Perl 6 - the future is here, just unevenly distributed

IRC log for #metacpan, 2016-06-02

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

All times shown according to UTC.

Time Nick Message
00:50 Khisanth joined #metacpan
04:16 oiami joined #metacpan
04:18 vanstyn_ joined #metacpan
07:48 cjm joined #metacpan
07:57 oiami joined #metacpan
07:58 Relequestual joined #metacpan
08:35 metacpan joined #metacpan
08:35 metacpan [cpan-api] mickeyn created mickey/queue_from_script_release (+1 new commit): https://git.io/vrxVc
08:35 metacpan cpan-api/mickey/queue_from_script_release 79cbca8 Mickey Nasriachi: Support queueing from Script::Release
08:35 metacpan left #metacpan
10:05 metacpan joined #metacpan
10:05 metacpan [cpan-api] mickeyn created mickey/update_mapping (+2 new commits): https://git.io/vrxSY
10:05 metacpan cpan-api/mickey/update_mapping 1865b0f Mickey Nasriachi: add missing 'version_numified' to the index
10:05 metacpan cpan-api/mickey/update_mapping db0a8bf Mickey Nasriachi: remove internal attribute 'section' from the index
10:05 metacpan left #metacpan
10:51 * ilmari wonders if there's a way to get a list of dists that depend on Foo but not Bar
11:13 csson joined #metacpan
12:11 castaway joined #metacpan
12:43 oiami joined #metacpan
12:50 mickey ilmari: yes, with a bool query (ES)
12:55 mickey curl -XGET "https://api.metacpan.org/release/_search" -d'{"query":{"bool":{"must":[{"term":{"dependency.module":"Moose"}}],"must_not":[{"term":{"dependency.module":"Try::Tiny"}}]}},"size":0}'
12:56 mickey # hits.total = 20265
12:56 mickey vs.
12:56 mickey curl -XGET "https://api.metacpan.org/release/_search" -d'{"query":{"bool":{"must":[{"term":{"dependency.module":"Moose"}}]}},"size":0}'
12:56 mickey # hits.total = 23935
12:59 mickey you can also use MetaCPAN::Client with es_filter for that
13:22 ribasushi joined #metacpan
14:35 oalders mickey++
15:04 ilmari mickey: cheers
15:04 ilmari there are apparently only four dists tha depend on Sub::Quote and not Moo
15:05 ilmari and two of them are MooX dists
15:06 ilmari oh, it returns all versions
15:08 ilmari ah, adding a status:latest term helps
15:12 ilmari also "fields":["distribution"]
15:30 mickey yeah, you can refine your query. that was just the core of X and not Y :)
15:34 ilmari my final query was this:
15:34 ilmari curl -s -XGET "https://api.metacpan.org/release/_search" -d'{"query":{"bool":{"must":[{"term":{"dependency.module":"Sub::Quote"}},{"term":{"status":"latest"}}],"must_not":[{"term":{"dependency.module":"Moo"}},{"term":{"dependency.module":"Moo::Role"}}]}},"size":100,"fields":["distribution"]}' | jq -r '.hits.hits[]|.fields.distribution'
15:34 ilmari jq++
15:35 ilmari the | in the jq expression is not strictly necessary
15:35 metacpan joined #metacpan
15:35 metacpan [metacpan-client] stevan opened pull request #39: Adding a `source` method to MetaCPAN::Client::File (master...master) https://git.io/vrpwK
15:35 metacpan left #metacpan
15:41 castaway joined #metacpan
15:43 metacpan joined #metacpan
15:43 metacpan [metacpan-client] mickeyn closed pull request #39: Adding a `source` method to MetaCPAN::Client::File (master...master) https://git.io/vrpwK
15:43 metacpan left #metacpan
15:43 metacpan joined #metacpan
15:43 metacpan [metacpan-client] mickeyn pushed 2 new commits to master: https://git.io/vrprj
15:43 metacpan metacpan-client/master 827b33b Stevan Little: Adding a `source` method to MetaCPAN::Client::File
15:43 metacpan metacpan-client/master 73dbcc3 Mickey: Merge pull request #39 from stevan/master...
15:43 metacpan left #metacpan
15:48 metacpan joined #metacpan
15:48 metacpan [metacpan-client] mickeyn pushed 1 new commit to master: https://git.io/vrpKf
15:48 metacpan metacpan-client/master 8fdc4bd Mickey Nasriachi: release 1.015000
15:48 metacpan left #metacpan
16:22 oalders did not know about jq. nice.
16:27 jberger and because cpan is crazy, https://metacpan.org/pod/yq
16:35 oalders didn't know about that either. :)
16:43 perigrin not just cpanis crazy ... preaction is crazy.
17:34 jberger that too
20:16 dolmen joined #metacpan
21:54 Grinnz_ huh, why is Changes the only thing listed under Other Files for this dist? https://metacpan.org/release/OCBNET/CSS-Sass-3.3.4
21:55 Grinnz_ browsing clearly shows it has all the usual things
21:56 haarg headings have to start with the version number
21:57 Grinnz_ headings?
21:57 Grinnz_ oh i mean, why isn't META.json/Makefile.PL/etc listed
22:04 haarg no
22:04 haarg in Changes
22:04 haarg the version heading lines have to start with the version
22:04 haarg or wait
22:04 haarg i misunderstood your question entirely
22:06 ether going to that release page crashed my firefox (!)
22:07 haarg i'm guessing that there are too many files in the dist and the filter is doing the wrong thing
22:07 Grinnz_ huh
22:07 Grinnz_ page loads pretty normally in chrome, just misisng all the stuff
22:08 haarg the page loads fine for me in firefox
22:09 ether oh my goodness there's a lot of files in this distribution.
22:09 ether find . | wc   -> 20907
22:09 haarg yeah
22:10 Grinnz_ :s
22:30 metacpan joined #metacpan
22:30 metacpan [cpan-api] oalders pushed 1 new commit to mi/es2: https://git.io/vrhMm
22:30 metacpan cpan-api/mi/es2 37f52c6 Olaf Alders: Merge pull request #489 from CPAN-API/mickey/queue_from_script_release...
22:30 metacpan left #metacpan
22:30 metacpan joined #metacpan
22:30 metacpan [cpan-api] oalders deleted mickey/queue_from_script_release at 79cbca8: https://git.io/vrhMY
22:30 metacpan left #metacpan
22:31 haarg the search for files includes anything with a filename starting with eg or ex
22:31 haarg but then that is later filtered by with perl code to exclude anything it t or xt
22:43 metacpan joined #metacpan
22:43 metacpan [metacpan-web] haarg created topic/no-t-other-files (+2 new commits): https://git.io/vrhyh
22:43 metacpan metacpan-web/topic/no-t-other-files 85c6636 Graham Knop: use Config::ZOMG rather than Config::JFDI
22:43 metacpan metacpan-web/topic/no-t-other-files 1bf6c3a Graham Knop: exclude t and xt directory from other files at search time...
22:43 metacpan left #metacpan
22:44 metacpan joined #metacpan
22:44 metacpan [metacpan-web] haarg force-pushed topic/no-t-other-files from 1bf6c3a to 90c2ba7: https://git.io/vrhST
22:44 metacpan metacpan-web/topic/no-t-other-files 90c2ba7 Graham Knop: exclude t and xt directory from other files at search time...
22:44 metacpan left #metacpan
22:44 haarg Grinnz_: https://github.com/CPAN-API/metacpan-web/pull/1700
22:44 Grinnz_ haarg++
22:45 haarg there's some questionable things about how we are searching, but for now i went for solving the immediate problem
22:47 Grinnz_ hmm, the ex/eg prefix stuff looks like it's trying to apply to directory names, not files, dunno if there's any way you can differentiate that in the query though
22:52 haarg the way it's written, it explicitly matches filenames
22:53 Grinnz_ oh so that's actually matching like example_foo.pl and not ex/foo.pl?
22:53 Grinnz_ or both?
22:53 haarg it's matching any file in a directory starting with ex (from the root, not a subdir) or any file starting with ex
22:54 Grinnz_ ah, the separate 'name' and 'path' conditions
22:54 haarg yes
22:55 Grinnz_ the filename part doesn't seem useful to me
22:55 haarg and then https://github.com/CPAN-API/metacpan-web/blob/master/lib/MetaCPAN/Web/Controller/Release.pm#L96-L97
22:56 haarg well, it makes more sense for example
22:57 haarg for now i'm ignoring the odd choice of things to include and just fixing the part where we return then throw away a bunch of them
23:12 mickey do we really need those files in the index to begin with... they also affect sorting in ES and generally inflate the index, just sayin'
23:13 mickey maybe this should be an api change?
23:14 haarg everything goes into the index
23:14 mickey i know it does, i ask why :)
23:16 mickey file index search on distribution=Moose gives "total": 152476 results... seems way too many entries than we would ever want to search through
23:16 mickey maybe i'm wrong... just trying to find the reason for it
23:16 haarg i'm not sure why we would want to exclude them
23:17 haarg the index size hasn't ever been an issue
23:17 mickey indexing time is one thing
23:18 haarg that only really matters for a full reindex
23:18 mickey and do we ever have a use case for file = "010_array_from_role.t"
23:18 haarg and once you've done that we need to know up front what every use case is
23:19 haarg or you have to keep reindexing a you make changes
23:19 haarg *as
23:20 haarg maybe someone might want to look at the count of test scripts across different dists
23:21 mickey sure, but those metrics can be an index in itself :)
23:21 haarg and again
23:21 haarg now you have to decide on your use cases first
23:21 haarg any they have to wait for us to implement them
23:21 haarg and then we need to do a full reindex
23:22 mickey k, i get your point

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