fish-shell: update to 3.5.1.

[triallax]

Testing the changes

• I tested the changes in this PR: YES

[paper42]

history_races test fails on x86_64-musl:

 61/236 Test  #64: history_races ............................***Failed    1.27 sec
Proceeding with target execution
Testing low-level functionality
Testing history race conditions
Error: Expected 1025 items, but instead got 1005 items
Error: Line '0 0' still left in the array
Error: Line '0 1' still left in the array
Error: Line '0 2' still left in the array
Error: Line '0 3' still left in the array
Error: Line '0 4' still left in the array
Error: Line '0 5' still left in the array
Error: Line '0 6' still left in the array
Error: Line '0 7' still left in the array
Error: Line '0 8' still left in the array
Error: Line '0 9' still left in the array
Error: Line '0 10' still left in the array
Error: Line '0 11' still left in the array
Error: Line '0 12' still left in the array
Error: Line '0 13' still left in the array
Error: Line '0 14' still left in the array
Error: Line '0 15' still left in the array
Error: Line '0 16' still left in the array
Error: Line '0 17' still left in the array
Error: Line '0 18' still left in the array
Error: Line '1 0' still left in the array
Encountered 21 errors in low-level tests

[icp1994]

Could try again with v3.5.1. I don't have a -musl system for the tests but they pass in my x86_64.

Edit: packaged in a separate masterdir for x86_64-musl and the tests passed.

[triallax]

@icp1994 sure, I'll update this PR tomorrow.

[triallax]

Updated to 3.5.1. With 3.5.1, the tests failed for me once (didn't take note of which test it was that failed, but I think it was history_races), but then I ran them again and they succeeded. The tests were done on x86_64 glibc. I think this is most likely just a flaky test that we can safely ignore.

Also, note that I removed the groff dependency, which I think is not needed.