[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Bug-gnulib] ssize_t and mingw32
From: |
Bruno Haible |
Subject: |
Re: [Bug-gnulib] ssize_t and mingw32 |
Date: |
Tue, 1 Jul 2003 12:34:01 +0200 |
User-agent: |
KMail/1.5 |
Paul Eggert wrote:
> Is there any reason to examine both <sys/types.h> and <unistd.h>?
> POSIX says that <sys/types.h> is the "home" of ssize_t, and I don't
> know of any host that defines it in <unistd.h> but not <sys/types.h>.
You're probably right. I had <unistd.h> included because earlier I used
this autoconf macro for testing the return type of read() and write(),
which are in <unistd.h>...
I've now changed the test to rely uniquely on <sys/types.h>.
> I think
> this is "over-porting", as I can't imagine any host that has readlink
> but not <sys/types.h>.
I agree, I haven't seen any platform in years that does not have
<sys/types.h>.
> The rest of gnulib includes <sys/types.h>
> unconditionally, so how about the following patch?
>
> Index: lib/xreadlink.c
Fine with me, but this is Jim's territory.
Bruno
- Re: [Bug-gnulib] ssize_t and mingw32,
Bruno Haible <=