This is the mail archive of the cygwin-developers@cygwin.com mailing list for the Cygwin project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [aenchevich@shaw.ca: compile error in types.h?]


On Sun, Mar 30, 2003 at 07:10:18PM +0200, Corinna Vinschen wrote:
>On Sun, Mar 30, 2003 at 11:39:57AM -0500, Christopher Faylor wrote:
>> Should the cygwin/types.h in cygwin/fs.h and cygwin/in.h actually be
>> sys/types.h?
>
>It's the simplest solution to include sys/types.h but in theory it
>shouldn't be necessary.  From where did you move the _t definitions to
>cygwin/types.h on 2003-02-10?

2003-02-10  Christopher Faylor  <cgf at redhat dot com>

        * libc/include/sys/types.h: Don't define __MS_types__ for Cygwin.
        Don't define some types under cygwin.

Basically, I was trying to move all control of cygwin's types into the
cygwin headers.

>The problem is that they use the double underscore definitions from
>sys/types.h.  Actually they should go to stdint.h, which can be
>included from cygwin/types.h and cygwin/in.h as desired.

Shouldn't that be 'inttypes.h', which includes 'stdint.h'?  That's how
linux seems to do it.  

I'll just check in the cygwin/types.h -> sys/types.h change for now.
Personally, I don't like what seems to be the convention of defining two
things with similar names and identical characteristics separately
rather than having one depend on the other, though.  That seems to
be how linux does things.

cgf


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]