<    March 2017    >
Su Mo Tu We Th Fr Sa  
          1  2  3  4  
 5  6  7  8  9 10 11  
12 13 14 15 16 17 18  
19 20 21 22 23 24 25  
26 27 28 29 30 31
00:04 chriswarren joined
00:13 chriswarren joined
00:16 scarfacedeb joined
00:16 nomicflux joined
00:33 MarioBranco joined
00:47 rschmukler joined
00:48 <alisdair> i should reread it now that i actually know how to use ecto
00:57 io_bora joined
01:01 markh joined
01:02 webdev007 joined
01:08 chriswarren joined
01:10 vkushner_ joined
01:11 MarioBranco joined
01:17 chriswarren joined
01:18 nighty- joined
01:22 nomicflux joined
01:31 rschmukler joined
01:36 lixmon joined
01:36 <lixmon> yoyo
01:38 tomterl joined
01:42 rschmukler joined
01:43 miwa joined
01:44 chriswarren joined
01:47 junsuiji1 joined
01:53 chriswarren joined
01:53 montanonic joined
01:55 rschmukler joined
01:56 <wsmoak> Wow, I didn't realize the forum wasn't public. :( Just got a "You need to log in to see that topic."
01:58 <wsmoak> no... I've logged out... and I can still see things...
01:59 montanonic joined
02:01 <wsmoak> but this "Visit Topic" link from my email archives requires a login: https://elixirforum.com/t/what-do-you-think-is-missing-or-needs-improving-in-elixir/2369/13
02:03 <Radar> wsmoak: I see the same problem. You're not crazy.
02:06 <wsmoak> is it _all_ of the email links? or only some of the categories? I'd say I would investigate further and report it but realistically I am not going to have time so if it annoys anyone else feel free to pursue it and thanks!
02:11 TheMoonMaster joined
02:11 chriswarren joined
02:14 Guest_ joined
02:16 MarioBranco joined
02:17 scarfacedeb joined
02:21 chriswarren joined
02:21 Ioyrie joined
02:26 rschmukler joined
02:30 duggiefresh joined
02:35 SouvikB joined
02:35 strixy joined
02:36 lmoura joined
02:36 raycoll joined
02:42 sfbw joined
02:57 griffinb_ joined
03:02 Codecast_ joined
03:07 refriedchicken joined
03:15 chriswarren joined
03:15 Haydos joined
03:15 Haydos joined
03:16 ernie joined
03:21 marcdel joined
03:24 chriswarren joined
03:25 chavito joined
03:25 rschmukler joined
03:32 sfbw joined
03:32 racycle joined
03:38 rschmukler joined
03:42 chavito_ joined
03:44 milad joined
03:45 gvaughn_ joined
03:46 <wwwd> I'm trying to run a singl test which runs from line # 26 to # 29. I know the next test is broken. When I run it with "mix test test/controllers/post_controller_test.exs:26" I get a an error "** (CompileError) test/controllers/post_controller_test.exs:32: undefined function post_path/2". Line # 32 is in the next test in the file. Am I doing something wroing, or is it expected that it will run all tests from test.exs:[line #] on?
03:46 jerel joined
03:51 rschmukler joined
03:52 larshesel joined
04:01 rschmukler joined
04:02 marcdel joined
04:10 carterparks joined
04:12 rschmukler joined
04:16 Duckily joined
04:19 scarfacedeb joined
04:24 BeerLover joined
04:28 rschmukler joined
04:31 bokayio joined
04:42 gvaughn joined
04:44 jkreeftmeijer joined
04:46 chriswarren joined
04:51 Tica2 joined
04:54 charlenopires joined
04:55 chriswarren joined
04:56 NeverDie joined
04:58 jerel joined
05:02 lessless joined
05:04 chriswarren joined
05:07 dec0n joined
05:14 chriswarren joined
05:15 m_m joined
05:23 chriswarren joined
05:25 meandi_2 joined
05:30 jamilbk joined
05:40 griffinb_ joined
05:46 jkreeftmeijer joined
05:52 rschmukler joined
05:55 MasterNayru joined
05:56 cemilowski joined
05:56 jkreeftmeijer joined
05:56 m_m joined
05:58 tomterl joined
06:00 Exagone313 joined
06:06 fernandomm joined
06:08 Reshi joined
06:12 rschmukler joined
06:12 <nyaray> What would be a good place to put a big literal used in testing?
06:19 cemilowski joined
06:20 jkreeftmeijer joined
06:21 <OliverMT> somewhere in test folder
06:21 <OliverMT> so it doesnt go into releases :)
06:24 <nyaray> yeah, but a helper just used for providing data or a private function?
06:25 <OliverMT> there is no more correct way than the other
06:26 <OliverMT> is this integration test data, so to say?
06:26 <OliverMT> I always advocate writing fixture data to a database and use it from there when testing :)
06:26 <OliverMT> unless this is a test BLOB for some parsing you are doing or similar
06:27 chriswarren joined
06:28 <nyaray> yes, it's integration test data, but it's sent to me by a service and I'm going to assume that the json is well-formed, thus allowing myself to skip parsing it.
06:28 <nyaray> doesn't make sense to store this in a database
06:28 <nyaray> thanks for the sound boarding!
06:30 gvaughn joined
06:34 vmoravec joined
06:36 chriswarren joined
06:40 scarfacedeb joined
06:41 griffinb_ joined
06:43 tuacker joined
06:44 scarfacedeb joined
06:45 tuacker1 joined
06:48 rschmukler joined
06:53 flaviodesousa joined
06:55 mark_66 joined
06:55 icanhazbroccoli joined
06:59 mattyw joined
07:01 wsieroci joined
07:03 chriswarren joined
07:08 BeerLover joined
07:09 flaviodesousa joined
07:13 montanonic joined
07:14 josevalim joined
07:17 milad joined
07:17 milad joined
07:17 stephen_m joined
07:19 Guest67__ joined
07:20 cemilowski joined
07:23 chriswarren joined
07:27 cemilowski left
07:29 nd__ joined
07:30 dani0_ joined
07:31 scarfacedeb joined
07:32 chriswarren joined
07:32 wsieroci_ joined
07:33 robinsjdotcom joined
07:35 akeating joined
07:36 refriedchicken joined
07:37 ZippoWeb joined
07:42 chriswarren joined
07:43 jmiven joined
07:50 tuacker joined
07:51 chriswarren joined
07:51 <cmk_zzz> elixirforums without javascript is ugly as but it is so much quicker it is not even funny.
07:53 <cmk_zzz> imagined how fast it would be without the inlined javascript
07:54 <cmk_zzz> too bad about the mailing list. That was even faster :/
08:00 chriswarren joined
08:06 MaSven joined
08:07 <micmus> cmk_zzz: I'm using elixir form primarily as a mailing list - i.e. I read all messages as email, only for writing go to theweb UI
08:07 <micmus> works quite well for me
08:09 <cmk_zzz> micmus: So if you create an account you can get all posts mailed to you?
08:09 <micmus> yup
08:10 <cmk_zzz> Sounds good. To some degree I'm not sure javascript heavy sites actually add any benefits. At least not as a consumer of content
08:11 sevenseacat joined
08:13 marr joined
08:16 scarfacedeb joined
08:16 Ghouli joined
08:24 <cmk_zzz> /win2
08:24 <cmk_zzz> fail
08:24 shadowjack joined
08:29 mattyw joined
08:29 <OliverMT> discourse is pretty shitty for being so popular
08:29 <OliverMT> I wonder if Firestorm will be any success at all
08:29 <OliverMT> it looks slightly stalled
08:36 scarfacedeb joined
08:36 rschmukler joined
08:41 griffinb_ joined
08:42 hbraun joined
08:42 <Nicd-> what forum is not pretty shitty? :P discourse is one of the better that I have tried
08:42 <Nicd-> I still run an SMF 2.1 beta forum… now that's a pain
08:43 colrack joined
08:44 inoas joined
08:47 robinsjdotcom joined
08:51 voltone joined
08:53 gvaughn joined
08:54 chriswarren joined
08:56 jkreeftmeijer joined
09:04 chriswarren joined
09:04 akeating joined
09:05 jkreeftmeijer joined
09:16 akeating joined
09:23 lexmag joined
09:26 BeerLover joined
09:31 rschmukler joined
09:32 akeating joined
09:32 <OliverMT> I run an SMF one too lol
09:32 <OliverMT> its private, does discourse cover that-
09:32 <OliverMT> _
09:32 <OliverMT> ?
09:35 <Nicd-> never administrated discourse but I would assume?
09:35 cemilowski joined
09:35 <Nicd-> kind of a basic feature
09:38 vmoravec joined
09:40 NeverDie_ joined
09:40 chriswarren joined
09:41 cemilowski left
09:41 NeverDie joined
09:43 BeerLover joined
09:47 nd__ joined
10:08 <celyr> Idk where all this hate about SMF come from but it used to be pretty straightforward
10:15 inoas joined
10:16 gmcabrita joined
10:16 chriswarren joined
10:17 NeverDie joined
10:19 renl joined
10:21 <renl> hi for phoenix 1.3 generators it doesnt create the belongs_to schema when we do referencing anymore?
10:27 nd__ joined
10:35 chriswarren joined
10:41 gvaughn joined
10:42 proteusguy joined
10:42 griffinb_ joined
10:43 Siyfion joined
10:45 griffin__ joined
10:48 m1dnight_ joined
10:53 m1dnight1 joined
10:55 jkreeftmeijer joined
10:56 DeadTrickster joined
11:05 BeerLover joined
11:09 robinsjdotcom joined
11:10 cristianR_ joined
11:11 chriswarren joined
11:12 robinsjd_ joined
11:13 vendethiel joined
11:15 cemilowski joined
11:19 rschmukler joined
11:20 chriswarren joined
11:23 scarfacedeb joined
11:23 dani0_ joined
11:25 jkreeftmeijer joined
11:32 dimitarvp joined
11:39 BeerLover joined
11:45 milad joined
11:46 sfbw joined
11:49 Guest____ joined
11:52 lessless joined
11:52 lessless joined
11:52 Guest____ joined
11:53 <nyaray> I thought the forum-ness attribute was inversely proportional to the shitiness of a site
11:53 <nyaray> s/attribute/property
12:00 pHcF joined
12:05 chriswarren joined
12:08 milad joined
12:14 rschmukler joined
12:14 chriswarren joined
12:15 nomicflux joined
12:19 wtetzner joined
12:24 chriswarren joined
12:27 dani0_ joined
12:30 griffinb_ joined
12:30 gvaughn joined
12:31 griffinb_ joined
12:31 griffinb_ joined
12:32 griffinb_ joined
12:33 chriswarren joined
12:36 nighty- joined
12:42 chriswarren joined
12:51 chriswarren joined
12:51 robinsjdotcom joined
12:53 dani0_ joined
12:57 ramortegui joined
12:57 rpip_ joined
13:00 chriswarren joined
13:01 Or1on joined
13:06 maxbeizer joined
13:06 ikcerog joined
13:16 Guest67_ joined
13:26 jkreeftmeijer joined
13:27 cemilowski joined
13:33 jeregrine joined
13:35 sfbw joined
13:43 rschmukler joined
13:46 chriswarren joined
13:46 chriswarren joined
13:49 jerel joined
13:55 chriswarren joined
13:56 cemilowski left
13:57 gvaughn joined
14:03 cdg joined
14:04 chriswarren joined
14:07 cdg joined
14:12 meandi joined
14:13 chriswarren joined
14:14 chriswar_ joined
14:15 <vans163_> when I have 2 modules in the same file. if the top module imports the bottom module. theres an error. Is there a way to tell the top module the other module is there, its just below it in the file?
14:17 chriswarren joined
14:18 rschmukler joined
14:19 <ciawal> put them the other way around? :P
14:22 <CvRoques1> If I understand correctly (I have not yet read the source), the compiler just starts compiling all files in parallel and when it blocks on a dependency, it waits until the dependency is available/has been compiled. When this runs into a deadlock (cyclic dependency or your case) it bombs out and you have to fix it. So the simplest way is to split the source file into two.
14:22 <vans163_> gotcha
14:22 rushsteve1 joined
14:30 dani0_ joined
14:32 ikcerog joined
14:33 racycle joined
14:34 griffinb_ joined
14:35 scarfacedeb joined
14:37 ekinmur joined
14:38 scarfacedeb joined
14:38 <InternetFriend> c
14:38 Codecast_ joined
14:40 raycoll joined
14:41 gvaughn_ joined
14:42 rushsteve12 joined
14:44 wsieroci joined
14:45 jgpawletko joined
14:51 scarfacedeb joined
15:00 meh` joined
15:00 tuacker joined
15:05 cdg_ joined
15:13 rschmukler joined
15:14 scarfacedeb left
15:14 wsieroci joined
15:14 wsieroci joined
15:22 raycoll joined
15:22 milmazz joined
15:23 rschmukler joined
15:27 __vy joined
15:28 chavito joined
15:29 <__vy> there we go.. cancelled my netflix subscription for an irccloud subscription
15:32 <adam12> I hate watching IRC. Never anything good on.
15:35 griffinb_ joined
15:35 larshesel joined
15:35 <__vy> the resolution is great though
15:38 amclain joined
15:38 montanonic joined
15:41 Havvy joined
15:53 hxegon joined
15:58 racycle joined
16:03 cemilowski joined
16:08 cemilowski left
16:08 akeating joined
16:11 chavito_ joined
16:15 m1dnight_ joined
16:16 chriswarren joined
16:16 chriswarren joined
16:17 chriswarren joined
16:18 chriswarren joined
16:19 chriswarren joined
16:20 squallstter joined
16:20 akeating joined
16:23 raycoll joined
16:24 kus_ubuntui686 joined
16:25 InternetFriend joined
16:27 mrogne joined
16:30 robinsjdotcom joined
16:31 dani0_ joined
16:33 chrisml joined
16:35 griffinb_ joined
16:36 <rozap> argh
16:36 <rozap> so i have two stages in gen-stage
16:37 <rozap> and the producer is emitting a list of lists of binaries
16:37 <rozap> and when i measure binary usage in the producer, like this
16:38 <rozap> (memory usage) https://gist.github.com/rozap/a1e170f1058d7e528d0b95b3570c8dee
16:38 <rozap> it says 25mb
16:38 <rozap> and then in the next stage, before anything has happened at all, in the handle_events function, i print the binary usage, and it's 8gb
16:38 <rozap> O_O
16:39 <rozap> anyone have any ideas?
16:39 <rozap> i feel like i'm losing my mind
16:39 asabil_ joined
16:40 <rozap> (and indeed, observer is saying the app is asking for like 20gb) (and it keeps OOMing in prod)
16:40 <rozap> well not prod, we haven't shipped it yet, but there's something crazy going on..
16:40 kau joined
16:43 <micmus> rozap: what's the size of the data you're sending?
16:44 <rozap> i'm benchmarking with a test dataset which is filling with many small binaries, like "value 0" "value 1" etc
16:45 jkreeftmeijer joined
16:45 <rozap> the producer is chunking the input stream (from a network socket) into rows and columns and sending them downstream to a consumer
16:45 <micmus> I'm mostly asking about the real size of the data, so it's possible to check what overhead there is
16:46 <rozap> the physical size of the *whole* input file is only 153mb
16:46 <rozap> but the first chunk of roughly 25mbv expands to 8gb, somehow
16:46 <rozap> when it gets passed in between genstages
16:46 marr joined
16:48 Duckily joined
16:48 ikcerog joined
16:50 nd__ joined
16:54 nighty- joined
16:56 Or1on joined
16:57 DeadTrickster joined
16:58 josevalim joined
16:59 <asonge> rozap: each message copies each binary <64B
17:00 <rozap> but genstage is batching the rows into a single message, so i'd only expect like 2x usage
17:01 <rozap> i did find a bug in that gist i sent - still fiddling with it to get a more accurate reading
17:03 montanonic joined
17:09 wsieroci joined
17:09 rschmukler joined
17:11 jeffweiss joined
17:14 jparise joined
17:15 refriedchicken joined
17:16 Duckily joined
17:17 rschmukler joined
17:20 chriswarren joined
17:21 InternetFriend joined
17:23 racycle joined
17:27 <jparise> anyone have an idea when Elixir 1.4.2 packages will be available on erlang-solutions.com?
17:29 Guest joined
17:35 tolerablyjake joined
17:42 montanonic joined
17:57 Duckily joined
18:00 ikcerog joined
18:00 Duckily joined
18:03 codestorm7 joined
18:04 Or1on joined
18:17 <josevalim> jparise: please ping them :) on twitter is usually enough
18:21 chriswarren joined
18:28 rschmukler joined
18:31 lexmag joined
18:34 s_kilk joined
18:43 drincruz joined
18:44 jesuis joined
18:44 <jesuis> maybe someone can help me out with (I think) phoenix_live_reload
18:45 <jesuis> is I reload a page it's pretty fast, but when I change a file, phoenix_live_reload kicks in and it takes around 10secs to load a page in Chrome :(
18:45 <jesuis> also in FF btw
18:49 wsieroci joined
18:49 <tuacker> jesuis: what file did you change? and what is the output you get in your iex (the 'compiling files' one)
18:50 rozap joined
18:52 <jesuis> just a /show.html.eex for example
18:52 <jesuis> Compiling 56 files (.ex)
18:52 <jesuis> ha, there is it :)
18:53 <jesuis> it's compiling 56 files when I touch one eex
18:53 <micmus> that shouldn't be happening normally
18:54 IRCFrEAK joined
18:54 <micmus> do you import the view that uses this template into some other modules?
18:55 IRCFrEAK left
18:55 <jesuis> nope, pretty standard
18:57 <tuacker> so I had this same problem a few days ago, the way I solved is was do `mix xref graph --format dot` in your project, then open the resulting xref_graph.dot file in a text editor and check where your view (ex user_view.ex) is on the right side of an "->"
18:57 <asonge> you're saving or touching a file that's causing a lot of other things to compile
18:57 <asonge> ^^^ that.
18:57 <tuacker> meaning "this" -> (depends on) user_view
18:57 <tuacker> found my issue real quick with that
18:58 HendrikPeter joined
18:59 <jesuis> I think I found it, you pointed me in the right direction
18:59 kyrix joined
18:59 maxxe joined
18:59 <maxxe> test
18:59 <HendrikPeter> pong back ;)
19:00 <jesuis> Compiling 1 file (.ex)
19:00 <jesuis> ha, I was using a module for rendering a page title which was importing shizzel
19:01 chriswarren joined
19:02 <jesuis> bam! from 10 sec's down to 600ms (still a bit slow), but know where to look now()
19:02 chriswar_ joined
19:02 <jesuis> thanks guys
19:02 Tallys__ joined
19:03 <HendrikPeter> sounds nice, whatcha building?
19:06 InternetFriend joined
19:08 IRCFrEAK joined
19:08 mattyw joined
19:09 <asonge> jesuis: referencing modules doesn't usually do that much on their own, but if you start `use`-ing stuff in those modules, that can bring in a few compile dependencies that can slow down compilation
19:10 IRCFrEAK left
19:10 dani0_ joined
19:14 m1dnight_ joined
19:20 wsieroci joined
19:23 <OliverMT> ericmj: does postgrex re-use prepared statements (query cache) ?
19:24 <ericmj> there isn't really a query cache in the driver
19:24 <ericmj> the database handles that if you prepare queries
19:25 <OliverMT> ah ok, so the answer is yes, but not like I ask
19:25 <OliverMT> :p
19:25 mattyw joined
19:25 rschmukler joined
19:26 GOLDPANK joined
19:26 <ericmj> yes :)
19:27 GOLDPANK left
19:35 m_m joined
19:36 rschmukler joined
19:37 griffinb_ joined
19:40 craigger joined
19:41 sfbw joined
19:41 dani0_ joined
19:41 bspatafora joined
19:41 _I_R_C_F_R_E_A_K joined
19:45 <bspatafora> I'm trying to migrate up to a new Ecto migration. The new migration gets picked up when I run `MIX_ENV=[env] mix ecto.migrate` when MIX_ENV is "dev" or "test", but is totally ignored when the env is "prod". All the other migrations run except the new one in prod. Any ideas?
19:46 jparise joined
19:48 icanhazbroccoli joined
19:50 <barttenbrinke> Did you commit it?
19:50 _I_R_C_F_R_E_A_K left
19:50 <barttenbrinke> Is there a way to convert the string part of this query to something that ecto understands :)?
19:51 <barttenbrinke> https://www.irccloud.com/pastebin/tUw3xh9e/
19:51 <bspatafora> barttenbrinke: Yes
19:52 cdg joined
19:52 <micmus> bspatafora: is there a chance the migration was already applied on prod? (migrations are identified by the number in the begging of the file name)
19:53 sfbw joined
19:53 <bspatafora> micmus: Nope. New app, so I was able to drop the prod database and try again fresh. Migrates everything but the last one. Very strange.
19:54 <micmus> barttenbrinke: from r in Shield.Role, join: p in assoc(r, :permissions), where: not p.id in r.deactivated_parent_permissions, preload: :permissions
19:54 <micmus> you can translate to the pipeline syntax, if you want, but I prefer this one :P
19:57 hashpuppy joined
19:58 milad joined
19:58 <barttenbrinke> @micmus Super! I want to pipe all the things :)
20:01 <barttenbrinke> micmus: I know out of 10 Elixir developers, 9 prefer the piped version. So maybe the documentation needs to change on that part.
20:02 <micmus> it seems they haven't been doing much joins :P
20:02 <micmus> with a lot of joins, the piped syntax becomes really hard to read
20:02 <micmus> (IMO)
20:02 <barttenbrinke> I guess they dislike Macros :)
20:02 <micmus> both are macros ;)
20:03 <barttenbrinke> Yes but the piped version looks like real code :)
20:03 drincruz joined
20:03 <barttenbrinke> Hmm, doe Ecto have a .to_query?
20:03 <barttenbrinke> does
20:04 <micmus> Ecto.Adapters.SQL.to_query
20:04 ikcerog joined
20:04 <micmus> uhg, to_sql
20:05 <micmus> barttenbrinke: were you asking about how to turn into sql? Or maybe something else?
20:05 <barttenbrinke> Thanks :) The query generaeted does not play well with the database :)
20:07 dpehrson joined
20:08 <dpehrson> Is there a way to make confirm's distillery release plugin use the distillery --env rather than MIX_ENV to figure out which `app.[env].conf` file to use?
20:08 <dpehrson> As far as I can tell doing `MIX_ENV=prod mix release --env=staging` will result in it using `app.prod.conf` still.
20:09 <dpehrson> (I don't have a staging MIX_ENV, it's simply a release environment with a few different config settings related to hostnames, etc.)
20:09 milad joined
20:09 drincruz joined
20:09 <bspatafora> micmus: Do you know by chance if there is any special logic around migrating when MIX_ENV is prod? Or where I might look to find such logic?
20:10 tvaalen joined
20:10 <micmus> bspatafora: no, it should all be the same. If you generate a new migration, does it behave that way as well?
20:10 <tuacker> bspatafora: is your migration file actually present on wherever you run the prod migration?
20:11 <bspatafora> micmus: Yes, I've tried recreating it with a different name. Or do you mean creating a new, new migration on top of the new one, to see what happens?
20:12 <bspatafora> tuacker: It is
20:12 <micmus> bspatafora: check the output of mix ecto.migrations
20:12 <micmus> this should list what ecto thinks is the status of each migration
20:12 <barttenbrinke> micmus: The Rails version of this is unreadable and the Ecto version does not work.. time to quit and refactor tomorrow :)
20:13 cemilowski joined
20:13 wsieroci joined
20:13 <micmus> barttenbrinke: anything more than "does not work"? :P
20:13 <bspatafora> micmus: List is as it should be in dev and test. But in prod, it does not even list the new migration, only the existing ones.
20:16 <barttenbrinke> https://www.irccloud.com/pastebin/SKe5KLD1/
20:16 <barttenbrinke> micmus ^^
20:17 j0nii joined
20:17 codestorm7 joined
20:17 <bspatafora> micmus: Probably a dumb question, but any chance the length of the migration name would matter? Oddly enough, I remember this being a problem in a different system in the past, but I think it was ActiveRecord
20:17 montanonic joined
20:17 <barttenbrinke> The NOT IN is different, and so it the roles aliassing in the last part.. which is weird
20:17 <micmus> barttenbrinke: I'd assume the WHERE (false) part is problematic - it's going to reject all rows :P
20:18 <tuacker> bspatafora: this shouldn't be it because that would make it fail in dev/test too..but what is the file ext of the migration? (should be .exs)
20:18 <bspatafora> tuacker: Is .exs
20:18 <bspatafora> Generated with echo
20:18 <bspatafora> echo*
20:18 <barttenbrinke> micmus: That is ont thing, but MySQL does not like the NOT id = ANY part
20:19 <barttenbrinke> micmus: I am not 100% that is actually valid mysql.. I expected it to be a NOT IN.
20:19 <barttenbrinke> micmus: And I don't understand where the WHERE FALSE is coming from
20:20 <micmus> barttenbrinke: x in ^[] will generate false
20:20 <micmus> (the empty list is important)
20:20 <barttenbrinke> Ah :)
20:20 <barttenbrinke> Well that is one part :)
20:20 <tuacker> bspatafora: final idea I have: did you push to prod with a release, are you sure prod is actually running the new code and not mistakenly a previous version of it?
20:21 m_m joined
20:24 <barttenbrinke> micmus: Yeah, the not in is translated incorrectly.
20:24 <barttenbrinke> ((p1.`id` NOT IN (r0.`deactivated_parent_permissions`))) vs (NOT (p1.`id` = ANY(r0.`deactivated_parent_permissions`)));
20:25 <barttenbrinke> Mysql throws a syntax error on that
20:27 <barttenbrinke> Let see if I can pass raw sql to Ecto.Query
20:27 <bspatafora> tuacker: You brilliant motherlover. That was it. I'm doing all this on prod, and I thought that running the ecto commands manually would use the files I'd changed. But it was using the files in my last release, I guess, and building a new one did it. I actually don't understand why this is, but problem solved and lesson learned. Thanks!!
20:28 jhulten joined
20:28 chriswarren joined
20:29 chriswa__ joined
20:29 arpunk joined
20:29 nahtnam joined
20:30 <micmus> And I have such a beautiful command for you :P
20:30 <micmus> diff <(mix ecto.migrations | sed -n -e '/---*/,/^$/p' | sed -e '1 d' -e '$ d' | awk '{ print $2"_"$3".exs" }' | sort) <(ls priv/repo/migrations)
20:32 <micmus> barttenbrinke: we should never generate invalid SQL, so this looks like a bug
20:32 <barttenbrinke> micmus: Yes I think it is a bug
20:32 <barttenbrinke> micmus: Though I have to admit the rails version is also super dodgy :)
20:33 <barttenbrinke> micmus: I've used fragment as a workaround for now.
20:34 <barttenbrinke> micmus: I'll try to file a bug on Ecto Tomorrow. Thanks for your help!
20:35 <bspatafora> micmus: Thanks for your help!
20:39 griffinb_ joined
20:48 jkreeftmeijer joined
20:49 markh joined
20:57 cemilowski left
20:59 xsmalbil joined
21:00 loglaunch joined
21:01 soud joined
21:02 rschmukler joined
21:06 montanonic joined
21:08 strykerkkd joined
21:08 stryker_kkd joined
21:09 luke_nukem joined
21:13 milad joined
21:14 PaReeOhNos joined
21:20 PaReeOhNos joined
21:22 tomku joined
21:24 balo joined
21:26 dimitarvp` joined
21:31 jkreeftmeijer joined
21:34 m_m joined
21:37 wsieroci joined
21:39 jkreeftmeijer joined
21:46 mitchellvanw joined
21:46 wtetzner joined
21:52 nd__ joined
21:53 fastjames joined
21:53 zv joined
21:56 bontaq joined
21:58 ikcerog_ joined
22:00 ikcerog_ joined
22:01 <bontaq> hiya, I'm wondering if I'm doing something dumb: I'm curling an endpoint w/ posting some json, and when a field has a ":" in it, poison errors out decoding it? like { "html_url": "http://" } will throw malformed request, but { "html_url": "stuff!" } is fine
22:07 ZippoWeb joined
22:09 rschmukler joined
22:16 io_bora joined
22:20 rschmukler joined
22:24 <bontaq> figured it out, just me being doofy and not encoding properly in the curl
22:34 rschmukler joined
22:41 jkreeftmeijer joined
22:45 rschmukler joined
22:45 <mloy> Hi, has anybody here tried writing Erlang/Elixir NIFs in Haskell (compiled to C)?
22:45 rozap joined
22:46 mcspud joined
22:49 Haydos joined
22:49 Haydos joined
22:50 ZippoWeb joined
22:59 markh joined
23:17 rschmukler joined
23:23 markh left
23:26 arkangel joined
23:26 DLSteve joined
23:28 aeliton joined
23:29 travis-ci joined
23:29 <travis-ci> elixir-lang/elixir#14111 (master - a0b6612 : Aleksei Magusev): The build passed.
23:29 <travis-ci> Change view : https://github.com/elixir-lang/elixir/compare/ddd92110a5eb...a0b6612c71d6
23:29 <travis-ci> Build details : https://travis-ci.org/elixir-lang/elixir/builds/216561204
23:29 travis-ci left
23:31 PaReeOhNos joined
23:36 PaReeOhNos joined
23:36 io_bora joined
23:39 griffinb_ joined
23:42 io_bora joined
23:42 gvaughn joined
23:47 wtetzner joined
23:48 smt joined
23:49 gvaughn_ joined
23:51 Or1on joined