[Linux-ia64] umode_t?

[Linux-ia64] umode_t?

[Eric Gillespie, Jr.]

Curves, a frontend to CVS (it's a Debian package, the closest i could
find to a web page is this: ftp://ftp.netcom.com/pub/el/elf/curves),
checks for umode_t with AC_CHECK_TYPE(umode_t, int).  Unfortunately,
it can't find umode_t because it's not in sys/types.h, but only in
asm/types.h.  Are programs supposed to know about umode_t?  If so, it
needs to be declared in sys/types.h somehow.

I'm bringing this up here because Mosberger's name is on the header
and it breaks curves on IA-64.

-- 
Eric Gillespie, Jr. <*> epg@progeny.com
Software Developer
Progeny Linux Systems - http://progeny.com
"I don't know what a Sawfish Window Manager is."

Re: [Linux-ia64] umode_t?

[David Mosberger]

>>>>> On Wed, 26 Sep 2001 13:14:17 -0500, "Eric Gillespie, Jr." <epg@progeny.com> said:

  Eric> Curves, a frontend to CVS (it's a Debian package, the closest
  Eric> i could find to a web page is this:
  Eric> ftp://ftp.netcom.com/pub/el/elf/curves), checks for umode_t
  Eric> with AC_CHECK_TYPE(umode_t, int).  Unfortunately, it can't
  Eric> find umode_t because it's not in sys/types.h, but only in
  Eric> asm/types.h.  Are programs supposed to know about umode_t?  If
  Eric> so, it needs to be declared in sys/types.h somehow.

  Eric> I'm bringing this up here because Mosberger's name is on the
  Eric> header and it breaks curves on IA-64.

Isn't the poinit of AC_CHECK_TYP(umode_t, int) that if umode_t isn't
defined already, it will get defined as an "int"?

You'll need to talk to glibc maintainers as to whether umode_t should
or should not be defined.  I don't keep track of the latest name space
rules.

	--david