Custom Query (1173 matches)

Filters
 
Or
 
  
 
Columns

Show under each result:


Results (148 - 150 of 1173)

Ticket Resolution Summary Owner Reporter
#5414 fixed Login-validator arbitrary and capricious mi
Description

Using gmg adduser I created an account for myself named "mi". However, when I tried to login using it, I was told, the login-field must be between 3 and 30 characters long.

This is wrong on many levels: . The authenticator should not be verifying the length of submitted login and password at all. Such checks might be appropriate for a new account-creation, but I was logging-in, not creating account. . Even if you disagree with the above, gmg adduser should've rejected the name as too short instead. . Not only is the lower limit of 3 too high -- a family installation can easily have one-letter accounts, the upper limit of 30 is too low as well. Though over 30 is unusual for an account-name, an e-mail address can easily exceed 30 characters.

I was able to login using my e-mail address instead of username, but this needs fixing (along with tests/test_auth.py).

#5413 fixed Make models on clean database with alembic, not from models.py Jessica Tallon
Description

Currently, when there is a new database with nothing that has been done, we skip all the sqlalchemy-migrate migrations and just make the models based off models.py, we then run the alembic migrations. This actually causes problems when we start doing structural changes in alembic as they have already been done in models.py so when the alembic migrations run, they error.

One solution could be to skip all the alembic migrations, however, I think a better solution is to not make any models off models.py and instead make the initial alembic migration create all the models in the state they're in right now and use alembic migrations to change them in the future, removing all of the code to construct models from model.py.

#5412 invalid Add "Accept-Ranges: bytes" response header and support for videos Mike Swierczek
Description

For streaming large (greater than 1GB) WebM videos, support for the "Accept-Ranges: bytes" header indicates to the browser that partial content requests are supported. This allows the viewer to jump around the video without having to wait for the whole video to load.

To see this in action, take a large (greater than 1 GB) WebM file and load it onto Mediagoblin, and then try to watch it from the middle. Use the download link to open the video in its own browser tab, and try to watch it in the middle. Again, it will hang for a long time as the data buffers. If you use browser tools to examine the request, the request will have one HTTP 200 response.

Host the same video on nginx and watch it through the browser. You will be able to jump around the content as you like. If you examine the network traffic, there will be an HTTP 200 response for the first request and then HTTP 206 Partial Content responses for each click to navigate.

I'm going to try to figure this out myself, but Python isn't my normal playground. A Mediagoblin developer will probably be much faster.

Batch Modify
Note: See TracBatchModify for help on using batch modify.
Note: See TracQuery for help on using queries.