Re: [PATCH v2] git-prompt: make colourization consistent

From: Joakim Petersen <joak-pet@online.no>
Date: Jun 3, 2022, 13:55 UTC
Message-ID: <09d485f5-e3f0-be10-7061-bff6ef09a99a@online.no>
In-Reply-To: <xmqqilpiistk.fsf@gitster.g>

On 03/06/2022 00:49, Junio C Hamano wrote:

Show 24 quoted lines

> Joakim Petersen <joak-pet@online.no> writes:
> 
>> Range-diff against v1:
>> 1:  e235caa7a8 = 1:  e235caa7a8 git-prompt: make colourization consistent
>>
>>   contrib/completion/git-prompt.sh | 1 +
>>   1 file changed, 1 insertion(+)
>>
>> diff --git a/contrib/completion/git-prompt.sh b/contrib/completion/git-prompt.sh
>> index 87b2b916c0..dfd6cef35f 100644
>> --- a/contrib/completion/git-prompt.sh
>> +++ b/contrib/completion/git-prompt.sh
>> @@ -286,6 +286,7 @@ __git_ps1_colorize_gitstring ()
>>   	if [ -n "$u" ]; then
>>   		u="$bad_color$u"
>>   	fi
>> +	p="$c_clear$p"
>>   	r="$c_clear$r"
>>   }
> 
> Has this been tested?
> 
> It seems to break a handful of tests in t9903 for me.
>

Oh, no I hadn't run the test suite, sorry. I must've gotten too caught up in other parts of the submitting process and simply forgot to run them. After looking into it, the reason why the tests fail is that $p is no longer empty when it is not set, so $f is no longer empty, leading to both $z and $p being inserted into $gitstring. This causes there to be three clear-colour sequences in the final output instead of the expected one. Wrapping the clearing of $p's colour in a check for emptiness like the other short state indicators fixes the tests. I'll submit a v3 shortly.