[Tinycc-devel] Re: Debian's tcc patches.

[Rob Landley]

By the way, FYI: I changed the default install path in my tree 
from /usr/local/bin to /usr/bin last night, and updated the examples and man 
page to match.

On Monday 26 February 2007 2:10 am, Romain Francoise wrote:
> Hey Rob,
> 
> Rob Landley <address@hidden> writes:
> 
> > I don't understand #2 and #6 don't really care enough about #4 to
> > ask (if I start caring about MacOS X I'll do so directly).
> 
> #4 is not about OS X, it's about GNU/kFreeBSD (a system where
> FreeBSD is used instead of Linux as the kernel, see the following
> page for info: http://www.us.debian.org/ports/kfreebsd-gnu/).

I'm sure both of the people using it will be happy that you have a patch to 
support it, but it doesn't interest me.  And it seems the clean way to do it 
for tcc would be one patch at the top ala:

#ifndef __FreeBSD__
#ifdef __FreeBSD_kernel__
#define __FreeBSD__
#endif
#endif

On a related note, I'm pondering stripping out most of the #ifdefs at some 
point and replaing them with the if (ENABLE) style I introduced to BusyBox a 
couple years back.  I.E. have some header file with things like:

#ifdef __FreeBSD__
#define USE_FREEBSD 1
#else
#define USE_FREEBSD 0
#endif

And then use normal C syntax for the tests and let dead code elimination yank 
the unneeded stuff without having different syntax checking (and potentially 
different brace nesting levels!)

But not before tcc gets better at dead code elimination. :)

> > Change #2: Why shouldn't install call libinstall?
> 
> Because I felt it was of limited use...  I can include it if you
> feel it's necessary.

I'm just trying to figure out whether or not your patch fixes a bug that I 
should merge.  (I changed the /usr/local/bin paths to /usr/bin last night, by 
the way.)  In this case, it seems the answer in this case is "no".

> > Change #6: What exactly is this trying to accomplish?  You
> > generate an error message if the token isn't {, but the error
> > message itself says that the braces are optional?  Could you give
> > me an example of what this fixes?
> 
> See http://bugs.debian.org/259619 for the reasoning behind this
> change.

Ah.  Thanks for the explanation.

Except the hunk I quoted from the patch doesn't fix this problem, it just 
provides a longer error message when it occurs.  Hmmm...  Lemme think about 
this one...

Rob
-- 
"Perfection is reached, not when there is no longer anything to add, but
when there is no longer anything to take away." - Antoine de Saint-Exupery