#django

mattmcc, okay, I'll try to refactor my code. thanks.

during a django test, all the data are rolled back? I'm, anything will be save in the test database ?

I mean*

I’m trying to add the current user to a many-to-many when they create an instance through the admin interface, but I don’t manage. Trying to do it like this: https://dpaste.de/iJEc

The idea is that users can only see projects where they are in the owners list - but it’s happened a few times that they forget to add themselves when creating a project and so can’t access it without the help of an admin. (Yeah, I know, but they want it to work this way)

Ah. save_related is what I should be looking for, probably.

bjornkri: you don't actually save anything, me thinks

after you add the user, i mean

Nope, that’s why I’m adding the user *after* saving

Can’t do it before the object has an id

bjornkri: there is a code snippet in the docs that's almost what you need. why don't you use that?

bjornkri: sure you can

Are you referring to https://docs.djangoproject.com/en/1.11/ref/cont... ?

hop_: Note I’m doing many-to-many, not foreign key

well, i don't know what owners.add() does, but i bet it doesn't persist immediately

It’s many to many. That’s how many to many works

bjornkri: ah, that add

well, guess what: "Using add() with a many-to-many relationship, however, will not call any save() methods"

sorry, missing the rest of the sentence

the gist is: you are right, i don't know what i'm talking about

Django raises an error if you do. Note that I’m working with *save_model* on the ModelAdmin, so it’s behaving differently.

Thanks anyway ;)

sorry for the sarcasm.

This did the trick: https://dpaste.de/kmqe

hi guys! I'm trying to change my db backend from sqlite to postgresql but on migrate I get : LINE 1: ...MN "reminder_offset" TYPE date USING "reminder_offset"::date

any ideas how to fix this ?

can I just squash every migration ever and make a schema from the current models and use that ?

cuci: sure you can reset your migrations as long as you don't need another db to be migrated from a passed state

cuci: if you have data you want to keep you have to split the migration

hop_: no data I wanna keep

then truncate the table before migrating, i think

the project changed very much since the initial state and many of the migrations are just trying out stuff till I got it done

well, sure you can squash anyway, but altering the column still won't work

with data in it

I shouldn't have any reminder_offsets

hm

Is there a manage.py create_database_from_current_models ?

so I can use just that ?

throw away all migrations and start from scratch?

yup

I feel bold

keeping migrations clean is ocd work, don't worry

anyone here deploying django as a container?

Hello! it's not strictly django, but I'm trying to deploy a django app to my small webserver using this guide http://michal.karzynski.pl/blog/2013/06/09/djan... , but supervisor fails. I've tried changing things in the supervisor config file but it still fails. any ideas? https://dpaste.de/sxpQ

vimes: "fails" tells us very little - dpaste traceback etc. please

...sorry my window truncated that!

My problem is that I didn't find any more logs, but now I tried to run the gunicorn script manually and got an error. I can paste the gunicorn script and error

vimes: that would be a good start - if you can't start it manually, chances are supervisor can't start it either

sounds like your sourcing of your virtualenv isn't right

I think the name of the sock was wrong, I changed it and got a different error now. https://dpaste.de/5UDi

seems like I'm missing secret key (this is just to test on server, it will be shut down once it works)

mmm but my settings has a secret key, the one that was generated when I created the project

probably its not reading your settings at all, and its just erroring hard on that specific [required] one

yes, my config says "DJANGO_SETTINGS_MODULE=fossekalltek.settings" but my secret key is in fossekall/settings/base.py. Can I just change it to "production.py" (that gets all the data from base.py) or do I have to set the folder too?

why is the reverse importet in the method definition here?: is https://dpaste.de/5bqj and why is the id coerced to a str?

why is reverse imported there? no idea. probably for no reason. It's unlikely to be avoiding a circular import. Why coerce to str? No idea, its probably not necessary?

[usually imports live outside the top of a file to avoid circular imports, usually because of package boundary crossing]

kezabelle, can you give an example of how you use get_absolute_path in your models?

hm.. I have all imports at the top. :)

lol wait, that example is from the docs?

yes: https://docs.djangoproject.com/en/1.11/ref/mode...