hello, I am using a package which is meant for Django 1.7, however, we use Django 1.6. Hence, I cannot find a way to do migrations. Is there maybe some way the migrations could be run anyway (I know for some packages back portability exists)
elenaoat is now known as lenochka
Quitta has quit
ojii2 joined the channel
djsnoops joined the channel
retailify joined the channel
Mask777 joined the channel
mattmcc
There's no graceful conversion from native migrations to South, no.
sn_ joined the channel
colinbits has quit
Mask7777 has quit
Koterpillar
lenochka: you can try removing migrations from it, it'll be syncdb'ed
lumidify joined the channel
mattmcc
Incidentally, both 1.6 and 1.7 are no longer supported.
lenochka
Koterpillar: hm, how do you mean? all the existing migrations that I have in 1.6?
Koterpillar
lenochka: remove only the migrations *that package* has
lenochka
mattmcc: ok... :/ we really need to upgrade then
Mask777`Work has quit
Koterpillar
however, you will have a mess
lenochka
Koterpillar: ah
TheTemplar joined the channel
Koterpillar
upgrade
lenochka
Koterpillar: maybe I will upgrade instead.. not sure how much work it requires though
Koterpillar
you can try finding an older version of the package
but really, upgrade, old django does not receive security updates
what package?
raininja has quit
mattmcc
The release notes are pretty thorough when it comes to backwards-incompatible changes.
lenochka
Koterpillar: I had it. But I need the newest.. it's pinax-stripe
will just upgrade.
Koterpillar
good
mattmcc
Well, it shouldn't come as a surprise that current versions of Django apps don't work with ancient versions of Django.
lenochka
mattmcc: yes, I know
sn_ has quit
djsnoops
Asked this earlier but got disconnected (weak wifi…). I'm trying to invalidate template fragment caching with the low level API, on a model's post-save signal. the problem I'm having is that my template fragment cache has a number of vary_on's (filters, pagination). how do i invalidate these properly? https://dpaste.de/63aa#L10
Lucifer_arma has quit
retailify has quit
deb has quit
mattmcc
Depending on your backend, you may be able to delete keys using a wildcard.
djsnoops
Wildcard how so?
TheTemplar has quit
onizo has quit
Lucifer_arma joined the channel
saurabhjha has quit
alexgordon has quit
saurabhjha joined the channel
mattmcc
It depends on the backend. Although neither memcached nor redis support it with a built-in operation, it requires some extra steps.
mastokley joined the channel
ilhami joined the channel
ilhami
hey
Gooer_ joined the channel
jladage joined the channel
Lucifer_arma has quit
jladage has quit
djsnoops
mattmcc I'm using memcached. At the moment I'm considering keeping a separate "version" cache key for the specific template and including it in the template fragment cache vary_on, which I can just simply increment the "version" to invalidate the previous cached data. What do you think?
lumidify has quit
Stranger6667 joined the channel
ilhami
502 - Bad Gateway
I have setup something wrong hehe
mattmcc
djsnoops: That could work, but I've gotta wonder how practical it is to be storing all these various permutations of your template.
Gooer_ has quit
Especially since you're only caching for a minute.
Gooer joined the channel
Gooer_ joined the channel
Lucifer_arma joined the channel
djsnoops
mattmcc: It's only at a minute for development testing purposes ;) The various permutations are from the different filters and sorting orders
Gooer_ has quit
Gooer_ joined the channel
mattmcc
Right, but are you sure it's a win?
For example, it wouldn't take many possible choices for the filters before you've got tens of thousands of cache entries, many of which may not be hit more than a couple times before they expire.