Printf: Fix printf hex alternate zero

[spineki]

Summary

• Fixed case where giving only 0 would crash printf due to an attempt to read it as an octal
• Fixed wrong prefix for lower and upper hex alternative format. Now, the "0x" is not shown when the value is 0.

Code

GNU:
> printf %#x 0
0

uutils: fails
now shows 0

Fixed an

Related issues:

original : #5810
part of other changes (PR): #5794
part of other changes: #5709
octal similar problem: #5807

[github-actions]

GNU testsuite comparison:

Skipping an intermittent issue tests/tail/inotify-dir-recreate (passes in this run but fails in the 'main' branch)

[samueltardieu]

I would prefer this PR not to be merged as-is if the larger #5783 PR is merged, as it rewrites the parser used in printf and it doesn't have of the single 0 interpreted as octal already.

[tertsdiepraam]

@samueltardieu I agree. For clarity, you're suggesting to only keep the 0x fix from this, right?

[samueltardieu]

@samueltardieu I agree. For clarity, you're suggesting to only keep the 0x fix from this, right?

Yes, and the associated tests.

[spineki]

I did the octal fix in the first place because printf would fail before reaching the 0x-prefix fix.
Thus, amongst the four new tests, those named xxxx_0 will fail without the octal fix, even if the 0x prefix fix is kept.

It means that if you want to keep a subset of this PR without the octal fix, you can

• keep tests on lower hex and upper hex (not the 0 variant), but the prefix fix for 0 will not be tested.
• keep the 4 tests, but the xxx_0 tests variants will fail.

[tertsdiepraam]

Maybe we could try to merge @samueltardieu's PR first and then merge this. That should make everything alright, shouldn't it?

[tertsdiepraam]

I've rebased this on top of main after merging #5783.

[github-actions]

GNU testsuite comparison:

Skip an intermittent issue tests/tail/inotify-dir-recreate (fails in this run but passes in the 'main' branch)