[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: address@hidden: Re: Your Emacs changes]

From: Eli Zaretskii
Subject: Re: address@hidden: Re: Your Emacs changes]
Date: Sun, 25 Jun 2006 00:34:21 +0300

> Date: Thu, 15 Jun 2006 22:21:12 +0900 (JST)
> To: address@hidden
> Subject: Re: Your Emacs changes
> From: Kyotaro HORIGUCHI <address@hidden>
> Sorry for delay in replying to your message.
> > We can install your Emacs changes now.  Would you please send
> > me the latest version of them, adapted to the current CVS
> > sources, with change log entries?
> Following patches are to make available make-network-process with
> :nowait on MS-Windows, for emacs that the revision of
> src/ChangeLog is 1.5136.

I applied your changes on my machine, but I found some disturbing
inaccuracies (described below), so I didn't yet commit the patch to
CVS, pending your answers to my questions below.

> +int (PASCAL *pfn_getsockopt) (SOCKET s, int level, int optname,
> +                           const char * optval, int *optlen);

You declare pfn_getsockopt, but you never initialize it: there's no
LOAD_PROC line for it.  Should we add a LOAD_PROC for it?

>  int
> +sys_getsockopt (int s, int level, int optname, char *optval, int *optlen)
This should be "void *", not "char *", as that's what process.c
expects.  Is it safe to use "void *" with your code?

Also, the linker issues the following warning about sys_getsockopt:

    Warning: resolving address@hidden by linking to _sys_getsockopt
    Use --enable-stdcall-fixup to disable these warnings
    Use --disable-stdcall-fixup to disable these fixups

What should we change in the code to avoid this warning?

> - --- process.c       1 Jun 2006 14:08:25 -0000       1.483
> +++ process.c 15 Jun 2006 13:08:01 -0000
> @@ -4831,5 +4831,5 @@
>             p = XPROCESS (proc);
> - -#ifdef GNU_LINUX
> +#ifdef GNU_LINUX || WINDOWS

This is a bad #ifdef (the preprocessor simply ignores everything after
GNU_LINUX).  While it is easy to fix the problem, this makes me wonder
whether the sys_getsockopt code was indeed tested by you, as this last
problem would ifdef away the call to getsockopt, if indeed this is the
code you used on your machine.

Finally, please provide a short test case that would demonstrate that
the new code is working (something that wasn't working in Emacs on
Windows before the patch, but should work after the patch).

Thanks in advance.

reply via email to

[Prev in Thread] Current Thread [Next in Thread]