Re: [PATCH] precompose_utf8: use a flex array for d_name

From: Junio C Hamano <gitster@pobox.com>
Date: Jul 3, 2026, 08:39 UTC
Message-ID: <xmqq8q7sjwkl.fsf@gitster.g>
In-Reply-To: <20260703050800.GA29216@tb-raspi4>

Torsten Bögershausen <tboegi@web.de> writes:

> Nice, thanks for the patch.  One minor nit/question: Do we need a
> test_have_prereq PERL in t/t3910 ?

Good question.

Show 11 quoted lines

>> +test_expect_success "status with long non-ASCII filename" '
>> +	test_when_finished "rm -rf long-utf8-status" &&
>> +	git init long-utf8-status &&
>> +	(
>> +		cd long-utf8-status &&
>> +		test "$(git config --bool core.precomposeunicode)" = true &&
>> +		long_utf8_name=$(
>> +			perl -e "print q(a) x 249, qq(\342\200\224) x 3, q(.md)"
>> +		) &&
>> +		test "$(printf "%s" "$long_utf8_name" | wc -c | tr -d " ")" = 261 &&
>> +		printf "content\n" >"$long_utf8_name" &&

I would say that if we are going to use this construct as-is, then we do need the prereq.

But as far as I can see, this is mostly to create a very long filename, which does not require perl at all, with 9 bytes of binary which could be easily done with printf with the ame backslash notation.

So, if we can fix the test, that would be preferrable.

Thanks.

Show 8 quoted lines

>> +		git status --porcelain=v1 >actual
>> +	)
>> +'
>> +
>>  test_expect_failure 'handle existing decomposed filenames' '
>>  	echo content >"verbatim.$Adiarnfd" &&
>>  	git -c core.precomposeunicode=false add "verbatim.$Adiarnfd" &&
>>