The web in a box - a next generation web framework for the Perl programming language

IRC log for #mojo, 2017-11-18

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

All times shown according to UTC.

Time Nick Message
00:58 marty joined #mojo
01:08 kaare_ joined #mojo
01:16 aborazmeh joined #mojo
01:20 ccakes_ joined #mojo
01:21 VVelox joined #mojo
01:25 Grinnz https://metacpan.org/release/DBOOK/Minion-Backend-SQLite-3.000
01:38 mohawk neat
01:39 mohawk just been looking over Mojo::SQLite and the example blog app there
01:39 mohawk conclusion: also neat
01:39 Grinnz that example is just a copy of the example shipped with Mojo::Pg :P
01:40 mohawk so i gathered from the commit message
01:40 mohawk still neat ;-)
01:48 ccakes_ joined #mojo
01:54 ccakes_ joined #mojo
01:57 ccakes_ joined #mojo
02:38 disputin joined #mojo
02:39 slimerzz joined #mojo
03:09 noganex_ joined #mojo
03:23 mohawk preaction, this is the gql schema generated so far for v3 of cpantesters swagger: https://github.com/graphql-perl/GraphQL-Plugin-Convert-OpenAPI/blob/master/t/01-schema.t#L17-L258
03:23 mohawk the Query is to be ignored, but it's getting there
03:33 preaction mohawk: this is pretty interesting. did some of these names require you to add docs/something to the OpenAPI spec?
03:34 mohawk no, why?
03:34 mohawk oh, things like LanguageName i generated since it's anonymous in openapi-land
03:34 preaction it seems like some of the identifiers for the classes are spot-on, but i know just last week that Swagger... yeah, that
03:35 preaction Swagger just had "Anonymous Thing ###" and i promised to jabberwok that i would correct that
03:35 preaction so, if you had already done so in the course of your labors, i was going to steal your work ;)
03:35 mohawk sorry ;-)
03:35 mohawk however, that seems like the right scheme to follow
03:35 preaction no worries. a slog of docs for me then!
03:35 mohawk if you do make proper types of them, the issue goes away ;-)
03:36 mohawk i expect that would need a v4, though
03:36 preaction yeah, i've got to clean that up a bit
03:36 preaction eh. i find cosmetic changes or non-breaking additions to be non-version-requiring
03:36 mohawk already your spec is more sophisticated than the "petstore" one the JS version uses, with more swagger features
03:37 preaction v4 is likely when the language becomes part of the path. right now, there's no api that lets you retrieve Perl 6 reports / statistics
03:37 vicash preaction: in ETL::Yertl v0.037 line 220 of ETL::Yertl::Adapter::influxdb when you multiply $epoch with 10**9 , perl converts it to scientific notation which your code prints in the line that gets sent to influxdb and it gets rejected.
03:37 mohawk well, if you zap all the anon types into proper types, i'd reckon that would want a new version
03:37 mohawk i'd suggest not worth doing
03:37 mohawk the really "correct" way forward is to switch completely to graphql :-)
03:38 preaction vicash: ... i saw that a few times in my tests, but for some reason it's not affecting me in production
03:38 mohawk with its famous lack of need to version APIs anymore
03:38 mohawk i cheerfully concede i may be biased here
03:38 vicash preaction: ok. i think it is a matter of the version of perl being used. I am using 5.20.2 in Debian jessie and maybe that's causing it and later perls dont
03:39 preaction i'm using 5.20.2 in Debian jessie :(
03:40 preaction oh, oh oh oh. i'm not using _timestamps at all_
03:40 preaction because influx defaults the timestamp to now()
03:40 preaction so, bug. let me try something quick
03:41 preaction vicash: what does this print on your machien? perl -E'say +time * 10 ** 9'
03:41 preaction er... or maybe we should go to #yertl because yertl isn't mojo?
03:41 vicash preaction: i just hacked the installed module to use sprintf("%ld", ($epoch * 10**9)) and it worked
03:41 Seth joined #mojo
03:42 preaction okay. PR incoming then? ;)
04:20 mohawk long guilty silence ;-)
04:21 preaction nah, we moved the conversation
04:21 mohawk fair
04:57 [1]mohawk joined #mojo
05:04 dboehmer joined #mojo
07:03 trone joined #mojo
07:29 Vandal joined #mojo
07:34 jamesaxl joined #mojo
07:54 sh14 joined #mojo
08:13 sh14 joined #mojo
08:21 dod joined #mojo
08:26 dod joined #mojo
10:45 karjala_ joined #mojo
10:47 geospeck joined #mojo
11:59 kes joined #mojo
13:38 karjala_ joined #mojo
13:56 vicash left #mojo
14:35 geospeck joined #mojo
14:43 sri does @ not work in file names on some windows systems? https://groups.google.com/d/msg/mojolicious/YXfeJpYSKwA/q8T4Qoz7AgAJ
14:44 maschine seems to work fine in Windows 10 and Windows 7
14:57 jamesaxl joined #mojo
15:16 zivester joined #mojo
15:26 pink_mist used to work just fine on winxp back when I was using that
15:27 pink_mist maybe not on fat32 filesystem, I only used ntfs
15:30 haarg more likely a make issue
15:30 haarg ^ is the escape character for cmd.exe
15:51 sri haarg: ah, so it's the @ that's tripping up the make
15:52 haarg i think what's happening is that it's trying to quote the @ for cmd.exe, but then using gmake, which doesn't use cmd.exe
15:52 haarg older strawberry used dmake
16:06 ChmEarl joined #mojo
16:06 good_news_everyon joined #mojo
16:06 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/vF9AA
16:06 good_news_everyon mojo/master a38cf23 Sebastian Riedel: fix installation problems with some versions of Perl on Windows
16:06 good_news_everyon left #mojo
16:10 good_news_everyon joined #mojo
16:10 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/vF9xU
16:10 good_news_everyon mojo/master e1f26e5 Sebastian Riedel: bump version
16:10 good_news_everyon left #mojo
16:53 preaction would it be useful to have a way to make the minion admin gui read-only? i need to sort out my auth strategy, but i also want to test the gui thing quick
17:18 mohawk the new firefox has actual >1 processes!!! but the icon in my taskbar now looks like it's infected
18:11 dod joined #mojo
18:30 trone joined #mojo
19:46 preaction Minion::Backend::mysql 0.11 now has support for the Minion 8 and the admin dashboard: https://metacpan.org/release/PREACTION/Minion-Backend-mysql-0.11
19:54 perler_another joined #mojo
20:02 dod joined #mojo
20:04 geospeck joined #mojo
20:05 jberger preaction++
20:10 geospeck joined #mojo
20:10 sri preaction++
20:11 sri one thing i forgot to add to the admin ui was a flash message after retry/stop/remove
20:11 sri right now it only shows the affected jobs (or none if everything went well with remove)
20:12 sri which can be a little confusing, that's what the flash is for
20:38 disputin joined #mojo
21:17 ghenry joined #mojo
21:19 kaare_ joined #mojo
21:52 disputin joined #mojo
21:57 mohawk preaction, Query fields are now https://github.com/graphql-perl/GraphQL-Plugin-Convert-OpenAPI/blob/master/t/01-schema.t#L124-L133 - just got to sort input types for the mutation and it should be ready
22:01 marcus preaction++
22:07 mohawk (highlight has slipped now i've also added descriptions from the schema - https://github.com/graphql-perl/GraphQL-Plugin-Convert-OpenAPI/blob/master/t/01-schema.t#L125-L142
22:10 Grinnz I did a thing... it may be a crazy thing... but it works https://github.com/Grinnz/Minion-Backend-Redis
22:12 pink_mist 0_o redis minion backend 0_o
22:14 mohawk forgive my dumbness: crazy?
22:15 Grinnz Redis is not a relational database so it needs a ... different approach to a lot of things
22:15 mohawk ha ha, ok
22:26 sri Grinnz: cool, how does it perform in the benchmark? :)
22:26 mohawk Grinnz, on the plus side, NoSQL is web-scale
22:26 mohawk i saw a 5-minute documentary about it
22:26 Grinnz that's what i need to check next :P
22:27 mohawk it may or may not also be rock-star tech; the jury is still out
22:28 sri most mainstream job queues use redis exclusively
22:28 sri *cough* sidekiq *cough*
22:29 Grinnz i'll also note, there are a ton of options for how to optimize things for how redis consumes them, and i certainly haven't explored them all yet
22:30 Grinnz the biggest thing i'm concerned about is theres not really a concept of 'migrations' so module upgrades will be tricky
22:31 sri i'd assume most new features would just end up creating a few more key/value pairs in redis
22:31 sri making upgrades easy
22:32 Grinnz not all values are strings... and it should, but then either it has to create the new structures on upgrade or account that they might not always be there
22:32 sri maybe annoying when you need to upgrade old jobs
22:32 Grinnz yeah thats what i mean
22:33 * sri nods
22:33 sri worst case your upgrade requires a queue reset
22:33 Grinnz right
22:34 sri not that i have plans for big new features at the moment :p
22:34 sri minion is fairly feature complete as far as job queues go
22:34 Grinnz i could also just not worry about that for now, and implement something when it's needed in newer versions
22:35 sri seems reasonable
22:35 sri most likely near future additions would be more stats for the ui
22:36 sri maybe a list of currently active locks
22:36 sri $backend->list_locks?
22:40 sri well, lets see what features people request for the ui :)
22:53 karjala_ joined #mojo
23:31 Grinnz think i've still got a lot of work to do on dequeueing efficiency :P

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