nheko: 0.6.4 -> 0.7.1

[doronbehar]

DONT MERGE, this PR should work once #73940 is merged.

Motivation for this change

https://github.com/Nheko-Reborn/nheko/releases/tag/v0.7.1

Things done

• Tested using sandboxing (nix.useSandbox on NixOS, or option sandbox in nix.conf on non-NixOS linux)
• Built on platform(s)
  • NixOS
  • macOS
  • other Linux distributions
• Tested via one or more NixOS test(s) if existing and applicable for the change (look inside nixos/tests)
• Tested compilation of all pkgs that depend on this change using nix-shell -p nixpkgs-review --run "nixpkgs-review wip"
• Tested execution of all binary files (usually in ./result/bin/)
• Determined the impact on package closure size (by running nix path-info -S before and after)
• Ensured that relevant documentation is up to date
• Fits CONTRIBUTING.md.

[rnhmjoj]

Oh, it has been released, finally. I had prepared the update but you beat me to it.
I don't remember having problems with boost but I'm on 20.03, here's the commit maybe it can be of help.

[doronbehar]

One difference I spotted is the use of boost171 vs boost172 ( 17x is used here). I tested with boost171 and still the same error. @rnhmjoj what's the meaning of:

    "-DBoost_NO_BOOST_CMAKE=TRUE"

here?

It fixes the issue but it feels hacky to me considering #73940 ...

[veprbl]

@doronbehar Did you try #85254?

[doronbehar]

@veprbl #85254 works just as well as #73940

[bqv]

I've been using 0.7.0ish for some days, i don't remember having an issue with mtxclient building?

[bqv]

Looks reasonable compared to the version I'm running, and nothing missing in that inputs list according to @deepbluev7

[veprbl]

@GrahamcOfBorg build nheko mtxclient tweeny

[bqv]

I've been using 0.7.0ish for some days, i don't remember having an issue with mtxclient building?

Nevermind, I remember it now.

@doronbehar You may find this useful - I'm running and talking with this as we speak. Specifically I think the part that fixes your boost issue is the symlinkJoin
https://github.com/bqv/nixos/blob/live/overlays/nheko.nix

[doronbehar]

That's no good - I'm pretty sure that your final derivation references both boost.dev and .out - Not good for closure size.

[doronbehar]

@veprbl the build won't work until either #85254 or #73940 are merged. I don't want to add:

    "-DBoost_NO_BOOST_CMAKE=TRUE"

Which fixes the issue here because that's not a real fix. Please 🙏 consider merging either #85254 or #73940.

[veprbl]

@doronbehar I can't just blindly merge any of these and I haven't had time yet to look at those in detail. I propose you use the workaround for now with a comment that it can be removed once #85254 and #73940 are resolved.

[deepbluev7]

Btw, you may want to include this commit in your package: Nheko-Reborn/nheko@d94ac86

I accidentally broke room joins in 0.7.x and released 1.7.1 just before I noticed that.

It will take a bit until the next release, that includes this fix.

[doronbehar]

@doronbehar I can't just blindly merge any of these and I haven't had time yet to look at those in detail. I propose you use the workaround for now with a comment that it can be removed once #85254 and #73940 are resolved.

OK, done.

Btw, you may want to include this commit in your package: Nheko-Reborn/nheko@d94ac86

Thanks @deepbluev7 for joining in and suggesting this, I appreciate your effort you are putting in Nheko.

Unfortunately, and I may need help with this @veprbl , I get this error:

I've applied the patch and everything seems to be OK.

[doronbehar]

@GrahamcOfBorg build nheko

[doronbehar]

I'm testing the darwin build with:

@GrahamcOfBorg build nheko

[deepbluev7]

The build error you get on Darwin looks like your clang is too old to support the deduction guides for std::array. We are only testing clang-6 in our CI. (Not sure, how you guys handle compiler versions.)

[doronbehar]

OK @deepbluev7 that certainly should be possible to fix but this needs a real Darwin machine. I added a comment that links here.

Besides that all is ready.

[deepbluev7]

@bqv clang-6 is just the minimum version, anything newer should work (or even apple clang in Xcode 10.2+). But the CI job used clang-5, which seems to be too old.

[rnhmjoj]

@bqv That's not that trivial, because we use mkDerivation by qt5... It's too much effort for me to test this, as my only means are GrahamcOfBorg.

libsForQt5.mkDerivation is very simple. Basically you can replicate it just by adding wrapQtAppsHook. See pkgs/development/libraries/qt-5/mkDerivation.nix.

[doronbehar]

@GrahamcOfBorg build

[doronbehar]

The Darwin build still fails, but differently:

https://logs.nix.ci/?key=nixos/nixpkgs.85922&attempt_id=1d08c688-e7d7-45c0-bb00-c25f49ce9651

[deepbluev7]

@doronbehar The current build fails, because some C++17 library features are only available in macOS 10.14.

[doronbehar]

@doronbehar The current build fails, because some C++17 library features are only available in macOS 10.14.

Thank you @deepbluev7 for replying. If so, we are almost there but I'm afraid I couldn't find a way to mark it as broken only if the minimum macOS version is not met... Hence I'm marking it as broken for every macOS version for now. If anyone knows a way to do that, please tell me. For now, the PR should be ready.

[doronbehar]

/ping @veprbl

[veprbl]

Passes nixpkgs-review on NixOS, runs, diff looks mostly okay, could use some (light) squashing.

[doronbehar]

Thanks @veprbl for catching all of these imperfections. I went through all the commits and made sure they all line up to your comments.