ratfactor/ziglings

Re-running `zig build` after fixing exercise results in confusing questions about patches

matgis opened this issue · 3 comments

matgis commented

(I am new to both Zig and Ziglings, so please bear with me if I'm doing it wrong and the following is expected behavior. 😊)

Seen on macOS Monterey 12.6.3 with Zig 0.11.0-dev.2613+b42562be7.

Steps to reproduce

  1. Run zig build in a fresh clone of the repository.
  2. Fix the reported issue in exercises/001_hello.zig by inserting pub before fn main.
  3. Re-run zig build to verify the fix, and hopefully proceed to the next exercise.

Expected outcome

The test passes and we move on to exercises/002_std.zig.

Actual outcome

Even before the Ziglings banner appears, I am presented with this question:

Reversed (or previously applied) patch detected!  Assume -R? [n]

After I chose the default response, a follow-up question appears:

Apply anyway? [n]

Selecting the default response prints a warning about the patch being rejected:

1 out of 1 hunk ignored -- saving rejects to file patches/healed/001_hello.zig.rej

Eventually I am told my fix to exercises/001_hello.zig is valid and we move on to exercise two. However, my responses don't take, and these questions are repeated for every fixed exercise every time I re-run zig build:

% zig build
Reversed (or previously applied) patch detected!  Assume -R? [n] 
Apply anyway? [n] 
1 out of 1 hunk ignored -- saving rejects to file patches/healed/001_hello.zig.rej
Reversed (or previously applied) patch detected!  Assume -R? [n] 
Apply anyway? [n] 
1 out of 1 hunk ignored -- saving rejects to file patches/healed/002_std.zig.rej

Additional information

These prompts seem to originate from the patch shell command, which appears to be called from the heal function in tests.zig. The heal function is itself called from addCliTests, which is called at the bottom of build in build.zig.

Workaround

Running the exercises individually with zig build -Dn=1, 2, 3, ... does not trigger the issue.

chrboesch commented

@matgis Thanks for this hint, that should not happen. I'll have a look at it.

  • ❤️2
chrboesch commented

@matgis Could you please try if it works now as expected?

matgis commented

Works great! Thank you! ❤️

  • 👍1