| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
This was missed in dbdbe6a when ordering of moderator notes was changed
on the model.
|
|
|
|
|
|
|
|
This way we can avoid potential duplicate-postings on a refresh, and it
makes tracking the code a bit more clear.
|
|
Indentation missed in a12e1b97 made it try that. Instead, it should fall
through and render the form, which will then show the proper error
message.
|
|
|
|
|
|
|
|
We can't verify it anymore, and this blocks anybody from saving a blog
that has twitteruser on it.
Twitter does not provide any way to verify the existance of a user
anymore, other than paying $$$$. So we'll just cut the field for now, to
make things work again.
This does not remove *existing* registrations of twitter users, but it
does make it impossible to register new ones.
Issues reported by several different people.
|
|
|
|
Commit c9f68d5f upgraded the site to bootstrap 4, amongst other things,
but completely missed that "labels" are now called "badges", and thus
need new CSS classes.
|
|
|
|
For debugging, it's helpful to know what the old value was of a given
blogurl, to make sure we are updating things correctly and only what
should be getting updated is getting updated.
|
|
Previously, we set blogurl if the prior blogurl was set to empty-string
('') but that change wasn't picked up as part of what needed to be
updated, so instead always use new_blogurl if we need to set the blogurl
as this ensures that we'll realize there was a change and will update
the database accordingly.
At the same time, also send an email in every case that the blogurl is
changed, even when it's being changed from being empty. This is perhaps
a bit noisy but it's been broken for so long that it's worth it to make
sure people are aware that it's now working and folks don't register new
blogs over and over again anyway.
Lastly, move the block that does the update and sets the new blogurl
value up higher so that if a new blogurl is set then the emails that are
generated from that run will use that new blogurl instead of the old
one (which could be confusing).
|
|
|
|
|
|
|
|
|
|
|
|
|
|
This way we'll purge the redirect whenever a post is hidden, so we'll
stop redirecting to it.
|
|
retreiving -> retrieving
|
|
|
|
Spotted by Jonathan Katz
|
|
Unfortunately, 6c44b60 very clearly de-optimized the query when running
on a system with millions of log-rows. Oops. So force the django ORM to
put the filter on the join clause thereby making it possible to use an
index, which should make it actually optimized.
|
|
In admin view we are looking at others people's blogs, so it is
confusing to have the bottom part of the page reflect the administrators
personal setup.
|
|
Basically make django generate the joins/prefetches/subqueries instead
of firing of hundreds of queries on the db. Probably doesn't actually
matter that much given how seldom these queries are run, but still.
|
|
For the regular user view, just show which account and email is used for
info.
For the administration view, show the username and email for each
individual blog, to help with tracking down misbehaving blogs.
And finally, in the moderation view, also include the email (it already
had the account name) to make it easier to contact an author.
|
|
|
|
In particular it was terribly labeled in the moderation interface. Make
it better and more consistent between the two interfaces.
|
|
This includes receiving updates using the new push api.
|
|
In particular, it is interesting to know when a blog that has failed
multiple times has actually succeeded more recent than the most recent
failure. So show this both on the dashbaord and on the individual blog
details, and include it in the email that's sent when a blog generates
large amounts of errors.
|
|
|
|
|
|
Mostly whitespace fixes, but also some bare-exception fixes.
Add a setup.cfg that specifies which pep8 settings we normally run with.
|
|
Remove unused imports, and stop using "import *"
|
|
|
|
Finally time to clean this repo up
|
|
Spotted by Guillaume Lelarge
|
|
|
|
|
|
|
|
|
|
|
|
|
|
We're going to need to make small modifications, and the new 6.x release
isn't ready and doesn't appear to be close.
|
|
|
