Perl 6 - the future is here, just unevenly distributed

IRC log for #metacpan, 2016-11-27

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

All times shown according to UTC.

Time Nick Message
00:16 oiami joined #metacpan
01:07 metacpan joined #metacpan
01:07 metacpan [metacpan-web] mickeyn force-pushed mickey/fix_leaderboard from cac1a53 to 98225cd: https://git.io/v1kY9
01:07 metacpan metacpan-web/mickey/fix_leaderboard 98225cd Mickey Nasriachi: fix leaderboard (read correct key from aggregations)
01:07 metacpan left #metacpan
04:16 haarg perl releases have been reindexed.  they should be showing documentation correctly now.
06:14 haarg normally cpanm does two queries to metacpan to get a download url.  for the shim, i'm tempted to have it give "invalid" results for the first query, including the download url.  then pick up that url in the second query and return it directly so i don't have to do another query.
07:54 neilb joined #metacpan
09:19 kentnl https://fastapi.metacpan.org/v1/download_url/Try::Tiny?version=%3E0.21,%3C0.27,!=0.24   # I don't understand why this gives 0.22 instead of 0.23
09:20 kentnl also, the key order of results changes for every request, is that intentional? Seems bad for caching
09:23 haarg that end point isn't cached, and i'm not sure if we will both doing so
09:24 haarg 0.23 is backpan
09:25 haarg backpan releases are only picked up for == versions
09:27 kentnl is there a way around that?
09:27 kentnl ie: like dev=1
09:27 kentnl reason: I'd hoped to eventually find a way to do "minimum version that satisfies constraint" queries somehow
09:28 haarg dev=1 also excludes backpan
09:29 haarg we could easily add a backpan=1 option
09:34 haarg we could also trivially give canonical json for the end point, but it doesn't really matter for caching.  if we did cache it, it would be based on input parameters, not output.
10:56 neilb joined #metacpan
12:03 metacpan joined #metacpan
12:03 metacpan [metacpan-api] ranguard pushed 1 new commit to master: https://git.io/v1kKr
12:03 metacpan metacpan-api/master c567b1c Leo Lapworth: Merge pull request #581 from metacpan/tsibley/backwards-compat-when-fields-provided...
12:03 metacpan left #metacpan
12:03 metacpan joined #metacpan
12:03 metacpan [metacpan-api] ranguard deleted tsibley/backwards-compat-when-fields-provided at f7850d3: https://git.io/v1kKo
12:03 metacpan left #metacpan
12:04 metacpan joined #metacpan
12:04 metacpan [metacpan-web] ranguard pushed 1 new commit to master: https://git.io/v1kKD
12:04 metacpan metacpan-web/master 2fe72f1 Leo Lapworth: Merge pull request #1823 from metacpan/mickey/fix_leaderboard...
12:04 metacpan left #metacpan
12:04 metacpan joined #metacpan
12:04 metacpan [metacpan-web] ranguard deleted mickey/fix_leaderboard at 98225cd: https://git.io/v1kKy
12:04 metacpan left #metacpan
12:11 oiami joined #metacpan
13:15 karjala I need to create a case-insensitive search for modules in MetaCPAN v1
13:15 karjala (please help)
13:15 karjala I don't know what to write in the metacpan explorer
13:17 karjala Where should I start looking for information to help me with this query I want to make?
13:18 karjala First of all, I think the URL I POST to should be "/v1/module" - is that correct?
13:22 karjala Here's what I ideally want:
13:24 karjala I want a query where I search for "mojolicious" (the module name, in lowercase or any case) and I get back just "Mojolicious" (the name of the only distro that authoritatively contains such a module)
13:24 karjala If anyone can help, I'd be grateful. Thanks
13:54 ranguard karjala: have you seen the download_url end point ... and does that help?
13:55 karjala I haven't seen that, ranguard
13:56 karjala Where can I find more information about it, ranguard ?
13:57 karjala Maybe if I used MetaCPAN::Client instead of the API directly, I wouldn't have all these questions
13:59 ranguard see neils latest post in homepage of blogs.perl.org
13:59 ranguard but yea we recomment the client where possible
13:59 karjala Ok
14:00 ranguard heading offline - leave questions here and others will pick up later
14:00 ranguard also goto fastapi.metacpan.org to see docs
14:40 Tempesta joined #metacpan
14:48 karjala I found a bug in Neil's second article:
14:48 karjala He says that if I go to https://fastapi.metacpan.org/v1/user/ while logged in on metacpan.org, I'll see my userinfo. I only see {}
18:00 karjala Is it possible to do "match" queries with MetaCPAN::Client ?
18:01 karjala What I want to do is a case-insensitive search with MetaCPAN::Client - is that possible? I believe that ES "match" queries are case-insensitive.
18:03 karjala How does the search in the frontpage of metacpan.org happen? It is case-insensitive
18:15 karjala It's not possible to do through MetaCPAN::Client, is it?
18:23 jberger joined #metacpan
18:34 ranguard mickey: ^^ ?
18:36 ranguard karjala: https://fastapi.metacpan.org/v1/search/autocomplete?q=Moose
18:37 ranguard karjala: we are moving more and more of the ES search logic to the API.. all of this is public here: https://github.com/metacpan/metacpan-api/tree/master/lib/MetaCPAN/Server/Controller
18:38 ranguard the most interesting bits being: https://github.com/metacpan/metacpan-api/tree/master/lib/MetaCPAN/Server/Controller/Search
18:46 karjala remember not to bring it just to the API, but also to MetaCPAN::Client. For example, autocomplete doesn't exist in MetaCPAN::Client, as far as I could check.
18:48 karjala I guess I can do matches with the "all" method, of MCPAN::Client
18:49 karjala (just saw the example in the distribution)
18:52 ranguard karjala: https://github.com/metacpan/metacpan-examples might also be of use
18:53 ranguard we've not even finished moving our front end over yet, very much a work in progress
19:01 karjala oh
19:02 ranguard but MC::Client is where we are aiming for all of this to go for public consumption, so please do open issues https://github.com/metacpan/metacpan-client
19:07 ranguard mst: content-type =~ /x-script/ || =~ 'text/x-' are now set to content-type text/plain on st.aticpan.org - please play
19:09 mst wget -S looks right. firefox is still giving me a download link, but I'm wondering if it's being daft and cached the wrong answer
19:09 mst (on the .xs file in the bug report)
19:12 ranguard mst: doh, I did 'purge all' but my password filler used the wrong 'submit' button in the form
19:12 ranguard try now
19:13 mst ranguard++
19:13 mst that now works perfectly
19:13 ranguard the 'Age: XXX' header is how long Fastly has had something in cache, so a useful one to look for
19:14 ranguard do open again if you find more edge cases
19:15 mst ah, you pasted my comment so I can skip the LGTM, ace
19:15 mst haarg++
19:17 ranguard for the PITA that ES is, I'm starting to like these snapshots
22:31 jdv79 wut
22:31 jdv79 ES is a PITA?
22:33 haarg yes
22:36 jdv79 whats better?
22:40 haarg there isn't really something better, but that doesn't prevent it from being a pita
23:35 karjala mickey: I replied in that github issue of mine you closed. Mentioning it, just in case you don't get automatically notified about it.
23:36 karjala I mean this one https://github.com/metacpan/metacpan-client/issues/58

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