Thanks for your efforts, @bjornfor!

Please be aware of the recommendations offered in #445447 (especially point 3 at the top, further explained in @LordGrimmauld’s comment).

nixpkgs-review result

Generated using nixpkgs-review.

Command: nixpkgs-review pr 450177
Commit: b9fb68b627ac173fadc00a1fce965040d662a8af

x86_64-linux

Please be aware of the recommendations offered in #445447 (especially point 3 at the top, further explained in @LordGrimmauld’s comment).

Thanks for the pointers, @panchoh!

taskwarrior2 isn't updated anymore (the current stable is 3.x), so I think cmakeFlag is OK.

BTW, I looked at making a patch/substituteInPlace, but it seems I have to add a new line somewhere in CMakeLists.txt with set(CMAKE_POLICY_VERSION_MINIMUM 3.5)[1], which means it isn't guaranteed to fail on the next update (even though 2.x isn't receiving any more updates), since it isn't changing an existing line..

[1] I first tried changing cmake_minimum_required (VERSION 3.0) to cmake_minimum_required (VERSION 3.5) in top of CMakeLists.txt with substituteInPlace, but that had no effect.

taskwarrior2 isn't updated anymore (the current stable is 3.x), so I think cmakeFlag is OK.

Ah, that alone clears it 👍

BTW, I looked at making a patch/substituteInPlace, but it seems I have to add a new line somewhere in CMakeLists.txt with set(CMAKE_POLICY_VERSION_MINIMUM 3.5)[1], which means it isn't guaranteed to fail on the next update (even though 2.x isn't receiving any more updates), since it isn't changing an existing line..

[1] I first tried changing cmake_minimum_required (VERSION 3.0) to cmake_minimum_required (VERSION 3.5) in top of CMakeLists.txt with substituteInPlace, but that had no effect.

Thanks for the clarification, and for the thorough job 🙏!

nixpkgs-review result

Generated using nixpkgs-review.

Command: nixpkgs-review pr 450177
Commit: b9fb68b627ac173fadc00a1fce965040d662a8af

x86_64-linux

cmake_minimum_required would need to be adjusted in all files for it to work.
Seeing as 2.x receives no more updates, the likelyhood of upstream pushing a version upgrading cmake (and our nixpkgs cli flag subsequently actually downgrading it) seems negligible, and fixing all these occurences would be a lot of find+replace. So i am willing to accept the CLI flag here.

cmake_minimum_required would need to be adjusted in all files for it to work.

Thanks! (I wondered why it didn't work for me after changing the entry in the root file -- didn't think of the possibility to have it more than once. Why would anyone have that directive more than once in a repo though?)

Why would anyone have that directive more than once in a repo though?

i didn't look at the exact reason in taskwarrior - prusa slicer has a similar situation going on, over there the reason is vendored deps having their own cmake files. Which (by logic of being vendored) use whatever cmake things upstream-upstream defined. In the case of taskwarrior, this is probably not actually vendor deps, but there is legitimate scenarios.