The Django version currently in use for EvalAI (1.11.23) is now being marked as insecure by requires.io. While 1.11.23 was a security release and is much safer than its previous versions, it can still be a major trouble in the future as Django will end support for all 1.x versions in April 2020.
The Django version should be updated to 2.2 so that we won't have to make major changes urgently in the API when a new vulnerability is discovered in Django.
However, this can be a big change, many dependencies will have to be updated to support the new version. Along with API updates for Django, the API for these dependencies will also have to be updated.
I think the best way to proceed would be to:
Edit:
Please test and make sure to verify the validity and compatibility of suggestions with EvalAI current master branch.
@nikochiko is this issue open?
i would like to work on this.
Yes, the issue is open!
But heads up, the maintainers haven't shared their opinions yet. And as it will be a big decision to bump django this is only a "spike" task for now.
@adrijshikhar Would you still like to work on it?
Can I take this up?
@devanshbatra04 I don't think @adrijshikhar or anyone else is working on this so you can take it. But it will be a lot of work. :+1:
@devanshbatra04 Sure!
@Rishabhjain2018 I think this issue is free as @devanshbatra04 is working on other issue #2686 . So can I take it?
@Suryansh5545 please go ahead, I have my exams so won't be able to work on either issues this week.
Hey @nikochiko @RishabhJain2018 I think porting complete backend from Django 1.11 to 2.2 is in itself a big task and would take time. So can we break it small tasks so that multiple people would be able to work on it?
@apoorvkhare07
Yes, the first task is to update Django form 1.11.23 to 1.11.28, I think @Suryansh5545 is already working on it.
I want to add that the official docs mention this:
If you鈥檙e upgrading through more than one feature version (e.g. 2.0 to 2.2), it鈥檚 usually easier to upgrade through each feature release incrementally (2.0 to 2.1 to 2.2) rather than to make all the changes for each feature release at once. For each feature release, use the latest patch release (e.g. for 2.1, use 2.1.15).
The same incremental upgrade approach is recommended when upgrading from one LTS to the next.
So we can have separate tasks for testing and adding changes for each move (1.11 -> 2.0 and 2.0 -> 2.1 and so on).
@RishabhJain2018, your input?
@nikochiko in previous issue I discussed with @RishabhJain2018 and there is a plan to upgrade from 1.11 . Also @apoorvkhare07 At this time I am unable to work on this issue so if you want then you can take this issue. After 3 march my exam will be finish then I will be able to work on it. If someone doesn't take it
hey @Suryansh5545 can I take this issue.If you are not working on this issue.
@ritikchauhan-01 Sure , go ahead
@ritikchauhan-01 are you still on it? Looking to contribute to this if I can.
@Rubix982 ok ,go ahead.