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

IRC log for #mojo, 2016-10-25

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

All times shown according to UTC.

Time Nick Message
00:53 irqq_ joined #mojo
01:02 disputin joined #mojo
01:05 aborazmeh joined #mojo
01:14 noganex joined #mojo
01:15 lluad joined #mojo
02:14 genio sri: is Chrome acting up for you lately in MacOS Sierra?
02:33 deserted joined #mojo
02:51 noganex_ joined #mojo
03:03 itaipu joined #mojo
03:11 tyldis joined #mojo
03:26 asarch joined #mojo
03:49 polettix joined #mojo
05:07 stryx` joined #mojo
05:18 xinming bpmedley: If possible, I do wish to do it that way, Since I don't understand how Mojo stack works, but it seems it'll cause more trouble than I thought, I'll choose the mojo way.
05:31 inokenty-w joined #mojo
05:38 polettix joined #mojo
06:21 sri genio: nope, better than ever actually
06:26 mbudde joined #mojo
06:31 sri it is a little annoying that perl has a hook for threads being started, but not for fork... which would be actually useful
06:33 sri you'd have to actually redefine fork to get something like it
06:48 parv joined #mojo
07:11 xinming sri: Is there a tiny worker script which I can test with?
07:11 xinming http://paste.scsys.co.uk/537677    <-- This is the worker script I tested, But the job doesn't dequeue successfully
07:13 dod joined #mojo
07:18 dod joined #mojo
07:23 stryx` joined #mojo
07:24 trone joined #mojo
07:25 Vandal joined #mojo
07:27 AndrewIsh joined #mojo
07:31 xinming ignore me, I connected to the wrong database. >_<
07:32 khfeng joined #mojo
07:42 osfabibisi joined #mojo
07:52 marty joined #mojo
08:15 rshadow joined #mojo
08:22 irqq joined #mojo
08:35 bpmedley xinming : https://gist.github.com/brianmed/770ae233a258d8dd5d23317ec6c4bdf0 <-- Does this help?
09:06 ashimema joined #mojo
09:14 polettix joined #mojo
09:21 dotan joined #mojo
10:12 Janos joined #mojo
10:22 itaipu joined #mojo
11:07 skrig joined #mojo
11:28 xinming bpmedley: My versioon doesn't work because I point to the wrong database, where the names have only 2 characters differ which makes me lost. I'll check the url you gave
11:29 xinming bpmedley: And thanks for the short version you gave, I checked the url you gave me yesterday, Which knows how it worked. Thanks for the smaller version. :-)
12:03 polettix joined #mojo
12:03 dotan_convos joined #mojo
12:11 lluad joined #mojo
12:23 dotan_convos joined #mojo
12:23 crab joined #mojo
12:24 Sebbe joined #mojo
12:24 ccakes joined #mojo
12:24 batman joined #mojo
12:24 Grinnz joined #mojo
12:24 itaipu joined #mojo
12:24 jabberwok joined #mojo
12:24 irqq joined #mojo
12:25 dotan_convos joined #mojo
12:26 ashimema joined #mojo
12:26 dotan_convos joined #mojo
12:29 Vandal joined #mojo
13:04 gizmomathboy joined #mojo
13:12 ramortegui joined #mojo
13:18 Janos joined #mojo
13:22 zivester joined #mojo
13:32 mcsnolte joined #mojo
13:36 aborazmeh joined #mojo
14:04 kaare joined #mojo
14:05 Gedge joined #mojo
14:05 PryMar56 joined #mojo
14:06 umask001 joined #mojo
14:07 sparky joined #mojo
14:10 CW joined #mojo
14:12 simbabque joined #mojo
14:12 ssm joined #mojo
14:12 stefan joined #mojo
14:14 Zx3 joined #mojo
14:16 upasana joined #mojo
14:17 Prajith joined #mojo
14:17 kivilahtio joined #mojo
14:18 upasana_ joined #mojo
14:23 vytas joined #mojo
14:26 salva joined #mojo
14:27 polettix joined #mojo
14:31 nicomen joined #mojo
14:32 jnbek joined #mojo
14:39 itaipu joined #mojo
14:47 kamyl joined #mojo
14:55 nic joined #mojo
14:56 disputin joined #mojo
14:58 rshadow joined #mojo
15:12 zivester joined #mojo
15:13 sri hmm, i guess you can't join multiple json fields in one query in sqlite
15:13 sri postgres is way ahead of sqlite when it comes to json
15:14 sri it's silly you can't even properly match against a json array
15:15 disputin joined #mojo
15:15 Grinnz sri: join as in string, or as in sql join?
15:16 sri sql join
15:16 sri basically i have a table with two json array fields, and i want to match against both of them in a query
15:17 Grinnz ah
15:17 Grinnz yeah would be really nice to get one or two more json array functions...
15:17 Grinnz the problem is sqlite doesn't have native arrays
15:18 sri only solution i see at the moment is to make two queries that each join one of the fields and combine the results manually :/
15:19 asarch joined #mojo
15:42 PopeF0 joined #mojo
16:08 kryp joined #mojo
16:31 Janos joined #mojo
17:04 disputin joined #mojo
17:07 PopeF joined #mojo
17:26 sri yea, not really a fan of json support in sqlite
17:26 sri too spoiled by postgres
17:28 Janos joined #mojo
17:30 suede sri: indexing on json keys in postgres is ~~so nice~~
17:30 sri i mean, it's better than pretty much every other database.... but still ;p
17:31 sri (sqlite)
17:31 suede What makes sqlite better? I just have a local postgres instance on a server in my house that I use for development :P
17:31 sri native arrays in postgres are what i miss most
17:31 Grinnz suede: use mysql or oracle for a while...
17:32 suede Grinnz: I used MySQL for years until I found Postgres
17:32 suede I don't want to switch back :P
17:32 suede I've never used Oracle though
17:32 sri people like sqlite because it doesn't require you to manage a database server
17:32 sri not that i really agree with that
17:32 * sri glares at coolo
17:33 jberger Phil21 will argue about pg deployment of course
17:33 sri for me sqlite is just an alternative to flat files
17:34 Phil21 I heard my bat signal
17:34 Phil21 someone is deploying pg? we can still save them!
17:34 sri :D
17:35 suede i've been using postgres for a year and a half now- good luck getting me off it
17:35 suede Faces of Postgres
17:35 Phil21 hehe
17:35 Phil21 my complaints are entirely to do with clustering and replication
17:35 sri valid complaints imo
17:35 sect0r joined #mojo
17:35 sect0r ,
17:35 sect0r http://rarbg.to/torrents.php?search=&amp;category%5B%5D=4
17:35 sect0r http://nzbindex.nl/release/139246417/502679-FULL-a.b.eroticaEFNet-ps-no44-0137-ps-no44.par2.nzb
17:35 sect0r http://nzbindex.nl/release/139525932/503252-FULL-a.b.eroticaEFNet-kks-fumesi2016-0148-kks-fumesi2016-sample.par2.nzb
17:35 sect0r http://nzbindex.nl/release/139525889/503251-FULL-a.b.eroticaEFNet-kks-homo2016-0138-kks-homo2016.par2.nzb
17:35 sect0r http://nzbindex.nl/release/139525895/503249-FULL-a.b.eroticaEFNet-kks-filo-0127-kks-filo.par2.nzb
17:35 sect0r was kicked by Grinnz: sect0r
17:35 suede i've never used either of those features since i have nothing that's that big that requires it :P
17:35 Grinnz finally, he did it in a channel i have op in :P
17:36 suede Phil21: i'm interested in hearing about your qualms with them though
17:36 suede because i want to know what i'm up against when i finally have to deal with it :P
17:36 Phil21 suede: well, my #1 complaint is there are 14 ways to do it, all of them different and all with different drawbacks
17:36 Grinnz hm, irccloud's /kickban is not very good
17:36 Phil21 so you have a lot of fractured dev efforts
17:36 Phil21 and it shows
17:37 Phil21 but basically, there is no "easy" way to out of the box do a master/master active replicating pair, for example
17:37 Phil21 where MySQL that is trivial - I can have a HA cluster of mysql up with no silly IP takeover tricks/etc. in an hour or two
17:37 Phil21 good luck making that happen with pg at any sort of scale
17:38 Phil21 pg is more about sharding and other things like that as well at scale - which I agree are good things - but those are application things
17:38 Phil21 if I had to just tldr it...
17:38 sri with every postgres release it looks like we are tiny bit closer to native master/master replication
17:38 Phil21 MySQL seems to be designed with ops in mind as the primary consumer, and pg seems to be developer focused
17:38 Phil21 and I think you see that manifest in both's featuresets and focuses
17:39 Phil21 sri: yeah, I agree with your statement for the past decade
17:39 sri heh
17:39 Phil21 kinda the problem :)
17:39 Phil21 it just never quite gets there
17:39 sri have you tried the commercial postgres distributions yet?
17:39 Phil21 it seems something is a release or two away, and then it gets trashed for new shiny
17:39 Phil21 haven't tried, no
17:39 Phil21 well, we tried one
17:40 Phil21 the open source version of it, I forget which
17:40 Phil21 jberger recommended it, he may remember
17:40 Phil21 it was... overly complex, and more or less required using their pre-packaged stuff and support imo
17:40 jberger Citus
17:40 Phil21 I'm sure it would have been much better for experienced pgsql guys though
17:40 Phil21 we came in blind
17:40 sri enterprisedb seems to be where it's at
17:41 Phil21 it's just something that is so simple on the MySQL end - wasn't used to fighting it so hard
17:41 Phil21 like it's an afterthought to app design for me usually - the SQL part is black-box HA to developers
17:42 Phil21 and it's difficult for me to provide that with any degree of confidence right now due to both inexperience and lack of a cohesive clusters methodology/ecosystem in pg
17:42 Phil21 shrug
17:42 Phil21 that's all :)
17:42 sri </rant>
17:43 sri i can't take databases without transactional ddl serious anymore
17:43 Phil21 hehe
17:43 Phil21 yeah, devs vs. ops :)
17:43 sri no matter how nicely they replicate broken state
17:43 Phil21 and I get it, the pg feature list is pretty awesome
17:43 lluad Doesn't the app need to be aware that MM replication is going on with mysql? To avoid non-dterministic updates and so on?
17:43 dod joined #mojo
17:44 Phil21 lluad: depends on the app, but no not with decent operational design in 99% of all cases
17:44 Phil21 I mean, you are correct in the sense I wouldn't operate bank ledgers on mysql's MM replication (or any replication for that matter)
17:45 Phil21 but what we do for the "typical" not-scale-but-HA cluster is A/B pair it
17:45 Phil21 we put haproxy in front of two mysql in M/M, then point the apps to the virtual IP on haproxy
17:45 Phil21 haproxy is then "locked" to one server as the master, 100% of queries it it - but if we failover we can immediately and failback and forth with zero worries
17:45 Phil21 it's the best of both worlds
17:45 lluad Oh, yeah. Just for hot standby rather than "real" MM replication is much easier.
17:46 Phil21 if you need MM replication for scale reasons, the conversation of course shifts quite a bit and gets a lot more nuanced
17:46 Phil21 yeah
17:46 Grinnz mysql's replication is dead to me since I learned it replicates temp tables and there's no option to disable that
17:46 Phil21 Grinnz: hrm, how could it not?
17:46 lluad Postgresql could really do with an out-of-the-box hot standby solution.
17:46 stryx` joined #mojo
17:46 Phil21 I assume you mean ephemeral temp tables
17:46 Grinnz temp tables specific to one connection...
17:46 sri rofl
17:46 Phil21 like how can it replay the query on the other master without making that :)
17:47 Phil21 it's simple and dumb, that's kinda why I like it
17:47 Phil21 just like my women
17:47 Phil21 jberger: if that gets pasted anywhere you're fired
17:47 Grinnz rofl
17:48 ribasushi Phil21: public log... https://irclog.perlgeek.de/mojo/2016-10-25#i_13463367
17:49 Phil21 heheh, well he knows how untrue that statement is but could troll me quite well with it if he wanted to :)
17:49 tyldis WEll Pg makes the case that you rarely really need multi master
17:49 * jberger keeps his mouth shut
17:49 Phil21 tyldis: you're not wrong actually, and this is why I initially agreed to move to pg ("how hard can active/passive HA be")
17:49 Phil21 but.. holy crap it's hard in pg :)
17:49 Phil21 comparatively
17:49 purl comparatively is uninteresting
17:50 Grinnz purl is not amused
17:50 Phil21 just to make a A/B cluster failover cleanly without manual interaction to fail it back again
17:50 Phil21 it's a non-trivial task
17:50 tyldis I always found MySQL replication easy to set up and even easier to break
17:51 Grinnz yeah, ours does break a lot...
17:51 tyldis Constantly out of sync
17:51 suede http://cppcms.com/pics/WPvsCPPCMS.png
17:51 suede sorry whoops
17:51 suede i meant to paste that in telegram not textual -.-
17:51 tyldis So I prefer the simpler aproach that Pg gives us. And it fits our usecase so far
17:52 Phil21 tyldis: what's that? just a single DB and manual failover if shit hits the fan?
17:52 Phil21 I mean, that's super tempting - I almost did it :)
17:52 tyldis Manual? Hot standby :)
17:52 Phil21 this all said.. jberger really did convince me on the superiority of the postgres feature list
17:53 Phil21 I mean, how did you make it work reliably :)
17:53 Phil21 that's what we have setup now as well, and while we have most of the bugs ironed out I still eyeball it suspiciously as I know it's going to bite us in the ass one day
17:53 tyldis :)
17:54 tyldis All writes are handled by a few core services, which makes it all a lot easier
17:54 tyldis As I said, it depends on your usecase
17:55 tyldis But MySQL has served me well and still does. Have 600 million records stuck in one ;)
17:56 Phil21 ah, yeah mysql sorry got confused :)
17:56 Phil21 and yeah, completely agree re: use-case
17:56 Phil21 I mostly rant about pg to annoy jberger these days
17:57 jberger you'll notice its was me that pinged you in
17:57 ribasushi Phil21++ #annoying jberger is a worthy goal
17:57 Phil21 well played
17:58 tyldis ;)
18:00 rshadow joined #mojo
18:10 cfedde disaster planning is a bit like medical insurance.  It seems great till you have to use it.
18:22 tyldis https://github.com/inway/mojo-rabbitmq-client/issues/1#issuecomment-256122019
18:23 tyldis Give the man some love!
18:34 jberger added a few comments, including inviting him here
18:34 jberger I would love to see a mojo rabbitmq binding, I wish I had some time to review it
18:35 jberger I won't at least until after Chicago.pm on thursday
18:44 irqq joined #mojo
18:58 tyldis jberger: Thanks.
19:04 Lee joined #mojo
19:18 kryp joined #mojo
19:22 rshadow joined #mojo
19:42 sri ohoh, the leaked images of the new macbook are scary
19:42 sri there really is no escape key :O
19:43 suede sri: was that a pun
19:43 sri http://www.macrumors.com/2016/10/25/images-of-new-macbook-pro-leaked/
19:43 suede oh no it was literal
19:43 suede why would they DO that
19:43 * sri is 100% serial
19:44 suede who needs a tardis when you have a universal cereal bus
19:44 sri it also confirms the same shallow shitty keys from the 12" macbook :(
19:45 suede :(
19:46 jberger wow
19:47 * lluad cries for the vi users
19:47 jberger with microsoft providing an actual unix environment and apple being actively hostile to vi(m) users I wonder if developers might actually start to migrate hardware
19:47 Lee so macbook keyboards have now gone from "shit" to "shit *and* useless"
19:47 jberger that would have been absolutely inconceivable one year ago
19:48 suede jberger: i would switch to Windows if the unix environment was decent enough
19:48 suede also i don't think the vim crowd is big enough that you'd see a great migration of any kind
19:48 jberger suede: have you tried WSL?
19:48 jberger also, I wouldn't be so sure
19:48 suede you'll have to define that term for me, sorry :(
19:48 jberger windows subsystem for linux
19:48 jberger you can run native ubuntu on windows now
19:48 suede Oh, duh
19:49 suede No i haven't yet
19:49 suede that's pretty cool
19:49 suede It's an actual linux subsystem right, not just a compat layer like Cygwin?
19:50 jberger right
19:50 suede yeah, i'd switch
19:50 jberger something that was built into NT4 but hadn't been exposed until just recently
19:50 jberger and re vim
19:50 jberger http://stackoverflow.com/research/developer-survey-2016#developer-environment-per-occupation
19:50 jberger check the different fields
19:51 jberger then look at the section below, see mac gain bigly over windows
19:51 sri the thought of the oled touchbar was intriguing, but removing the escape key is total madness
19:51 jberger (hehe)
19:51 suede sri: they are making all kinds of crazy decisions lately
19:51 sri they could have just gotten rid of the caps lock
19:51 ribasushi jberger: learning to speak presidentially? :D
19:51 jberger yup
19:51 jberger yuuugge
19:51 suede like removing the headphone jack when it doesn't change anything
19:52 suede on the new iphone
19:52 sri i saw an article comparing tim cook to ballmer, and i guess i have to agree
19:52 jberger suede: they are going to (or probably already are) selling usb c headphones
19:52 ribasushi such a nasty jberger
19:52 tyldis Or introduce DRM
19:52 sri he knows how to squeeze the last cent out of existing products
19:53 jberger ribasushi++
19:54 Lee i'll take the headphone jack over the esc key on a new macbook (as long as the jack still has the optical out function)
19:55 sri apple is now basically ballmer's microsoft
19:59 itaipu joined #mojo
20:07 Grinnz apple removed something else that people actually needed? http://www.troll.me/images/winter-is-coming/brace-yourself-the-memes-are-coming.jpg
20:14 sri yea, the memes are coming https://pbs.twimg.com/tweet_video/Cvo2w1kUEAAiwHG.mp4
20:21 sri https://twitter.com/Gidgerby/status/791011089651789824
20:21 sri hahahahaha
20:21 purl LOLCON 5 reached.
20:21 Grinnz LOL
20:21 suede LOL
20:22 genio oh, no! I can't charge and hit ESC at the same time! :)
20:23 Grinnz and listen to music
20:27 lluad joined #mojo
21:02 lluad joined #mojo
21:40 sri can anyone confirm this? https://github.com/kraih/mojo/issues/1011
21:57 jberger sri: are route conditions on headers expensive?
22:28 Janos joined #mojo
22:51 itaipu joined #mojo
23:17 lluad joined #mojo
23:21 Adura Can't reproduce on Win 7 64bit Strawberry 5.20, sri.

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