Perl 6 - the future is here, just unevenly distributed

IRC log for #metacpan, 2016-09-27

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

All times shown according to UTC.

Time Nick Message
05:03 oiami joined #metacpan
06:59 neilb joined #metacpan
08:08 oiami joined #metacpan
08:12 Relequestual joined #metacpan
08:15 neilb joined #metacpan
09:22 Relequestual joined #metacpan
10:05 osfabibisi joined #metacpan
10:27 metacpan joined #metacpan
10:27 metacpan [metacpan-web] mickeyn created mickey/move_code_from_templates (+1 new commit): https://git.io/vijVq
10:27 metacpan metacpan-web/mickey/move_code_from_templates c502e9c Mickey Nasriachi: author page: moved logic from template to module
10:27 metacpan left #metacpan
10:45 ranguard cheers
10:52 zamanf Is there a way to convert firefox cookies to lwp format?
11:04 ranguard zamanf: hi - you probably want #lwp :)
11:09 Tempesta joined #metacpan
13:10 neilb joined #metacpan
13:30 Relequestual joined #metacpan
14:25 Khisanth joined #metacpan
15:25 metacpan joined #metacpan
15:25 metacpan [metacpan-api] mickeyn created mickey/wip_new_endpoints (+1 new commit): https://git.io/vPemd
15:25 metacpan metacpan-api/mickey/wip_new_endpoints ed082c6 Mickey Nasriachi: created /author/by_user endpoint to replace web direct query
15:25 metacpan left #metacpan
15:59 metacpan joined #metacpan
15:59 metacpan [metacpan-web] mickeyn force-pushed mickey/move_code_from_templates from c502e9c to 965590c: https://git.io/vPeZ9
15:59 metacpan metacpan-web/mickey/move_code_from_templates 965590c Mickey Nasriachi: author page: moved logic from template to module
15:59 metacpan left #metacpan
16:31 metacpan joined #metacpan
16:31 metacpan [metacpan-web] oalders pushed 1 new commit to master: https://git.io/vPe4n
16:31 metacpan metacpan-web/master 01cab45 Olaf Alders: Merge pull request #1776 from metacpan/mickey/move_code_from_templates...
16:31 metacpan left #metacpan
16:58 neilb joined #metacpan
17:42 ranguard mickey: so, our API (metacpan-api code base) does not seem to support DELETE, other than for user and user/favs
17:42 ranguard oalders: ^^
17:43 oalders right. so we need to whitelist DELETE for the scroll ids?
17:44 ranguard we are throwing 500's on the current API - all relating to scroll
17:44 ranguard so this has never worked as far as I can see
17:44 oalders :(
17:45 oalders ranguard: could you open an issue for that?
17:45 ranguard or I can just hide it at fastly and we get on with other stuff :) ?
17:46 oalders heh. can we do both?
17:46 oalders that may be one of the reasons that ES needs the odd reboot
17:47 oalders you can label it for meta::hack
17:47 ranguard ok, I have a feeling that it's using DELETE + body, which I'm not sure fastly supports we might be lower level that we need to 'fix' it
17:49 ranguard there would need to be some serious security checks here, as this is effectivly giving access to deleting on our main index :(
17:49 oalders sounds good.  it's at least wroth looking into.  we'd have to lock it down for sure
17:55 ranguard oalders: happen to know what gets returned for the delete :) ? - long shot I know :)
17:56 oalders https://www.elastic.co/guide/en/elasticsearch/reference/current/search-request-scroll.html
17:56 oalders that doesn't actually show the return values
18:01 * ranguard wishes fastly had a notes field against changes... just seen I've been blocking all DELETE requests
18:01 * ranguard doesn't remember doing that :(
18:02 oalders i'll see if i can get a return value for that DELETE
18:03 ranguard it was just so I can fake it, but not required :)
18:03 ranguard the clients on our live are getting 500, and on fastapi a 405 and no one's complained
18:04 ranguard logging++ (I'm sending any request with 400+ error to sumlogic at the moment for fastapi)
18:07 oalders ranguard: the successful DELETE returns "{}"
18:07 ranguard thanks
18:07 oalders ranguard: if you want to play with it, you can grab a scroll id via https://fastapi.metacpan.org/v1/author/_search?scroll=5m&search_type=scan&size=100
18:09 ranguard wonder why only /_search/scroll gets DELETE requests
18:10 ranguard guess many no one searching on authors with a client that does scrolling?
18:12 oalders i didn't quite get what you're asking
18:12 oalders the ES docs point to /_search/scroll for DELETE
18:13 ranguard oh, for _any_ search on _any_ index
18:13 oalders right. you just need the id
18:13 * ranguard assumed it was per index, but guess it doesn't have to be
18:13 * ranguard nods
18:13 ranguard _EVEN_ more important to get security right :)
18:14 oalders you can search across multiple indexes, so that makes sense
18:14 oalders *indices
18:14 * ranguard nods
18:25 zamanf joined #metacpan
18:27 mickey ranguard: fwiw, we don't need caching on these DELETEs
18:28 ranguard mickey: I don't want them in my error logs is more the reason I did it
18:28 ranguard still trying to debug the issues Tux and I get every couple of days
18:28 ranguard but thanks :)
18:29 mickey well, yeah :)
18:29 mickey the DELETEs are sent by the client (actually by S::Es) to free up resources on the server
18:29 mickey and ES only support the ids through the body in several forms... but no support for a URL param for the id
18:30 mickey afaik
18:30 * ranguard updates ticket, Fastly support DELETE+Body
18:30 mickey we can limit the scope it needed, i don't think we need any other DELETE by clients
18:32 * ranguard nods
19:33 ranguard dyn.com++ # Just extended our DNS account by 4 years for free
19:35 oalders https://twitter.com/metacpan/status/780853127570305025
19:36 ranguard nice
19:36 alh https://twitter.com/Dyn now :)
19:36 ranguard doh :)
19:36 alh (They've switched a few times x.x)
19:37 oalders ah, it autocompleted with the correct logo, so i just assumed...
19:38 alh heh
22:28 batman joined #metacpan
22:32 batman joined #metacpan

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