Re: [PATCH v3 3/4] t1300: add more tests for whitespace and inline comments

[Dragan Simic <dsimic@manjaro.org>]

On 2024-03-20 07:46, Dragan Simic wrote:
> On 2024-03-20 07:42, Junio C Hamano wrote:
>> Dragan Simic <dsimic@manjaro.org> writes:
>> 
>>> +test_expect_success 'setup whitespace config' '
>>> +	tr "X" "\011" >.git/config <<-\EOF
>>> +	[section]
>>> +		Xsolid = rock
>>> +		Xsparse = big XX blue
>>> +		XsparseAndTail = big XX blue
>>> +		XsparseAndTailQuoted = "big XX blue "
>>> +		XsparseAndBiggerTail = big XX blue X X
>>> +		XsparseAndBiggerTailQuoted = "big XX blue X X"
>>> +		XsparseAndBiggerTailQuotedPlus = "big XX blue X X"X
>>> +		XheadAndTail = Xbig blue
>>> +		XheadAndTailQuoted = "Xbig blue "
>>> +		XheadAndTailQuotedPlus = "Xbig blue "
>>> +		Xannotated = big blueX# to be discarded
>>> +		XannotatedQuoted = "big blue"X# to be discarded
>>> +	EOF
>>> +'
>> 
>> If you want to write tests where leading and trailing whitespace
>> matter in them, making these invisible characters visible is a good
>> way to convey your intention to your readers.
>> 
>> 	sed -e "s/Q/	/g" \
>> 	    -e "s/^|//" \
>> 	    -e "s/[$]$//" <<-\EOF
>> 	|[section]
>> 	|	solid = rock  $
>> 	|	sparse = tab and space Q $
>> 	EOF
>> 
>> This is still true even if we assume there is no patch corruption
>> while the e-mail is in transit.  It is safe to assume that the
>> receiving end of your patches uses "git am --whitespace=fix" and a
>> common way to protect against it is to use the opposite of "cat -e"
>> as a convention, additional to the "'|' is the left edge of paper"
>> and "Q stands for HT" conventions.
> 
> Agreed, will make it so in the v4.  I already had some thoughts
> about protecting the trailing whitespace, and making obvious it
> wasn't included by mistake.
> 
> Though, I'll have to use 'X' istead of 'Q' for HTs, because the
> option names already contain 'Q' characters.

Oh, I just saw that you've already picked this patch up in the "seen"
branch.  Would you prefer if I make this change and submit the v4, or
to perform the change in the already planned follow-up patches, which
would also clean up some other tests a bit?