Re: CVS commit: src/tools/compat

[christos%zoulas.com@localhost (Christos Zoulas)]

On Oct 19,  3:11pm, yamt%mwd.biglobe.ne.jp@localhost (YAMAMOTO Takashi) wrote:
-- Subject: Re: CVS commit: src/tools/compat

| > On Feb 3, 11:43pm, YAMAMOTO Takashi wrote:
| > } Christos Zoulas wrote:
| > } > On Oct 18,  2:45pm, yamt%mwd.biglobe.ne.jp@localhost (YAMAMOTO Takashi) 
wrote:
| > } > 
| > } > | it includes headers in /usr/include, which don't necessarily know 
about
| > } > | _NETBSD_TOOLS, doesn't it?
| > } > 
| > } > I changed the needed headers to "know" about the define. mtree passes 
-Wall
| > } > now.
| > } > 
| > } > christos
| > } 
| > } only if your /usr/include is new enough to know it?
| > } i don't think that host tools should expect supports from
| > } host environment.
| > 
| >     They don't.  The support comes from our compat library.
| > 
| > }-- End of excerpt from YAMAMOTO Takashi
| 
| _NETBSD_TOOLS is a specific support from host environment, isn't it?
| 
| i don't see why bother to use host headers here, rather than
| making compat_defs.h provide necessary prototypes.

The autoconf tests look if the host library contains the functions and
assumes that if it does, then the headers must provide it too. Autoconf should
really not assume that. On the other hand it is better to expect that
the host OS will provide both headers and implementation, this is why
I did it this way.

christos