Perl 6 - the future is here, just unevenly distributed

IRC log for #metacpan, 2014-10-23

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

All times shown according to UTC.

Time Nick Message
00:08 klapperl joined #metacpan
00:18 klapperl_ joined #metacpan
00:23 klapperl joined #metacpan
00:49 vanstyn joined #metacpan
00:55 vroom joined #metacpan
01:12 Mso150 joined #metacpan
01:18 klapperl_ joined #metacpan
01:35 vroom joined #metacpan
04:15 vroom joined #metacpan
04:27 SineSwiper joined #metacpan
04:41 vroom joined #metacpan
05:00 vroom joined #metacpan
05:12 andreea joined #metacpan
06:11 heytitle joined #metacpan
06:40 Mso150 joined #metacpan
06:42 SineSwiper joined #metacpan
07:00 heytitle joined #metacpan
08:50 Mso150 joined #metacpan
10:11 neilb joined #metacpan
11:05 szabgab joined #metacpan
11:05 szabgab hi again
11:05 szabgab how can I fetch the source of a file?
11:05 szabgab http://api.metacpan.org/v0/file/_search?q=path:Build.PL%20AND%20release:Config-Model-Itself-1.241&limit=1
11:05 dipsy urgh. long url. Try http://tinyurl.com/nymuehf
11:06 szabgab gives me all the meta information of the file, but I am not sure how to fetch the source
11:08 oiami joined #metacpan
11:10 szabgab Oh, I see it is http://api.metacpan.org/source/DDUMONT/Config-Model-Itself-1.241/Build.PL
11:14 oiami joined #metacpan
12:21 oiami joined #metacpan
12:22 vroom joined #metacpan
14:40 vroom joined #metacpan
15:07 klapperl joined #metacpan
15:50 vroom joined #metacpan
16:10 vroom joined #metacpan
16:13 ribasushi is there a ticket that tracks the "all backpan flags are gone"?
16:14 ribasushi I can't find one, I just wanted to subscribe so I get updates in my mailbox
16:14 ranguard ribasushi: not that I know of, please open and add examples
16:15 ribasushi ranguard: um... examples... it's all of it, I thought you were aware
16:15 ribasushi (at least oalders seemed to be)
16:15 * ribasushi files ticket
16:17 ranguard I know there were some indexing issues, all the tickets I've seen oalders has fixed, but think that was more things NOT being indexed, rather than having a missing backpan flag
16:18 ranguard oalders: ^^ please add to your list to look at
16:19 ribasushi ranguard: the backpan flag is basically gone from everywhere
16:19 ranguard :(
16:21 ribasushi ranguard: https://github.com/CPAN-API/metacpan-web/issues/1399
16:21 dipsy [ The backpan flag appears to be lost for the entire archive after the reindex · Issue #1399 · CPAN-API/metacpan-web · GitHub ]
16:21 ranguard thanks
16:33 Mso150 joined #metacpan
16:37 trs oalders, rwstauner, and I briefly talked about the backpan status yesterday
16:37 trs the re-index was run without --detect_backpan, so everything in our mirror was treated as "on cpan"
16:39 vroom joined #metacpan
16:45 ribasushi trs: is another reindex running now...?
16:51 trs ribasushi: I don't believe so.
16:51 ribasushi :(
16:53 trs I'm not involved enough recently to know if that's something we can do on the live box.  I know it takes a while...
16:56 neilb joined #metacpan
16:58 trs I think the right answer is likely a new script which doesn't reindex everything, but just sanity checks the status field of each indexed release (using the same method as --detect_backpan) and updates the releases's files as necessary too.
16:58 trs this would be much quicker than reindexing, I believe.
16:58 trs unfortunately, I can't justify metacpan as $work.
17:05 Mso150 joined #metacpan
18:04 Mso150 joined #metacpan
18:29 Mso150 joined #metacpan
18:52 oalders i'll get this fixed this evening as trs described it
18:52 oalders that will be faster
18:52 oalders and i'll just generally be handy to have
19:05 Mso150 joined #metacpan
19:41 dpetrov joined #metacpan
19:50 szabgab joined #metacpan
19:50 szabgab hi
19:51 szabgab as far as I remember at the early days of MetaCPAN we were asked to upload a file called author-1.0.json to PAUSE with some initial config data
19:51 szabgab Is that file still in use?
19:52 szabgab Would it be a good idea to remove it and to encourage others to remove it so there won't be some old data confusing someone?
19:52 Mso150 joined #metacpan
20:08 trs szabgab: those files are still parsed by metacpan
20:12 szabgab oh. why?
20:12 dipsy oh. why is it ace
20:13 oalders szabgab: https://github.com/CPAN-API/cpan-api/blob/master/lib/MetaCPAN/Script/Author.pm#L103
20:13 dipsy [ cpan-api/Author.pm at master · CPAN-API/cpan-api · GitHub ]
20:14 oalders you can use it to set some author defaults
20:15 oalders and it persists in CPAN, so you don't require an API to get at the info
20:15 oalders not sure how many people still use it
20:15 szabgab But if I log in to MetaCPAN, and change some values in MetaCPAN then these two will disagree, right
20:15 szabgab ?
20:15 oalders right
20:16 szabgab so imho it is an accident waiting to happen
20:16 szabgab but maybe it is just too late for me tonight
20:16 ranguard someone, mst? argued people needed the file so they didn't _have_ to use metacpan
20:17 ranguard as MC shouldn't be a repository of all data
20:17 ranguard mst: ^^ am I making that up?
20:17 oalders ranguard: no, that's correct
20:18 szabgab but if someone uses metacpan, for that person this is just confusion
20:19 mst ranguard: you aren't
20:19 oalders well, after you've updated your account for the first time, it should never fall back to the .json file
20:19 oalders so it's ok if they get out of sync, from our perspective
20:19 mst szabgab: perhas the UI needs to make it more clear that the .json file is now going to be ignored
20:20 mst but the idea of having a commonly understood means of disseminating author data through our standard approach for disseminating information seems like a useful thing to have as well
20:21 mst also
20:21 mst if somebody's editing their information, reminding them that they should probably schedule their author.json for deletion is a good thing as well
20:21 szabgab and provides an opportunity for confusion just as the data in whe 00whois.xml vs the data MetaCPAN provides
20:22 szabgab mst: that's what I was asking
20:22 mst you made it sound like you wanted author.json support removed from MetaCPAN
20:23 mst I absolutely agree that there's potential for confusion, but "delete one of the features" seems like an imperfect way to resolve that
20:25 mst I think, basically, anybody continuing to use author.json should use -only- author.json, and anybody wanting to switch to using the metacpan interface to edit it should schedule their author.json for deletion
20:25 mst and we should find some way to make sure people realise that
20:25 mst szabgab: does that seem like, assuming we can find a sufficiently effective 'some way' UI-wise, it'd be a decent resolution?
20:26 szabgab yes
20:27 oalders if you really want to keep the two in sync, you could just upload a new author.json with the contents of http://api.metacpan.org/author/SZABGAB
20:28 szabgab oalders: I am not really interested in having author.json, but if someone is then maybe this could be explained
20:28 oalders but really it was just a way for people to edit their author data before there was a UI
20:28 szabgab in the FAQ?
20:28 oalders that sounds like the right place for it
20:29 szabgab thanks for clarifying this.
20:29 szabgab Now there is a similar issue with the account info on PAUSE
20:30 szabgab If I am not mistaken that was the source of some of the fields in the MetaCPAN account but then if I edit in either places, they will get out of sync, right?
20:30 oalders right. same prinicple
20:31 oalders *principle
20:32 oalders i saw your ticket about PAUSE info. i'm not sure what purpose it would serve to provide the underlying PAUSE data
20:32 szabgab so I wonder if it would be a good idea to keep pulling the data from PAUSE and having it in some other fields that shown in the MetaCPAN account but are read only
20:32 ranguard oalders: please stop reading/responding to IRC/github etc and go fix the indexing :)
20:32 szabgab with some explanation what is that data and where to update
20:33 oalders ranguard: nice :) it's actually non-trivial
20:33 oalders szabgab: but to what purpose?
20:33 oalders for example, if i've changed my contact info in metacpan, that's how i'd like to be contacted
20:34 szabgab well, for one so I won't have to use 00whois.xml directly when re-implementing SCO :)
20:34 oalders ok, that's _1_ use :)
20:34 szabgab oalders: right, but SCO still shows the one that you set in PAUSE
20:34 oalders but nobody has ever asked for this in 4 years
20:34 oalders szabgab:ah,i see
20:34 szabgab and by this time you (ok, me) forgot where is that value from
20:35 oalders szabgab: https://metacpan.org/pod/PAUSE::Users
20:35 dipsy [ PAUSE::Users - interface to PAUSE's users file (00whois.xml) - metacpan.org ]
20:35 oalders i'd start there
20:36 szabgab so IMHO at leas some warning should be in the MetaCPAN account that will tell the users where does SCO take the values from
20:36 trs szabgab: start providing patches for those warnings...
20:36 oalders that could go in the FAQ, i guess
20:37 oalders probably SCO should state where it gets the values from ;)
20:37 oalders not sure if we should be documented 3rd party behaviour
20:37 szabgab trs: you know, first I try to make sure I understand the situation ...
20:38 szabgab oalders: and thanks for that module link!
20:38 oalders when in doubt, i always check to see if neilb has already implemented what i need
20:39 neilb :-)
20:39 szabgab oalders: that seems to be a good process
20:39 ether xdg has been building up a common interface to PAUSE files that use a variety of backends
20:39 szabgab ok, I think I had enough irc for quite some time now
20:39 oalders ether: is it on CPAN?
20:39 ether yes.. 1sec
20:40 szabgab thanks
20:40 ether was it just me or did it sound like szabgab doesn't realize that PAUSE *is supposed* to be the authoritative information source??
20:41 ether oalders: it is CPAN::Common::Index
20:41 oalders ether: thanks!
20:47 grantm so maybe the metacpan profile page should offer an "Export as author.json" link :-)
20:47 oalders :)
20:48 neilb_ joined #metacpan
20:48 oalders the metacpan explorer link from the author page pretty much does that. we could just make it clearer
20:48 ether how many author.json files are there?
20:50 grantm ether: 3
20:51 oalders yeah, i thought it was 3 at first too
20:51 oalders but there are versioned files
20:51 oalders ie author-1.1.json
20:51 oalders 110 of those, but many authors have multiples
20:54 * grantm bows to oalders' superior investigative skills :-)
20:54 oalders i just had a vague memory of the versions :)
20:59 mst basically: if we can do it, I'd like to avoid confusing users -and- retain author.json support
21:05 vroom joined #metacpan
21:26 trs the api author document does contain some metacpan-specific stuff, like user: "sWuxlxYeQBKoCQe1f-FQ_Q"
22:10 fedelippo joined #metacpan

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