Re: [PATCH] libc: Replace i386/sys/fenv.h symlink with an #include shim

Corinna Vinschen via Newlib <[email protected]>
Newsgroups gmane.comp.lib.newlib
Message-ID <YEolJ7/[email protected]>
On Mar 11 08:13, David Macek via Newlib wrote:
> > > > This doesn't work.  The machine-specific header will get installed to
> > > > $DESTDIR/usr/include/sys/fenv.h, and it will now include a non-existing
> > > > file.  We have to keep the symlink or to duplicate the file.
> > >
> > > How come the newlib/libm/machine/i386/fenv.c non-symlink works?  Is it
> > > installed differently?
> >
> > It's not installed at all.  It's a source file so the include only has
> > to work at build time.  The header file on the other hand will be
> > installed on the target.  As such, the result has to work on the target.
> 
> Ah, right.  But then I'm confused how a symlink can work in the first
> place.  I guess some part of the build process copies the contents
> instead of the file.
> 
> More to the point, how big of an issue would be to duplicate the file?

I'm not hot on duplicating an identical file, it just raises maintenance
cost.  I'm currently looking into a solution to share the file without
duplicating it.

> > > Coincidentally, is newlib/libc/machine/ft32/stpcmp.S file correct?
> >
> > I have no idea, you have to ask the guys knowing that chip type.
> 
> I didn't read the instructions, it's just the label looks wrong. :)

Oh, right, I see what you mean.

> Should I start another thread?

Yes, ideally addressing the FT32 CPU in the subject.


Thanks,
Corinna
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.