Re: [PATCH v2 2/2] [GSOC] ref-filter: add %(raw) atom

From: Junio C Hamano <gitster@pobox.com>
To: ZheNing Hu <adlternative@gmail.com>
Cc: ZheNing Hu via GitGitGadget <gitgitgadget@gmail.com>,
	Git List <git@vger.kernel.org>,
	Christian Couder <christian.couder@gmail.com>,
	Hariom Verma <hariom18599@gmail.com>,
	Karthik Nayak <karthik.188@gmail.com>,
	Felipe Contreras <felipe.contreras@gmail.com>,
	Bagas Sanjaya <bagasdotme@gmail.com>, Jeff King <peff@peff.net>,
	Phillip Wood <phillip.wood123@gmail.com>
Subject: Re: [PATCH v2 2/2] [GSOC] ref-filter: add %(raw) atom
Date: Tue, 01 Jun 2021 18:54:51 +0900	[thread overview]
Message-ID: <xmqqa6o9ucyc.fsf@gitster.g> (raw)
In-Reply-To: CAOLTT8REM7tF6ojdmnsJt7-RE6a8oX4RFXVcL215YnYTx6b62w@mail.gmail.com

ZheNing Hu <adlternative@gmail.com> writes:

>>     ... because the host language may not support a NUL in the variables
>>     of its string type.
>>
>
> I agree. But host language not only support NUL but also some Non-ASCII
> character and Non-UTF-8 code:

Yup, rephase with "a NUL" with "arbitrary binary data" and you got
what I meant.  Thanks.

>> > +     const char *s = buf->buf;
>> > +     size_t cur_len = 0;
>> > +
>> > +     while ((cur_len != buf->len) && (isspace(*s) || *s == '\0')) {
>> >               s++;
>> > +             cur_len++;
>>
>> Is NUL treated the same as a whitespace letter for the purpose of
>> determining if a line is empty?  WHY?
>
> Well, there seems to be no correction here. But is it true that memory
> like "\0abc" is considered empty?

That sample has 'a' or 'b' or 'c' that are clearly not part of an
"empty" string and irrelevant.  After all, a string " abc" is not
treated as empty in the original implementation, either.

You are treating a block of memory with e.g. " \000 " (SP NUL SP) as
an "empty line" just like you do for "   " (SP SP SP), but I think we
should treat it more like " \001 " or " \007 ", i.e. not an empty
string at all.