| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| | |
|
| | |
|
| |\
| |
| |
| | |
This probably broke stuff
|
| | |
| |
| |
| |
| |
| |
| | |
This field is required to store the progress of the default
resolution of the media. So, we now store the total progress of transcoding
the video in 'transcoding_progress' and progress of transcoding the default
resolution in 'main_transcoding_progress'.
|
| |\ \ |
|
| | | | |
|
| | |/
|/| |
|
| |/
|
|
|
|
| |
In https://bitbucket.org/zzzeek/alembic/issues/419/ we were suggested
use connection instead of engine. This should fix an issue reported
via ml.
|
| |
|
|
|
|
|
|
|
| |
This is a necessary step towards deprecating sqlalchemy-migrate.
* mediagoblin/db/migrations/versions/52bf0ccbedc1_initial_revision.py
(upgrade): Update for adding tables. Mostly generated via Alembic's
autogenerate tools.
(downgrade): Added a comment.
|
| |
|
|
|
|
|
|
| |
This allows us to make use of automatically generated migrations, at
least for core.
* mediagoblin/db/migrations/env.py (target_metadata): Switch to gmg's
Base.metadata.
|
| |
|
|
|
|
|
| |
This is important (probably) for long-running plugin branches.
* mediagoblin/db/migrations/script.py.mako: Updated to upstream
Alembic version of this file.
|
| |
|
|
|
|
|
| |
It seems there was a commit for a while where the migration was
making Report.object_id NOT NULL and this caused an errror when
a report deleted the associated object (media). This migrtion
checks it's nullable and if not, alters it so it is.
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
| |
This shouldn't really effect much but it is a needed change for the future
this changes the Notification.object_id to be the ID of the Comment (the link
table to the comment object) rather than TextComment (the comment object itself).
This is needed as now comments can be other things, other than TextComment.
|
| |
|
|
|
| |
This update changes migrations to wrap db.execute results in a list to
keep the connection from closing for sqlite migrations.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
for alembic
For some reason the alembic connection wasn't seeing any tables that existed
however the existing connection works well and we shouldn't be creating a brand
new one when we have one ready to use. I've used a little bit of a hack due to our
old version of alembic.
After 0.7.5 of alembic it offers a `Config.attributes` dictionary which is designed to
allow you to pass your connection/engine/whatever to the env.py config script so you're
not required to create new ones. As we're on an older version I just create a dictionary
with the same name and use it as otherwise documented. It seems this is the suggested
fix for #5395 and it fixes #5398 too.
|
| |
|
|
|
|
|
| |
When a user deleted media or other objects any notificationn objects or reports
would cause errors as the object doesn't exist anymore. This now removes them or
sets them to None as expected by the code. This also adds some code to the base
deletion code to make sure this happens when an object is deleted.
|
| |
|
|
|
|
| |
Just like we would have in sqlalchemy-migration inspect_table works and
is better than constantly redefining a table in the migration. This switches
the migration to use that method.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
When an item is deleted it should be removed from all collections, this commit
makes that happen. It's got two changes:
1. Adds the code so when an object is soft deleted, it's automatically removed from
all collection items
2. Add a migration to fix this issue for those who have tombstones (Graveyard objects)
in their collections because of this bug.
This commit requires you to run a migration
|
| |
|
|
| |
Use this migration as a parent for yours.
|
| |
|
|
| |
This can be moved to docs/source/devel/migrations.rst.
|
| |
|
