#django

isn't that better?

kartik: Better than what?

vlt: I'm storing data in variables because valid.capacity didn't work. Better than putting them in context?

vlt: *better than simply passing them

vlt: as individual variables

__marco: Probably want to put it in quotes.

kartik: That’s totally up to you.

mattmcc, sure

Hey is there a way to use the striptags template filter but keep line breaks?

Any reason why my logger doesn’t catch errors inside of a celery delay method? If I make the method synchronous it catches it, but not when it’s a @shared_task

maks25: celery and logging -> fun ;)

maks25: check out CELERY_HIJACK_ROOT_LOGGER or similar

iirc there's all kinds of faff involved

yeah celery default logging is fucked

you can unfuck it but it's a little bit of magic incantations

ugh, I was hoping it’d just sorta work :(

this can probably do the trick

thanks hylje I’ll try it out

why is it fucked?

celery does its own logging config that may not be what you want

hylje: I’m getting no where…all I want is to wite the stack trace for level warning and above for celery tasks.

write to log file*

Hi all, if I want to alter my DRF result with a function (eg using GroupBy)- do I do it in views or serializers? And do I have to use viewSets or something else?

hi all! I'm trying to upgrade django to 1.9 and faced a problem. The problem stems from django's system check framefork and the fact that it does import urlconf on startup. This all happens during(or even before) apps registry is populated with models. And in views we heavily importing models

right now I'm not sure how to proceed further: either disable all url-related checks or move models imports from views to ... - there's no answer where I could move those, and I'm very against doing all imports lazily, either inside views or deferring imports to actual usage time

omg I’m such an idiot, no wonder it wasnt logging. Celery wasnt running on my dev machine

hi all

can anyone tell me

if i've got a Foo.objects.all() in memory and im' doing something with it (ie. paginating over records etc..)

and then another process somewhere adds another record to that Model...

what happens to the collection i've got going...?

what seems to be happening is that it...gets reset or something?

does anyone have any experience with this?

its the only thing i can think of which explains my scenario

and i can't find anything online about it

what do you mean it gets reset?

unit923, depending on where that new model in results set. if queryset already has put that into it's cache or if by some reason new object somewhere in the beginning of your results because of ordering - you won't see new object

hmm ok, not seeing the new object is totally fine...

wait.. do you know what would happen if one of the ones in that set (which i'm paginating over) is *updated*?

querysets are lazy. when you first time access results from it - all set is cached in memory (or one-by-one if you were using `.iterator()`). if there's cache - no query to db takes place on subsequent accesses

what would happen to what?

to the whole set

unit923, nothing would happen if the object is already in memory

so i'm doing for foo in Paginator(Foo.objects.all(), batch_size)

in a management command

and quite regularly it just completely stops processing the loop, with no error given...

unit923, https://docs.djangoproject.com/en/1.11/ref/mode...

however, the website part of the project is regularly adding/updating records to the model which is being paginated over

so i was wondering if somehow this was the cause of the Pagination just..."stopping"...

unit923, maybe all memory exhausted by the operation? and it just suffering from resources, which you observe as "stops processing the loop"

but it continues to the rest of the management command, it doesnt crash or throw memory exceptions

aaah, then there's just no results

and they're in small batches of 20 records

ahh ok..

how is that possible though

because it seems to fail in the middle of the batch

I've faced similar problem recently. But I've beed doing filter, and changing data in database while filtering goes. that did result in leaps in processing

try adding some stable order, like `order_by('id')` and verify you don't change data you're filtering against

hey thats a good idea...

ok, so i've got heaps of debug output...

it says "total 46"

batch size of 20

Progress: 20 out of 46