@zooba Could you have a look a this? I won't have any effect until we upgrade to SQLite 3.35.0, but it won't do any harm applying it now.

This PR is stale because it has been open for 30 days with no activity.

Is there anyone else from the Windows team who can review this?

I can confirm from a quick scan that this looks OK. I haven't built Python with this PR, though, so that's all I can say, sorry.

Thanks, Paul. The CI should fail for win32 and win64 if it didn't build, right?

I'd assume so, yes.

LGTM as well, but I do think we want to hold off on this until we are in fact using a version that it affects. If we merge this before then, we're giving the impression that the functionality is there when it actually isn't.

You could get fancy and include a version check, but that may be more trouble than it's worth.

LGTM as well, but I do think we want to hold off on this until we are in fact using a version that it affects. If we merge this before then, we're giving the impression that the functionality is there when it actually isn't.

That depends on how we phrase the NEWS item, I guess. However, adding this now will make it slightly easier to try out the current nightly SQLite builds. OTOH, those who do so are very likely to just add the compile definition themselves, so I guess the real gain is negligible. I'm fine with holding off until 3.35.0. We can include this when we're upgrading the Windows build to 3.35.0.

UPDATE: The Windows installer will be updated in #25641

You could get fancy and include a version check, but that may be more trouble than it's worth.

Totally agree.

FYI, rebased onto master bco. #24797.

@zooba: now that #25641 is merged, would you mind merging this as well? It's already approved by members of the Windows team.

Thanks @erlend-aasland for the PR, and @ambv for merging it 🌮🎉.. I'm working now to backport this PR to: 3.10.
🐍🍒⛏🤖

@ambv: Please replace # with GH- in the commit message next time. Thanks!

GH-25892 is a backport of this pull request to the 3.10 branch.