IRC log for #metacpan, 2013-11-26

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

All times shown according to UTC.

Time Nick Message
01:27 mst ether: I'm not assuming. that conversation was years ago.
01:32 ether mst: are we not talking about the same thing then?
01:32 mst ether: oh, the case thing rather than the GET with body, sorry
01:33 * mst thought clinton was a mac user
01:35 ether yes, I was talking about the new Elasticsearch.pm upload, last month
01:35 ether in the github ticket he suggested that either he didn't realize/forgot about the case insensitivity situation with macs, or that he figured PAUSE would just somehow do the right thing
01:35 ether I'm not sure what he expected the right thing was, though
02:20 klapperl joined #metacpan
02:30 dipsy joined #metacpan
02:53 klapperl_ joined #metacpan
03:33 jayallen_ joined #metacpan
03:45 klapperl joined #metacpan
04:19 klapperl_ joined #metacpan
04:25 klapperl joined #metacpan
04:59 preflex_ joined #metacpan
05:26 klapperl_ joined #metacpan
06:41 [Sno] joined #metacpan
08:14 dpetrov_ joined #metacpan
10:16 metacpan joined #metacpan
10:16 metacpan [metacpan-web01] ranguard pushed 2 new commits to master: http://git.io/Nd74Eg
10:16 metacpan metacpan-web/master 5361483 Michael Jemmeson: Added Lanyrd and Meetup to list of profile sites
10:16 metacpan metacpan-web/master 4204f97 Leo Lapworth: Merge pull request #1004 from mjemmeson/mjemmeson/more-profile-sites...
10:16 metacpan left #metacpan
10:16 dipsy [ Comparing 4d896714803c...4204f97ddf99 · CPAN-API/metacpan-web · GitHub ]
12:04 klapperl joined #metacpan
12:19 klapperl_ joined #metacpan
12:56 Farow joined #metacpan
13:19 andrefs joined #metacpan
13:25 andrefs joined #metacpan
13:42 jpn joined #metacpan
13:45 andrefs joined #metacpan
13:47 clintongormley joined #metacpan
13:47 ranguard clintongormley: thanks for joining :)
13:47 clintongormley heya
13:47 ranguard clintongormley: following on from our chat....
13:47 clintongormley ok, so to summarise the various issues:
13:48 ranguard 1) Can Es.pm be changed to do a POST if there is a Body?
13:48 ranguard 2) If you have users that require GET, can that be an option (with a 'This is against the RFC' mention in the docs) ?
13:49 clintongormley 1) some people disallow POST requests as a security issue, so switching GET->POST when there is a body would break things for them.
13:49 clintongormley would want to make this a configurable option.  also can specify the body as ?source= as well
13:49 clintongormley but then you are limited to some arbitrary qs length, usually 4096
13:50 * ranguard would suggest the default should be RFC compatable
13:50 clintongormley 2) I can't release a new ES.pm without screwing up Es.pm, because i didn't realise that CPAN was case-insensitive
13:51 clintongormley 3) Es::Compat exists, which gives the old ES.pm API on top of the new Es.pm internals, but does't support an AnyEvent backend yet.  (short on time)
13:51 ranguard could you release ES.pm and then a Es.pm after that?
13:51 clintongormley yeah, that'd make me even more popular ;)
13:51 * ranguard doesn't know the ramifications
13:51 andrefs joined #metacpan
13:52 clintongormley so you mentioned rt, who are the other big users of api.metacpan?
13:52 ranguard or maybe just delete ES.pm completly (I'm just offering options) if there is Es::Compat?
13:53 ranguard clintongormley: I know cpanm uses it, but think that's directly rather than through Es.pm
13:53 clintongormley ranguard: the only reason i haven't is because of the AnyEvent backend thing, leaving it there for people (specifically metacpan)
13:53 clintongormley is it directly? not using ES.pm?
13:54 * clintongormley looks at the code
13:55 ranguard https://github.com/miyagawa/cpanminus/blob/devel/lib/App/cpanminus/script.pm#L627 <- direct
13:55 dipsy [ cpanminus/lib/App/cpanminus/script.pm at devel · miyagawa/cpanminus · GitHub ]
13:55 clintongormley ranguard: note https://github.com/miyagawa/cpanminus/blob/devel/lib/App/cpanminus/script.pm#L635
13:55 dipsy [ cpanminus/lib/App/cpanminus/script.pm at devel · miyagawa/cpanminus · GitHub ]
13:55 clintongormley he's using a get there
13:56 clintongormley which wouldn't work via fastly
13:56 ranguard yea, but no BODY
13:56 ranguard so it would work
13:56 clintongormley ah right, yes
13:56 ranguard the issue isn't GET, it's GET + Body that isn't RFC compliant
13:57 ranguard just to be clear Es.pm doesn't support AnyEvent?
13:57 clintongormley correct
13:57 ranguard or jsut the Es::Compat bit?
13:57 clintongormley no Es.pm doesn't
13:57 clintongormley yet
13:57 rafl is it really? i've never seen a place where the spec forbids request bodies for get requests
13:58 clintongormley rafl: it's an age old bone of contention
13:58 clintongormley it doesn't disallow bodies with get requests
13:58 clintongormley but roy fielding thinks they should be ignored
13:58 rafl *nod*
13:58 rafl i wasn't aware of that
13:58 ranguard rafl: http://stackoverflow.com/questions/978061/http-get-with-request-body/15656884#15656884
13:58 dipsy urgh. long url. Try http://tinyurl.com/lgvn2ad
13:58 dipsy [ rest - HTTP GET with request body - Stack Overflow ]
13:58 clintongormley in the promised rewrite (promised 4 years ago) he said he would make that clearer
13:59 clintongormley and eg JS doesn't allow a body with a get request
13:59 clintongormley to my mind, using a GET request for search seems to be the more restful thing to do.
13:59 ranguard clintongormley: so, what's the likely hood that you'd be able to add AnyEvent to Es this year?
14:00 clintongormley ranguard: low. i'm working against the clock to meet my deadlines for the book right now, and adding AE is not simple
14:00 clintongormley also i'd prefer to use Promises instead of AE directly
14:01 clintongormley early next year sure, but right now difficult
14:01 clintongormley or even over christmas :)
14:01 clintongormley ranguard: what about cpan-api, does that use ES.pm with the AE backend? or not?
14:02 ranguard so, how about move all ES.pm to backpan, MC (and others) can carry on installing from there, but it's clear that people need to upgrade, and if AnyEvent is an issue they can supply a patch?
14:03 ranguard so that gets Es vs ES out of the way... ?
14:03 clintongormley any reason to move it to backpan as opposed to just leaving it where it is?
14:03 clintongormley it has the deprecated message on it
14:04 clintongormley (true, not as prominent as i would like)
14:04 ranguard I guess, just confusing
14:04 ranguard ok, so we ignore ES...
14:04 clintongormley gaah, what a mess
14:04 clintongormley <-- his fault
14:05 ranguard what about Es, can you default to POST if there is a Body, but let that be configurable for your GET only users?
14:06 clintongormley so currently no, but i could add support for that, even if it is just accepting a method parameter in the search request
14:07 clintongormley $es->search(method=>'POST', body=>{....}) etc
14:07 clintongormley want to talk to the other ES client authors, so we all implement the same thing
14:08 ranguard ok, could you let us know soonish please, and try get them to go with te default of POST :)
14:08 clintongormley the poor es.js author has more issues to deal with, eg some browsers don't accept a body with DELETE requests either...
14:09 clintongormley i'll open an issue on our list now, so we can get some resolution on this
14:09 clintongormley i'll get back to you as soon as i have anything
14:10 ranguard clintongormley: cheers
14:10 clintongormley sorry i can't just give you a quick solution :(
14:11 ranguard clintongormley: don't worry, I'll keep nudging you until there is one :)
14:12 clintongormley heh
14:12 * clintongormley changes his nick, email address, physical location, etc
14:33 Farow|2 joined #metacpan
16:21 trs left #metacpan
17:13 Farow|2 joined #metacpan
19:12 Mike-PerlRecruiter_ joined #metacpan
19:12 ranguard so, as ES/Es isn't really going to work (e.g. Es can't be released, ES doesn't have AnyEvent), would releasing MetaCPAN::ES (Es + patch for GET + Body) be reasonable?
19:23 [Sno] joined #metacpan
19:42 jayallen joined #metacpan
20:04 Farow|2 joined #metacpan
20:12 ranguard https://github.com/CPAN-API/MetaCPAN-ES - will give it a bit before doing anything
20:12 dipsy [ CPAN-API/MetaCPAN-ES · GitHub ]
21:08 daemon joined #metacpan
21:11 avar joined #metacpan
21:19 ranguard Clinton's working on sorting Es, but agrees this seems like a a best worst approach as long as we can point to Es for non-AnyEvent
21:31 oalders ranguard++
21:49 tianon joined #metacpan
23:02 Farow joined #metacpan
23:54 trs joined #metacpan
23:56 trs joined #metacpan

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