Perl 6 - the future is here, just unevenly distributed

IRC log for #metacpan, 2014-02-25

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

All times shown according to UTC.

Time Nick Message
02:41 klapperl_ joined #metacpan
06:28 [Sno] joined #metacpan
06:28 kentnl joined #metacpan
06:35 kentnl joined #metacpan
06:54 ether_ joined #metacpan
07:10 neilb joined #metacpan
07:58 neilb joined #metacpan
08:55 neilb joined #metacpan
09:08 neilb joined #metacpan
11:07 [Sno]_ joined #metacpan
11:13 tobyink joined #metacpan
11:13 tobyink I saw it. ;-) https://metacpan.org/source/NEILB/Acme-Playpen-0.05/lib/Acme/Playpen/Utilities.pm#L8
11:13 dipsy [ lib/Acme/Playpen/Utilities.pm - metacpan.org ]
11:14 neilb I'm just surprised it too you 5 releases! :-)
11:14 neilb s/too/took/
11:17 tobyink 0.05 releases. That's only a twentieth of a release.
11:18 tobyink While I'm here, a vaguely #metacpan question...
11:19 tobyink How can I prevent Type::Tiny::_HalfOp from being listed so prominently. I don't want it indexed, yet I also don't want to remove its pod. https://metacpan.org/release/TOBYINK/Type-Tiny-0.039_08
11:19 dipsy [ Type-Tiny-0.039_08 - tiny, yet Moo(se)-compatible type constraint - metacpan.org ]
11:49 neilb tobyink: don't know if this will work, but you could no_index the module, but include a separate .pod file. I wonder if MetaCPAN would still show the documentation link?
11:51 omega how can I log into the webapp when I run it on my mac?
11:51 omega metacpan web app that is
11:51 omega if I select github for instance, I get sent to port 5001
11:54 ranguard run the webapp on port 5001 - login (will redirect back to live I think) but after that you should be ok on your local 5001
11:54 ranguard omega: ^^
11:54 ranguard plackup -p 5001
11:56 tobyink neilb: I've been using provides rather than no_index, but from what I can tell, that would still put Type::Tiny::_HalfOp in the same category as the genuine bits of pure documentation (which are already .pod files, and not included in provides).
12:03 omega ahh, ok, thanks :)
12:08 neilb tobyink: I think maybe I'm misunderstanding what you're trying to do. If you look at https://metacpan.org/release/NEILB/Acme-Playpen-0.06, the ::Utilities module is no_index, but has a separate .pod. That's not the effect you're after?
12:08 dipsy [ Acme-Playpen-0.06 - ltjustgt "mucking about" with dist stuff via Dist::Zilla in this $VERSION :-) - metacpan.org ]
12:09 tobyink Acme::Playpen::Utilities shows up in the "Documentation" section of the release though.
12:10 tobyink Which is where Type::Tiny::_HalfOp already is. I *don't* want it there.
12:10 neilb you want it to have pod, but that pod not be listed in the release page. If it's internal, why not documentwith comments and don't put any pod in it?
12:13 tobyink It's internal, but still potentially useful for people subclassing, etc. I still want it to be properly documented. I still want it to have a pod LICENCE section for Software::License to be able to find. Etc.
12:16 tobyink I want to have my cake and eat it.
12:16 tobyink And I want the moon.
12:16 tobyink On a stick.
12:25 klapperl joined #metacpan
12:44 tobyink neilb: that's certainly closer.
12:45 neilb yeah, somewhat downgraded, which is sort of what you're after
12:47 tobyink So perhaps I need to figure out a way for my toolchain to add packages like this to "no_index". Though really MetaCPAN should be ignoring "no_index" because I provide a "provides" bit in META.json.
12:52 neilb no_index means the module's not listed in 02packages.details.txt, which is input to MetaCPAN. Though I suspect it doesn't take that as truth, and just interprets the dist's contents, since 02packages.details.txt is only updated hourly. One for the MetaCPAN dudes now ...
12:53 ranguard rwstauner: ^^
12:53 ranguard oalders: ^^
14:02 haarg tobyink: remove the NAME section maybe?
14:03 haarg although i'm not certain how metacpan will react to that specifically
14:03 tobyink Hmm... this is what I do when adding pod to test cases. So I suppose that's a possibility.
14:09 rwstauner ask neilb about that ;-)
14:10 neilb :-)
14:21 neilb If you don't have a well-formed abstract, then the module gets displayed differently by MetaCPAN.
14:22 rwstauner oalders++ # brevity
14:22 oalders i'm just lazy :)
14:22 oalders as soon as i submitted my comment, yours appeared
14:23 rwstauner ditto
14:23 oalders i had some problems with the test suite earlier as well, so it's really hard to see what works and what doesn't
14:24 rwstauner looks to me like the docs and test suite haven't been updated to match the new api
14:25 rwstauner or the internals just aren't working
14:32 oalders i think it's a bit of both :)
14:34 haarg the tests haven't been updated at all so it's hard to know
14:58 haarg neilb: re: CPAN::ReverseDependencies, Elasticsearch.pm should be used, not ElasticSearch.pm
14:58 neilb Yeah, monken just told me that too :-)
14:58 neilb Oh, I see what you mean
14:58 neilb great, two different modules with a case difference in the name?!
14:59 haarg yeah
14:59 haarg it's not a great situation
14:59 * neilb has been bitten by that (on his mac) more than once already...
14:59 haarg but Elasticsearch is the new version which will see continued development
15:00 neilb roger that
15:01 oalders does anyone have any working code with Elasticsearch yet?
15:02 * neilb watches the tumbleweed roll by
15:04 tobyink One of my main reason for writing LV.pm was that Lvalue.pm and lvalue.pm only differed in case.
15:04 tobyink :-(
15:10 oalders neilb: regarding mo's comments about MetaCPAN::API, there's a version which uses Elasticsearch.pm under the hood that's in the pipeline
15:10 oalders that will use scrolling
15:11 neilb how long 'til it pops out of said pipeline? ;-)
15:11 neilb Or should I just get on with Elasticsearch.pm for now...
15:11 oalders well, this issue is a blocker https://github.com/xsawyerx/metacpan-api/issues/20
15:12 dipsy [ s/ElasticSearch/Elasticsearch/ · Issue #20 · xsawyerx/metacpan-api · GitHub ]
15:12 oalders so, i can't give a timeline
15:12 oalders if you want to switch to Elasticsearch, that's totally fine and maybe you'll have some feedback on it :)
15:12 oalders ElasticSearch also will work
15:13 oalders but you'd need to move away from it at some point in the future
15:13 oalders i tried briefly to get Elasticsearch to work with the API, but it wasn't working for me.  didn't put a lot of time into it, though
15:19 dolmen Glad to learn that I must avoid the Elastic[sS]earch modules as they will not install properly on Windows...
15:19 oalders :(
15:20 oalders well, if you use carton with metacpan-web, it should install properly
15:21 dolmen if you have one of them installed you will not be able to get the other if your filesystem is case insensitive
15:22 dolmen well, it looks like one is the successor of the other...
15:24 rwstauner indeed
15:46 oalders ElasticSearch vs Elasticsearch feels like a conversation we'll be having for a long time :(
15:49 haarg could inline Elasticsearch::Compat as ElasticSearch, and it would mostly be solved
15:53 rwstauner yeah, assuming it actually is compatible
15:53 * rwstauner doesn't know
15:54 haarg would be better than how things are now
15:55 rwstauner likely
15:55 rwstauner and not the first time a release broke stuff and consumers just had to update their stuff
16:37 dolmen haarg: Elasticsearch::Compat clearly says that it does'nt provide full compatibility: asynchronous API with AnyEvent are not available anymore
16:38 dolmen I can't understand why an API would go that way nowadays...
16:38 oalders i think async support is on the way
17:13 oalders this looks good to me https://github.com/CPAN-API/metacpan-web/pull/1113  any objections?
17:13 dipsy [ Fix some bootstrap/layout issues on account pages by omega · Pull Request #1113 · CPAN-API/metacpan-web · GitHub ]
17:27 bowtie_ joined #metacpan
17:34 metacpan joined #metacpan
17:34 metacpan [metacpan-web] oalders pushed 6 new commits to master: http://git.io/wc1LLw
17:34 metacpan metacpan-web/master 7c9868c Andreas Marienborg: Change styling of identities buttons...
17:34 metacpan metacpan-web/master 449f41a Andreas Marienborg: Move some custom style from style.css to account.less...
17:34 metacpan metacpan-web/master e383499 Andreas Marienborg: Change the add/remove/check links to buttons...
17:34 metacpan left #metacpan
17:34 dipsy [ Comparing 000032a94255...f10001a481da · CPAN-API/metacpan-web · GitHub ]
17:35 [Sno] joined #metacpan
17:46 neilb joined #metacpan
19:09 daxim mo, are you serious?  https://github.com/CPAN-API/metacpan-web/issues/1112#issuecomment-36015511  is this some carnival joke I'm not getting?
19:09 dipsy [ improve TLS configuration · Issue #1112 · CPAN-API/metacpan-web · GitHub ]
19:24 oalders daxim: thanks for reporting that. is your issue with mo's response to the choice of cipher suite?
19:28 daxim no, my issue is that the server enables SSLv2 which is broken. did you (admins) even look at the report? you have no actual security
19:28 daxim speaking of trade-offs for speed is just ridiculous
19:28 daxim if you want speed, don't fuck around with the ciphers, eat the low-hanging fruits first
19:29 daxim use better cacheing headers: <http://redbot.org/?descend=True&amp;uri=https://metacpan.org/author/DAXIM>
19:29 dipsy [ REDbot: <https://metacpan.org/author/DAXIM> ]
19:29 daxim encourage your users to upgrade their clients so that they take advantage of session tickets (diagnostics at <https://howsmytls.com/>)
19:32 oalders i think the caching should be taken care of when we switch to fastly on the front end
19:33 oalders i do remember the discussion way back about that particular cipher we had been using being slow on nginx
19:52 * ranguard comments as well - agree with daxim
19:53 oalders let's do what we need to do to get this sorted. i'll defer to people who know this stuff better about what exactly we need to do
19:59 * ranguard doesn't know if all of these options work on the old nginx we've got - will have to test sometime I guess
20:02 neilb I'm trying to get my head around Elasticsearch and failing. Here's my code so far for doing reverse dependencies lookup: https://gist.github.com/neilbowers/d22009d68fa70c1d8455
20:02 dipsy [ mcpan-elastic-reverse-dependencies ]
20:02 neilb How / where do I put in the equivalent of '/search/reverse_dependencies/'.$distname?
20:08 trs hrm, I'd have some code that does that to show you, but it's on a computer which is at home at not powered on right now.
20:08 trs meanwhile, you can look at the api source: https://github.com/CPAN-API/cpan-api/blob/master/lib/MetaCPAN/Server/Controller/Search/ReverseDependencies.pm#L12-L21
20:09 dipsy [ cpan-api/lib/MetaCPAN/Server/Controller/Search/ReverseDependencies.pm at master · CPAN-API/cpan-api · GitHub ]
20:09 trs and https://github.com/CPAN-API/cpan-api/blob/10cb8c43772e4d5f1fff51676b4f09e109dc4748/lib/MetaCPAN/Document/Release.pm#L193-L202
20:09 dipsy [ cpan-api/lib/MetaCPAN/Document/Release.pm at 10cb8c43772e4d5f1fff51676b4f09e109dc4748 · CPAN-API/cpan-api · GitHub ]
20:09 trs you need to map a release to all modules it provides, and then do a second query to find distributions and/or releases depending on those modules.
20:09 trs note that metacpan's current reverse_deps endpoint doesn't take into account module versions, iirc.
20:19 neilb thanks - i'll have a look. I'm not planning on taking module versions into account, at least at first
20:36 rwstauner i don't know enough about Es.pm to be able to tell you if you could use it in combination with that mc api endpoint
20:37 rwstauner specifically where you'd put that uri part
20:45 oalders the uri would just be where you'd normally name the index, i think?
20:46 oalders sorry, the type
21:47 chansen joined #metacpan

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