Perl 6 - the future is here, just unevenly distributed

IRC log for #metacpan, 2016-12-01

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

All times shown according to UTC.

Time Nick Message
00:42 jberger karjala++
00:48 xenu so there's a dist which only has a developer release ( http://search.cpan.org/~pevans/stringification/ )
00:48 xenu it's visible on sco but not on metacpan
00:48 xenu why?
00:53 haarg because we don't show developer releases in search
00:53 haarg https://metacpan.org/release/PEVANS/stringification-0.01_004
00:53 haarg we have plans to address this
00:55 xenu thanks!
02:06 damog joined #metacpan
03:25 oiami joined #metacpan
03:29 metacpan joined #metacpan
03:29 metacpan [metacpan-web] oalders created oalders/sponsors (+1 new commit): https://git.io/v1sPv
03:29 metacpan metacpan-web/oalders/sponsors d447846 Olaf Alders: Adds Advance Systems to sponsor list.
03:29 metacpan left #metacpan
03:46 Guest69 joined #metacpan
04:03 metacpan joined #metacpan
04:03 metacpan [metacpan-web] oalders force-pushed oalders/sponsors from d447846 to 6721f89: https://git.io/v1s1Y
04:03 metacpan metacpan-web/oalders/sponsors 6721f89 Olaf Alders: Adds Advance Systems to sponsor list.
04:03 metacpan left #metacpan
04:18 oiami joined #metacpan
04:26 metacpan joined #metacpan
04:26 metacpan [metacpan-web] tsibley pushed 2 new commits to oalders/sponsors: https://git.io/v1sM6
04:26 metacpan metacpan-web/oalders/sponsors 2723f59 Thomas Sibley: Rework sponsor grid into a 3x3 for aesthetics
04:26 metacpan metacpan-web/oalders/sponsors dc8fa63 Thomas Sibley: Better positioning of text in sponsor grid on front page...
04:26 metacpan left #metacpan
04:45 metacpan joined #metacpan
04:45 metacpan [metacpan-web] tsibley pushed 1 new commit to oalders/sponsors: https://git.io/v1sDD
04:45 metacpan metacpan-web/oalders/sponsors dd0830f Thomas Sibley: Add Olaf's blog post
04:45 metacpan left #metacpan
04:57 oiami joined #metacpan
05:19 oiami joined #metacpan
06:43 oiami joined #metacpan
07:34 nakiro joined #metacpan
07:42 nakiro joined #metacpan
08:22 Tempesta joined #metacpan
09:28 osfabibisi joined #metacpan
09:39 neilb_ joined #metacpan
09:51 vanstyn joined #metacpan
09:54 GumbyNET7 joined #metacpan
10:00 simcop2387 joined #metacpan
10:01 castaway joined #metacpan
10:08 Tempesta joined #metacpan
10:08 Guest69 joined #metacpan
10:08 grantm joined #metacpan
10:08 ribasushi joined #metacpan
10:08 jberger joined #metacpan
10:08 kentnl joined #metacpan
10:13 neilb joined #metacpan
10:22 karjala mickey: you wrote you would look into ways of solving the inability to match case-insensitively. Should I open a bug for this on github?
10:23 nakiro joined #metacpan
10:24 mickey no need. this is not a code bug. this is an index mapping issue. i am looking into solving this but it will take some time to get a solution live
10:27 oiami joined #metacpan
10:31 Tempesta joined #metacpan
10:32 nakiro joined #metacpan
10:34 nakiro joined #metacpan
10:57 GumbyNET5 joined #metacpan
11:49 GumbyNET7 joined #metacpan
13:18 Khisanth joined #metacpan
13:30 Relequestual joined #metacpan
13:44 karjala https://metacpan.org/pod/release/SEBAPOD/Mojo-RabbitMQ-Client-0.0.6/lib/Mojo/RabbitMQ/Client.pm
13:44 karjala the release name appears grey
13:45 karjala after so many days. why? is it not "latest"?
13:47 karjala it's also not searchable from the searchbox
13:49 haarg it isn't in 02packages
13:49 Relequestual joined #metacpan
13:51 haarg nor 06perms
13:51 haarg not sure why, but metacpan is behaving as expected
14:01 metacpan joined #metacpan
14:01 metacpan [metacpan-web] haarg created haarg/release-nonlatest-with-latest-module (+1 new commit): https://git.io/v1GPm
14:01 metacpan metacpan-web/haarg/release-nonlatest-with-latest-module da4a57e Graham Knop: use version specific release links for non-latest releases...
14:01 metacpan left #metacpan
14:43 neilb_ joined #metacpan
14:45 edward joined #metacpan
15:38 metacpan joined #metacpan
15:38 metacpan [metacpan-api] mickeyn created mickey/wip_mapping_deploy_simplified (+2 new commits): https://git.io/v1GA3
15:38 metacpan metacpan-api/mickey/wip_mapping_deploy_simplified 4827658 Mickey Nasriachi: Deploy indices/mappings without ElasticSearchX::Model...
15:38 metacpan metacpan-api/mickey/wip_mapping_deploy_simplified c159ab3 Mickey Nasriachi: module.name,documentation --> analyzed
15:38 metacpan left #metacpan
15:50 metacpan joined #metacpan
15:50 metacpan [metacpan-api] mickeyn created mickey/fix_prove (+1 new commit): https://git.io/v1Gp5
15:50 metacpan metacpan-api/mickey/fix_prove ea8486d Mickey Nasriachi: bin/prove: unset the env var that breaks things
15:50 metacpan left #metacpan
16:08 metacpan joined #metacpan
16:08 metacpan [metacpan-web] mickeyn created mickey/wip_mapping_deploy_simplified (+1 new commit): https://git.io/v1Ze5
16:08 metacpan metacpan-web/mickey/wip_mapping_deploy_simplified 565c26f Mickey Nasriachi: adapt search to updated mapping
16:08 metacpan left #metacpan
16:11 neilb_ joined #metacpan
16:56 edward joined #metacpan
17:05 neilb joined #metacpan
17:27 neilb joined #metacpan
17:54 oalders We should make it more obvious that there are keyboard shortcuts on the front end.
17:55 trs I only discovered them while at meta::hack. :-)
17:55 trs (and I did fix their rendering)
17:56 * trs just submitted a PR for Pod::Cpandoc to use the v1 API
18:13 oalders trs++
18:52 mickey oalders: nice post
18:56 ranguard evening all
18:56 oalders mickey: thanks!
18:56 oalders ranguard: allo
18:57 ranguard oalders: out of interest is there a timeframe on 06perms - or do we need to organise another hackathon soon to get it wrapped ;) ?
18:57 oalders ouch!
18:58 oalders there are a couple of unrelated errors in the branch.  i could factor that logic out
18:58 oalders *however*, mickey wants to put this in its own index
18:58 oalders and i'm not sure how to do that yet (not that i've asked)
18:58 oalders so possibly another hackathon if i get distracted, but seems quite doable ;)
18:59 mickey i think that after the last attempt to split the indices... this option is on hold
18:59 mickey but we can add it to the main index
18:59 ranguard oalders: heh, fair enough, nothing wrong with needing another hackathon btw, just wondered how much was left
18:59 oalders that makes it easier for me.  i'll take another look at the branch then
18:59 mickey shouldn't block you
18:59 oalders ranguard: well, no blockers.  i'll just get the branch passing and we can move forward
18:59 ranguard :)
19:00 oalders then look into some kind of front end integration
19:01 * ranguard nods
19:01 mickey ranguard: we can work on it together Sat. morning :)
19:03 ranguard :)
19:05 mickey reyjrar: did you look into that issue that we had with the index splitting? was wondering if you know of workarounds to that problem
19:06 ranguard mickey: to confirm, our writes would be going to the correct index, it's just that a GET with an ID requires you to query the correct index (rather than an alias) as well?
19:06 ranguard e.g. would our ESx::Model would handle the writing to an index name other than the alias?
19:10 auggy joined #metacpan
19:14 mickey yes i think that was the problem... for writing we'll bypass the model and use our own scroller
19:14 mickey *bulk_handler
19:16 mickey the model writes to the alias... we want to make sure the data goes to the right index
19:18 ranguard so should we be looking to (phase out/remove) ESx::Model? - or keep that just for reads.. or ?
19:18 mickey yes, i think we should
19:18 mickey it's another layer that doesn't keep up with ES breaking changes :)
19:19 mickey and i probably broke it anyway :)
19:19 ranguard :)
19:19 mickey see https://github.com/metacpan/metacpan-api/pull/595
19:19 mickey 1st step to removing the need for it
19:20 ranguard oh, you are ahead of me :)
19:20 mickey i have an agenda :D
19:20 ranguard mickey++
19:27 mickey ranguard: stupid question, but can't we deal with the GET requests with rewrites?
19:28 mickey we know based on the type which index should be queried
19:34 ranguard mickey: heh, I was going to suggest the same!
19:35 ranguard though I guess you could have a ID based query in a POST body? - so we'd have to do it server side (rather than at fastly)?
19:39 mickey yeah, was thinking server-side
19:40 mickey didn't think of the post requests though... might be a bit trickier
19:40 ranguard I guess 1st thing would be to add some tests that use the ID directly, which we don't seem to have :) ?
19:41 mickey ranguard++ # for volunteering to write tests :)
19:41 ranguard I think we also need to de-couple mc-api controllers from ESx::Model
19:41 * ranguard is the sysadmin dude ;)
19:41 mickey qa :)
19:42 mickey i swear there was a slash there
19:42 * ranguard goes afk (not related honest) :)
19:42 mickey the decoupling is going to be more painful... i tried removing the special attribute keys... total failure
19:43 mickey we call ->model in too many places
19:43 mickey but generally, we can slowly move towards that goal
19:47 oalders slowly would be nice if we can't do it immediately
19:52 mickey it would be too big a change
19:53 mickey and doesn't have to be all at once
20:07 oalders can someone remind me what the gray link for dist name means? https://metacpan.org/pod/HTML::Restrict
20:07 oalders that module no longer appears on my author list
20:07 oalders it's indexed, authorized and latest
20:13 trs oalders: I _thought_ the grey release-version meant it wasn't latest.
20:13 oalders hmmm
20:13 trs https://metacpan.org/release/HTML-Restrict → 404
20:13 trs https://metacpan.org/release/OALDERS/HTML-Restrict-2.2.3
20:14 trs which has "status": "cpan" in ES
20:14 trs https://explorer.metacpan.org/#/286c41375f856d407f862c5ddea41028
20:14 trs so, the release status isn't "latest", but "cpan"
20:14 oalders trs++
20:14 oalders also, how are those gists even working?
20:14 oalders i got rid of the authentication (i thought)
20:15 oalders i'll do a manual re-index
20:15 trs No clue!  :)
20:17 oalders re-indexing fixed it :)
20:17 oalders we really need some QA that periodically checks for stuff like this
20:17 trs isn't the latest process supposed to run every night to do that?
20:17 trs I think it does.
20:19 oalders Yep, it _is_ supposed to.  But I think it only grabs the last 24 hours.
20:25 mickey we could add to it a query for all distibutions that have 'cpan' status but no 'latest' to also be included regardless of time
20:32 mickey hmm... we're setting status for a release, but our data source is based on modules. so we get distribution/relase info from querying the 'file' type by module name... i can see where this can go wrong
20:33 mickey we could probably simplify the logic
20:36 ranguard mickey: tests look like they should be ok! - I've created https://github.com/metacpan/metacpan-api/issues/596 which is from live at the point we split out the types from the cpan_v1_1 index
20:39 metacpan joined #metacpan
20:39 metacpan [metacpan-api] oalders pushed 1 new commit to master: https://git.io/v1Zbl
20:39 metacpan metacpan-api/master 1989805 Olaf Alders: Merge pull request #590 from metacpan/tsibley/document-indexer-flags...
20:39 metacpan left #metacpan
20:39 metacpan joined #metacpan
20:39 metacpan [metacpan-api] oalders deleted tsibley/document-indexer-flags at 41d9475: https://git.io/v1Zb4
20:39 metacpan left #metacpan
20:42 metacpan joined #metacpan
20:42 metacpan [metacpan-web] oalders pushed 1 new commit to oalders/sponsors: https://git.io/v1ZNZ
20:42 metacpan metacpan-web/oalders/sponsors 7664e32 Olaf Alders: Advance Systems needs a text link.
20:42 metacpan left #metacpan
20:43 metacpan joined #metacpan
20:43 metacpan [metacpan-api] oalders pushed 1 new commit to master: https://git.io/v1ZNa
20:43 metacpan metacpan-api/master b4a74ee Olaf Alders: Merge pull request #594 from metacpan/mickey/fix_prove...
20:43 metacpan left #metacpan
20:43 metacpan joined #metacpan
20:43 metacpan [metacpan-api] oalders deleted mickey/fix_prove at ea8486d: https://git.io/v1ZNV
20:43 metacpan left #metacpan
20:44 metacpan joined #metacpan
20:44 metacpan [metacpan-web] oalders pushed 1 new commit to master: https://git.io/v1ZNy
20:44 metacpan metacpan-web/master 76e39fc Olaf Alders: Merge pull request #1828 from metacpan/haarg/release-nonlatest-with-latest-module...
20:44 metacpan left #metacpan
20:44 metacpan joined #metacpan
20:44 metacpan [metacpan-web] oalders deleted haarg/release-nonlatest-with-latest-module at da4a57e: https://git.io/v1ZN9
20:44 metacpan left #metacpan
20:45 mickey oalders++ #mergingspree
20:45 dolmen joined #metacpan
20:47 oalders :)
20:48 Grinnz http://1.bp.blogspot.com/-yNVFqlVAwjc/UXc8amCbFPI/AAAAAAAABVk/d1RT89fycBk/s320/merge%20road%20sign.png
20:49 trs mickey: the logic can definitely be simplified, c.f. the indexing doc I wrote during meta::hack :)
21:16 metacpan joined #metacpan
21:16 metacpan [metacpan-web] oalders pushed 1 new commit to master: https://git.io/v1ne2
21:16 metacpan metacpan-web/master 7e65a9b Olaf Alders: Merge pull request #1826 from metacpan/oalders/sponsors...
21:16 metacpan left #metacpan
21:16 metacpan joined #metacpan
21:16 metacpan [metacpan-web] oalders deleted oalders/sponsors at 7664e32: https://git.io/v1nea
21:16 metacpan left #metacpan
21:57 jdv79 ping
21:59 trs jdv79: pong
21:59 jdv79 thanks
22:08 neilb joined #metacpan
22:32 neilb joined #metacpan
23:01 haarg at some point here i want to split the MetaCPAN::Server::Controller code out into roles and apply them to the subclasses
23:02 haarg avoid the silly hacks we have in the controllers that don't represent es types
23:07 oalders +1
23:11 trs haarg: yes, jberger and I both wanted that too while looking at moving the web search endpoint over.
23:11 trs ++++
23:11 * trs went through the same pain when I inherited the primary project at $work a few years ago
23:12 haarg it probably needs a couple roles, as there are things that use a primary type and ->model, but don't want the endpoints
23:12 trs Yup.

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