[FEATURE REQUEST] Microsoft, give some love to this project!

[spapas]

Is your feature request related to a problem? If so, please give a short summary of the problem and how the feature would resolve it
Yes, the problem seems that Microsoft does not seem to care about this project anymore. Django 5.1 is out for 6 months, yet the project still does not support it.

Describe the preferred solution
Microsoft give some love to this project and finally support Django 5.1 and 5.2 which will be out in 2 months. Only a little is needed to fix the basics.

Describe alternatives you've considered
Fork the project so it can be supported by the community. There already is the https://github.com/jazzband organization that can support abandoned projects.

However there are comments from Microsoft PMs: #418 (comment) that the project is not abandoned. If it ain't abandoned then please prove it by commiting 10 minutes of your time to fix the urgent issues of properly supporting newer Django versions. If it is abandoned the give it to the community. Microsoft is a multi billion company with hundreds of thousands of employes. I can't believe that you can't commit 10 minutes of your time,

Additional context

• The Compatibility issue with Django 5.1 #418 (Django 5.1 compatibility) is open for almost 4 months. The solution is really simple (merge a working PR) but nobody is doing it. This is only a press of a button and creating a new release.
• Django 5.2 will be out in a couple of months. I am already trying the existing code with Django 5.2b and it works fine so supportign Django 5.2 should be like 5 minutes work (update setup.py so Django<6). Why nobody is doing it ?
• Please notice that Djagno 5.2 has a very important change (Composite PKs) so is a very important update for a lot of people.

[Wedge009]

Please notice that Djagno 5.2 has a very important change (Composite PKs) so is a very important update for a lot of people.

I only just found out about this, so yes, being tied to Django 5.0 is a major hindrance.

[David-Engel]

We have not abandoned this project. The team has just been busy trying to address other pain points around the Python + SQL experience. If you want a sneak peek, check out the new Python SQL driver (in alpha): microsoft/mssql-python

That team will be looping back to this Django connector as soon as they can to add support for Django 5.2 (LTS). But to be transparent, I don't think that will happen until Q3, at the earliest.

[hannylicious]

I find it pretty infuriating that microsoft took over this project from a user who did well keeping it up to date and keeping things current - only to let it sit on a shelf and get stale, essentially telling the Django community 'you aren't that important.'

Lots of users have been hamstrung by having to stay on 5.0 and not move to 5.1 (and potentially be testing for 5.2); this has been the case since early October '24.

6 months later and here we are, still wondering about updates, only to be told it will be another 4 months (at the earliest, which in corporate speaks means "this will not happen").

Django support for 5.0 will end before microsoft is intending to get around to updating this driver.

This makes it impossible for people to be proactively testing their applications for Django updates (for 5.1 and 5.2), and additionally it means that people will have to be running apps on a django version that won't receive security updates, yikes.

Microsoft is really putting users in a bind here by ignoring this project. Because it's delaying development and forcing organizations to ignore security if they are on 5.0, which is unacceptable in my opinion.

For those of us who have built things using 5.x of Django and use those features we are left in a bit of a pickle:

• no security updates to the version
• no testing future versions to address issues that might arise ahead of time (putting us behind)
• no ability to roll back to 4.2 for security because we are utilizing some of the new 5.x features

It's upsetting. I wish there was even a little thought given to all of this when microsoft decided it's fine to just completely ignore the only real project keeping django devs connecting to their databases.

Please either support this project the way it should be and give it the attention it deserves, approach the previous dev to see if he wants to take it back over or just let the community know they're on their own again so someone can take the initiative to provide decent support/updates again.

There is a PR that fixes the lib and allows use with 5.1 that has been opened for almost this whole time. Someone else did the work and microsoft can't even just verify the changes are okay and accept it to allow us to move on with our lives/development/security? Unreal.

[hannylicious]

@David-Engel please have someone (at the very least) look at the open PR (where someone already did the work) so it can get accepted and the rest of us can move on with our work and quit having to worry about when Microsoft will actually give the project they took over (and the only project that can do what this does) any kind of attention it needs.

Seriously, the timeline you provided is an obvious security issue that does not even work with Django's release timeline at all. If you (or Microsoft) actually care about this community at all, you would at the very least get that open PR accepted so we can move on to a more secure version before it's out of date. The PR has been open and ready for TEN months. Almost an entire year. If this can't be accepted in the short future? I think it's painfully obvious that you/Microsoft really don't care about this community at all. Then we (the community) will just take over the project so it can be maintained appropriately.

[belthesar]

@David-Engel Thank you for providing information regarding future release expectations. I trust that you and the SQL team at Microsoft are working hard, and can only imagine how hard it is to support a wide swath of MS SQL clients across so many languages. However, the reprioritization of support for this driver, and more importantly, the lack of adequate notice about this reprioritization, leaves a slough of folks in a pretty tough situation.

That team will be looping back to this Django connector as soon as they can to add support for Django 5.2 (LTS). But to be transparent, I don't think that will happen until Q3, at the earliest.

This timetable is incredibly risky for anyone that moved to Django 5.X with the belief that the mssql-django support cadence for Django Short Term Support releases that had been historically supported by the driver would continue apace. By not prioritizing support for future Django STS and a tandem release for the next major Django LTS versions, you're essentially asking product teams that have chosen to use MSSQL as their DB backend to trade application security by eschewing access to critical bug and security fixes for database backend compatibility.

By not surfacing this information sooner, you've lead a lot of companies to make bad choices that they could have avoided by either staying on Django 4.2 (albeit at the detriment of the features and functionality around Django 5, not the least of which being significant performance upgrades) or chose a database backend that is supported first party by the Django project. Ultimately, the way that this communication has been handled gives me, and should give others, pause about continued investment into MSSQL for Django, and potentially other projects as well.

If the Q3 2025 timetable for providing support for mssql-django is real, then clear notice, perhaps in the project's README, should be provided. Additionally, in lieu of first party support from your team leveraging the new mssql-python version, some resources dedicated to working with community PRs to help limp along with the existing architecture would at least allow downstream users to update to supported versions of Django for security and stability fixes.

[thibaudcolas]

Just a heads’up Django 5.2 is due for release this week. As it’s a Long Term Support (LTS) release, a lot of people will be wanting to move over to it directly from Django 4.2. This month is also the end of support for Django 5.0, and Django 5.1 will only be on extended support (see Django’s supported versions). In Q3 of this year we’re probably looking at around 50-60% of the Django ecosystem expecting to be using Django 5.2.

[iamjameswalters]

Historically, they tend to release a new version 3-4 months after each Django LTS. I know that's a bit frustrating to users who want to keep up with latest Djangos, but I don't think it's out of line to move from LTS to LTS if you're relying on this tool.

[hannylicious]

@iamjameswalters that's inaccurate and verifiably false information. Go look at the releases... releases for 3.1 support, 3.2 support, 4.0 support, 4.1 support, 4.2 support. Sometimes with updates every month.

They got to Django 5.0 and stopped.

All maintainers of this driver previously supported all minor versions in between when microsoft took it over, they continued that release cadence.

Not to mention the obvious difficulties presented by only having support on LTS version releases. Inability to prepare codebases for deprecated features and functionality? That's without a doubt a terrible idea and most likely why what you are saying has never been the case.

[iamjameswalters]

You're right, I misspoke--it's been roughly 3-4 months after each release, not just LTS releases. But as @David-Engel shared above, it looks like there's a broader python driver for mssql in the works. My guess would be that eventually mssql-django will be rewritten to wrap it. This is ultimately good news, and is a path for making this package easier to maintain.

[hannylicious]

Yes. I agree that it would be great if it were all wrapped up into one driver that took care of everything - but that wasn't the case before, and from what I have read, this new driver doesn't appear to be/include a connector with Django. They have said they would be coming back to work on this - so I highly doubt the new driver will solve this pain point.

But It's been almost 1 year since anything was last done on this driver and they've said it will be another 4 months before they will look at it at the earliest (so odds are it will be later). So they have been skipping intermediate releases which does nothing but put other people 'behind' in terms of trying to ensure their code/features will not break with updates. Even when this driver was being managed by random individuals from the community - it received more attention and updates than microsoft is currently giving it, which is sad given that they previously gave it a whole team of people and for the last year they can't even let one person manage the PRs from the community. Disgusting display of poor management, in my opinion, given how essential this is to the django community who use mssql.

Microsofts behavior here is unacceptable as it puts people in danger and if they have apps depending on django 5 features? They are forced to either downgrade (potentially breaking the app) or use a version of django that is unsupported and no longer seeing security updates. That is absolutely unacceptable to me and I don't feel I am alone in that given the support in posts above.

If I had any deeper experience with this driver and SQL - I would gladly fork this and begin maintaining it with regularity (I still may). Microsoft has lost all of my hope and credibility - especially after they claim to 'love the django community' and then do this to those very people.

[belthesar]

My guess would be that eventually mssql-django will be rewritten to wrap it. This is ultimately good news, and is a path for making this package easier to maintain.

@iamjameswalters If that's been the roadblock to up to date releases since Django 5, then yes, I wholeheartedly agree. The crux of the issue for downstream users though isn't that the backend module is lagging because it's unsupportable in its current state (which we can only infer, as we have not been directly told as such), but that downstream users weren't given adequate warning that backend development would not continue apace due to a plan to rebuild it against the up and coming mssql-python driver.

At the end of the day, we all know that this is OSS that is being provided gratis to the community, and we have to be thankful that they are putting in the work. We can peer behind the veil a bit and also say that this goodwill is designed to feed more people into the MSSQL ecosystem, whether it be on-prem or in Azure, so it's not all out of the goodness of hearts, but it's also not necessarily a bad thing. However, it's also fair that when Microsoft puts their name on something, there is a higher expectation than if the maintainer is a non-salaried development community. Expectations were set, whether explicitly or implicitly, and when they were not met, the answers received in explanation for the miss leave downstream users in a very compromising situation.

At this point, I don't expect there to be a reprioritization of efforts to accelerate support. It's been a couple weeks since my last post in the thread asking for support for the community PRs that exist to bring them over the finish line, and 5 days ago @David-Engel commented on the draft PR linking to an earlier comment that is also revealing: #418 (comment)

This project hasn't been abandoned. But we did lose some individual contributors

It seems like the various reductions in force that are happening all over the tech world are impacting the support of this backend. I would imagine this will also impact support going forward with the backend, which adds more concern with relying on it in the future.

[Wedge009]

...this is OSS that is being provided gratis to the community, and we have to be thankful that they are putting in the work.

...it's also fair that when Microsoft puts their name on something, there is a higher expectation than if the maintainer is a non-salaried development community.

Both are valid points. However, in my case, the organisation I'm in already has commercial arrangements with Microsoft and when I was brought in to write a Django project to work with a legacy SQL Server database, I noted that this was a Microsoft-endorsed project and thought (perhaps naïvely) there would be a degree of stability as a result. A simple regression (or bug) as #411 - and many others besides - has yet to be tended to.

My experiences with the SQL Server connector and the project I'm working on has led me to be a strong advocate in the organisation for moving away from SQL Server as soon as possible, even if this lack of support is only temporary. Microsoft isn't the only database vendor the organisation uses, and I expect we should be able to find better stability with one of Django's officially-supported databases.

In the meantime, however, we are stuck with SQL Server and the sooner there is functional support and compatibility with Django 5.2, the better. I'm sure I'm not alone in looking forward to an updated release in Q3 2025 as promised...

[hannylicious]

In the meantime, however, we are stuck with SQL Server and the sooner there is functional support and compatibility with Django 5.2, the better. I'm sure I'm not alone in looking forward to an updated release in Q3 2025 as promised...

We are in a similar boat. This has me turned off enough that I have recently suggested we move onto other database options that exist with much better support.

Don't hold your breath on the release in Q3. It wasn't promised, but rather that was 'the earliest' they could look at it. Which is corporate speak for: 'yeah, maybe. most likely not, but maybe.'