Perl 6 - the future is here, just unevenly distributed

IRC log for #metacpan, 2016-06-01

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

All times shown according to UTC.

Time Nick Message
01:02 mattp joined #metacpan
01:47 ilbot2 joined #metacpan
01:47 Topic for #metacpan is now MetaCPAN Developer VM https://github.com/CPAN-API/metacpan-developer | Chat logs available at http://irclog.perlgeek.de/metacpan/ | Can't find your module on MetaCPAN? https://metacpan.org/about/missing_modules
05:09 vanstyn joined #metacpan
05:38 kentnl joined #metacpan
07:19 oiami joined #metacpan
08:05 Relequestual joined #metacpan
08:12 metacpan joined #metacpan
08:12 metacpan [github-meets-cpan] mickeyn pushed 1 new commit to master: https://git.io/vrbiQ
08:12 metacpan github-meets-cpan/master ba99e03 Mickey Nasriachi: updated snapshot
08:12 metacpan left #metacpan
10:05 sawyer so, hmm... i'm trying to track some stuff
10:06 sawyer `MetaCPAN::Client->new->module("if")->distribution` -> "if"
10:06 sawyer which leads to http://cpan.metacpan.org/authors/id/I/IL/ILYAZ/os2/582+/perl_mlb.zip
10:07 haarg yeah
10:07 sawyer which is an archive of 5.8.2 (which i assume is for os2)
10:07 haarg unfortunately indexing perl releases is one of our biggest weaknesses
10:07 sawyer i would hope that module("if")->distribution would lead to the latest stable perl
10:07 haarg perl_mlb shows up way too much :/
10:07 sawyer you can't find "perl_mlb" in metacpan.org web
10:08 sawyer we can ask ilya to delete it, but we also shouldn't think "oh, that's where if.pm is from" - i think that's a bug
10:09 haarg asking to delete it is solving it from the wrong end really
10:10 sawyer see the word "but" and the following text, in my previous line :)
10:10 sawyer as in, i think that, while it can be done, it's (as you say) not the actual problem
10:11 haarg i haven't dug into the indexing model to figure out how this should be fixed, so on this point i'm just a voice in the noise
10:15 sawyer but a handsome voice nonetheless! :)
10:17 haarg https://metacpan.org/release/RJBS/if-0.0606 more fun
10:18 sawyer already spotted and pinged rik :)
10:18 haarg shows as unauthorized even though it's the indexed version in 02packages
10:20 sawyer yeah
10:22 haarg and also rjbs has comaint
10:22 haarg maybe just needs a reindex
10:30 sawyer http://perldoc.perl.org/perlos2.html
10:31 mickey the client query needs to be inspected, there are 482 matches for module "if" in various distributions
10:35 haarg the metacpan site is wrong in the same way
10:40 mickey haarg: of course, the client fetches "https://metacpan.org/module/if" :)
10:41 mickey i was just querying the ES to get the numbers
10:54 ranguard sawyer: also asking to delete it won't do anything as we are a backpan
10:55 ranguard or rather I guess it might do _something_ but not what you expect
10:55 mickey ranguard: I think sawyer's point was that "https://metacpan.org/module/if" needs to go to the right place... we can special case all the 'perl' related stuff
10:55 ranguard mickey: fair point
11:08 ilmari https://explorer.metacpan.org/ won't load any of the examples
11:08 ilmari nothing in the js console either
11:09 ranguard ilmari: 401 from api.gh.com for the gists
11:09 ilmari ah, yes
11:09 ilmari just noticed in the network tab
11:09 ilmari message:"Bad credentials"
11:09 ilmari documentation_url:"https://developer.github.com/v3"
11:10 ilmari oauth token expired?
11:13 ranguard ilmari: I know they don't like a public token any more, can't work out if they block it or not
11:16 ranguard ilmari: http://irclog.perlgeek.de/metacpan/2016-02-03 <- knew I'd seen it before :)
11:16 ranguard we do now have postgres btw :)
11:17 ranguard mst:  is even working on the replication / stuff
11:19 ilmari \o/
11:21 sawyer ranguard: i was wrong, it shouldn't be deleted, as it's documented as to-be-used on os2
11:29 ranguard sawyer: heh, so we need a search that's OS aware :) ?
11:35 sawyer ranguard: no, we need to know that the distribution for if.pm is either "if" (on CPAN thanks to rjbs - I was wrong to suggest it being deleted) or latest stable "perl" distribution
11:35 sawyer i'm sure others are wiser at determining which
11:35 sawyer i would take upstream
11:36 sawyer so, if if.pm is both in CPAN and "perl", but the upstream is "perl", then latest stable "perl"
11:36 sawyer instead, it picked some variation of 5.8.2 that it found in a zip file of a specific user
11:36 sawyer the reason it didn't fail as UNAUTHORIZED is that ILYAZ is a pumpking, so he's allowed to upload stuff like that as AUTHORIZED
11:37 ilmari dual-life modules should never be indexed in perl
11:37 sawyer (interestingly, though, rjbs' release of "if" distribution) *does* appear as UNAUTH
11:37 sawyer ilmari: good point, you wouldn't bring "perl" as a depedency (only version of it)
11:37 ilmari and it is indeed in no_index in perl's META.json
11:37 sawyer but some pragmas that are listed as prereqs *are* only "perl" dist, like "strict", "warnings", "deprecate", etc.
11:39 ilmari see Porting/makemeta for what's included in no_index
11:39 ilmari RJBS _does_ have comaint on if
11:39 ilmari https://pause.perl.org/pause/authenquery?pause99_peek_perms_by=me&amp;pause99_peek_perms_query=if
11:39 sawyer ilmari: he should
11:40 sawyer https://metacpan.org/release/if shows unauthorized though
11:40 sawyer that's another issue i'm throwing out there :)
11:40 ilmari but did he have it when the dist was indexed?
11:41 sawyer oh... yeah, you suggested reindexing, i forgot that!
11:41 ilmari it is indexed as his release, though
11:42 ilmari $ ack '^if\b(?!::)' ~/.cpanm/sources/http%www.cpan.org/02packages.details.txt
11:42 ilmari if                               0.0606  R/RJ/RJBS/if-0.0606.tar.gz
11:42 ilmari but bizarrely ether uploaded earlier versions after rjbs uploaded that one
11:43 ilmari (according to the version dropdown on metacpan, at least)
11:43 ilmari and those _are_ unauthorized
11:50 sawyer hmm..
11:51 ilmari so it looks like it's metacpan that's confused, not pause
16:29 ether ilmari: I was backfilling earlier versions, since they'd never been released to cpan
16:29 ether I tried to talk rjbs into doing it but he wasn't that interested :)
16:29 * ether apologizes for the confusion that imparted on the software!
20:47 garu joined #metacpan
20:48 garu hi everyone! quick question: when I look for MojoX::Log::Log4perl, it finds 0.11, uploaded yesterday. I click on it, and it opens 0.10. What gives?
20:48 garu did I screw up some metadata?
20:51 Grinnz_ "click on it"?
20:51 Grinnz_ ah, probably caching, try ctrl+shift+R
20:53 dolmen joined #metacpan
20:54 garu ha!
20:54 garu sigh
20:55 garu beaten by cache. This is really not my day
20:55 garu thanks!
20:55 garu odd that I don't think I've ever seen that behaviour before in my browser (on metacpan)
20:55 Grinnz_ there really should be something that keeps cache from still being cached when a new version has been indexed, but not knowing anything about the situation I'm in no place to make suggestions
20:55 Grinnz_ It's done it as long as I can remember
20:56 garu maybe I've ctrl-shift-r the other times and just can't remember it correctly
20:57 haarg we have plans for how to improve the caching situation, but it hasn't been done yet
20:58 haarg basically all of the pages need better cache headers (must-revalidate), we need to tell fastly to cache indefinitely, and then manually purge as needed
20:58 Grinnz_ mmm
20:58 garu do they have a purge api?
20:59 haarg yes
20:59 haarg and you can tag pages such that we could purge everything for a given dist
20:59 garu I wonder how straightforward it would be to send a request whenever a new module is indexed
21:00 garu (a purge request)
21:00 haarg that's basically what we need, but it would be to purge the specific dist
21:01 haarg there's no per-user data in the pages anymore so we don't need to drastically change anything
22:14 jdv79 joined #metacpan

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