From: Jonathan Tan <jonathantanmy@google.com>
Date: Tue, 30 May 2023 17:29:33 GMT
Subject: Re: [PATCH v2 3/7] sane-ctype.h: move sane-ctype macros from git-compat-util.h
Message-ID: <20230530172933.2781693-1-jonathantanmy@google.com>
In-Reply-To: <4d29e5d4-6ce3-c099-15a7-e89ad787780d@gmail.com>

Phillip Wood <phillip.wood123@gmail.com> writes:
> > Any specific reason for the "sane-" prefix? I think it would make more
> > sense if it was just named ctype.h: see below.
> 
> I don't have a strong opinion either way but "sane-ctype.h" makes it 
> clearer why we're doing something different to the standard library.

Ah...that's true.

> >> -extern const signed char hexval_table[256];
> > 
> > And this one has nothing to do with ctypes or sanity, but rather, what's
> > considered to be a hex character. I think we need another patch to move
> > this to hex.h.
> 
> Isn't "what's considered  to be a hex character" related to ctypes and 
> sanity though as it is used by the "sane" definition of isxdigit().
> 
> >> -#define isxdigit(x) (hexval_table[(unsigned char)(x)] != -1)
> > 
> > Same for this one.
> 
> I'd much rather keep all of our "sane" ctype replacements in one place 
> as Calvin is proposing. isxdigit() is defined in <ctype.h> so it should 
> be in our "sane" version of that header.

Good point.