utils/gzip: set mtime = 1 when mtime == 0.

[carlocab]

• Have you followed the guidelines in our Contributing document?
• Have you checked to ensure there aren't other open Pull Requests for the same change?
• Have you added an explanation of what your changes do and why you'd like us to include them?
• Have you written new tests for your changes? Here's an example.
• Have you successfully run brew style with your changes locally?
• Have you successfully run brew typecheck with your changes locally?
• Have you successfully run brew tests with your changes locally?

---

This allows us to bottle formulae whose source modified time is indeed
0.

Needed for Homebrew/homebrew-core#123724.

[BrewTestBot]

Review period will end on 2023-02-21 at 14:06:16 UTC.

[carlocab]

One thing worth considering here: making this change might mean that this is a workaround we'll have to carry around forever, since changing this will change the checksum of some bottles (and hence bottling would not be reproducible).

However, this might be a property of any workaround we try, so I'm not sure what the right thing to do is.

If the above is correct, then the comment will need to be amended accordingly.

[Bo98]

How many bottles does this affect in practice? If it's just one or two that's getting (semi-)frequent bottle updates anyway then it's probably ok.

I hope we're not stuck with Ruby 2.6 for that long (Ruby 2.7 has the bugfix) but I guess we'll see come June.

[carlocab]

So far that I know of? Only one at Homebrew/homebrew-core#123724. But throwing an error here is relatively new, so there could be more. Presumably @alebcay spotted some others which is what prompted throwing an error here.

[MikeMcQuaid]

@carlocab Does this change the checksum of any existing bottles?

I hope we're not stuck with Ruby 2.6 for that long (Ruby 2.7 has the bugfix) but I guess we'll see come June.

I think we could bump the requirement for CI/dev-cmd before that but probably want to assume end-users may be running it for a while longer.

[alebcay]

So far that I know of? Only one at Homebrew/homebrew-core#123724. But throwing an error here is relatively new, so there could be more. Presumably @alebcay spotted some others which is what prompted throwing an error here.

I noticed this because my initial sanity testing with GzipWriter (in brew irb) was with mtime of 0. That's how I discovered the corner case.

I actually didn't know of any actual need we had for setting mtime of 0, this is the first I'm hearing of it.

[carlocab]

@carlocab Does this change the checksum of any existing bottles?

If it does, it would be because their checksums were not reproducible. @alebcay also points this out. Making this change will make their checksums reproducible (as long as we never remove the bit that does mtime = 1 if mtime.zero?).

[alebcay]

I'll mention that prior to #14316, the logic for setting mtime on the gz file is the same as in this current implementation. The bug would have caused the explicitly requested zero mtime to be ignored and the resulting bottles would not be reproducible as a result. In this case we are just catching and throwing an error about this previously silent failure.

[MikeMcQuaid]

Thanks again @carlocab!

[BrewTestBot]

Review period skipped due to critical label.