git: 2.17.1 -> 2.18.0, round 2

[dtzWill]

• Tested using sandboxing (nix.useSandbox on NixOS, or option sandbox in nix.conf on non-NixOS)
• 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 nox --run "nox-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)
• Fits CONTRIBUTING.md.

---

[dtzWill]

@layus is the hero behind this! See: #42376 (comment)

Anyway I haven't tested this with the breakage it caused on various PR's, will do so shortly.

[dtzWill]

LGTM -- tried nix-build -A lollypop.src --check (might need to run without --check first time) with/without the fix from @layus, seems to work with and fails without. \o/

[dtzWill]

LGTM but I'd prefer someone else to merge / confirm since I/we got it "wrong" the first time around :).

[worldofpeace]

I tested it too and it seems to work ✋

[ciil]

lgtm

[xeji]

the chicken says: staging, maybe?

[dtzWill]

Sounds like a plan, no rush here AFAIK anyway.

[dtzWill]

@GrahamcOfBorg build git gitFull

[FRidh]

What was causing the failures with the last major bump causing it to be reverted 3 times? Is it a reasonable risk here as well? Is it perhaps an idea to test it by building a smaller job?

[dtzWill]

What was causing the failures with the last major bump causing it to be reverted 3 times? Is it a reasonable risk here as well? Is it perhaps an idea to test it by building a smaller job?

Good question--after that silliness we (well, @layus mostly, IIRC) fixed and started always running git's (installCheck) tests to ensure things at least mostly work as expected.

These tests would have caught the problem we encountered last time, re:#38387.

Things were made worse by updates being merged separately (perhaps separate PR's? I don't recall) onto both master and staging, and we kept reverting as the two were merged into each other :(.

The update to 2.18.0 was reverted once already, but not due to general usage failure in the package itself but because our fetchgit apparently relied on behavior upstream changed-- but so IMO this is different since git was working "as intended". Since breakage resulting from git working as expected seems unlikely, I'd say staging is appropriate but if folks disagree I certainly won't stand in the way of performing additional testing first.

[GrahamcOfBorg]

Failure on x86_64-darwin (full log)

Attempted: git, gitFull

Partial log (click to expand)

cannot build derivation '/nix/store/zhdncxbqspnak1i3kn6sfl6bzn9w0f91-opensp-1.5.2.drv': 6 dependencies couldn't be built
cannot build derivation '/nix/store/kck48jwbwip7b4b9xsz9g6ixj029ymb4-serf-1.3.9.drv': 10 dependencies couldn't be built
cannot build derivation '/nix/store/92cll8baa995mfympkqkzizxq65xr38p-docbook2X-0.8.8.drv': 16 dependencies couldn't be built
cannot build derivation '/nix/store/g8axfs3rwdlj2f1nhsdcqs2knfgk8464-libXrender-0.9.10.drv': 6 dependencies couldn't be built
cannot build derivation '/nix/store/17p7x39ql012knmhf16ndkqbbby4ha7b-subversion-1.9.7.drv': 12 dependencies couldn't be built
cannot build derivation '/nix/store/ilc1khy1pld25c3sv5n5ck4flrfyhn3i-git-2.18.0.drv': 29 dependencies couldn't be built
cannot build derivation '/nix/store/85z1jahrsd2r6w5fj10hv9kpv5j1h3s9-libXft-2.3.2.drv': 8 dependencies couldn't be built
cannot build derivation '/nix/store/4bsihm57ns8rga954dz88ygcz52w1mr0-tk-8.6.6.drv': 6 dependencies couldn't be built
cannot build derivation '/nix/store/8d53fx7mihp3mzc5hhg4vlni1wrdipvq-git-2.18.0.drv': 32 dependencies couldn't be built
�[31;1merror:�[0m build of '/nix/store/8d53fx7mihp3mzc5hhg4vlni1wrdipvq-git-2.18.0.drv', '/nix/store/ilc1khy1pld25c3sv5n5ck4flrfyhn3i-git-2.18.0.drv' failed

[GrahamcOfBorg]

Success on x86_64-linux (full log)

Attempted: git, gitFull

Partial log (click to expand)

Files=816, Tests=19219, 216 wallclock secs ( 5.06 usr  1.05 sys + 287.42 cusr 275.58 csys = 569.11 CPU)
Result: PASS
make clean-except-prove-cache
make[2]: Entering directory '/build/git-2.18.0/t'
rm -f -r 'trash directory'.* 'test-results'
rm -f -r valgrind/bin
make[2]: Leaving directory '/build/git-2.18.0/t'
make[1]: Leaving directory '/build/git-2.18.0/t'
/nix/store/7zqshcw39jn3cv8dk5a08rjgl4hl4zxg-git-2.18.0
/nix/store/9c88lp6mwvakckpsf1alh5zp5qs8n7z1-git-2.18.0

[GrahamcOfBorg]

Success on aarch64-linux (full log)

Attempted: git, gitFull

Partial log (click to expand)

Files=816, Tests=19219, 285 wallclock secs (81.41 usr 17.18 sys + 2735.20 cusr 6858.77 csys = 9692.56 CPU)
Result: PASS
make clean-except-prove-cache
make[2]: Entering directory '/build/git-2.18.0/t'
rm -f -r 'trash directory'.* 'test-results'
rm -f -r valgrind/bin
make[2]: Leaving directory '/build/git-2.18.0/t'
make[1]: Leaving directory '/build/git-2.18.0/t'
/nix/store/lp7rciszywn3jmwmmfprd0g5rjx1bhyi-git-2.18.0
/nix/store/7bs2p7dlq9vdnydsmwwy83hbgbrnj1r5-git-2.18.0