Re: syncing Emacs from sources maintained elsewhere

[Eli Zaretskii <eliz@gnu.org>]

> Date: Thu, 27 Jan 2011 01:22:42 -0800
> From: Paul Eggert <eggert@cs.ucla.edu>
> CC: emacs-devel@gnu.org, bug-gnulib@gnu.org
> 
> On 01/26/2011 07:43 AM, Eli Zaretskii wrote:
> 
> >   . Is it true that time_r.c is needed only by mktime.c?
> 
> Yes, for now, but once I add the strftime module, time_r.c will also be needed
> by strftime.c, and in that case MinGW will definitely need to worry about
> time_r.c.
> 
> strftime is quite likely to be added.  I've already coded it up and tested
> it, and it works fine on POSIXy platforms.  I plan to add it to Emacs
> soon.

Thanks for the heads-up.  Does gnulib's strftime support MinGW?

> Once it's in, Emacs's format-time-string function can start supporting
> higher-resolution time stamps.

Is this enabled for MinGW as well?

> >   . Is it true that stddef.in.h, unistd.in.h, and time.in.h are needed
> >     only on systems which have "issues" with the corresponding
> >     standard headers?
> 
> The "issues" are either with the headers, or with the implementations of
> functions declared by the headers.
> 
> >     is it okay to compile gnulib sources in
> >     lib/ without editing these 3 into the corresponding *.h files?
> 
> If the MinGW headers and implementations are good enough, you should
> be OK without the .h files.  I'd be surprised if that were true, though.

Where can I find some docs regarding what is considered to be "good
enough"?

Thanks.