[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [MiNT] MiNTlib and FreeMiNT



On Tue, 2004-09-14 at 02:43, Philipp Donzé wrote:
> Hi,
> 
> 
> >> Further I had to change for some tools who missed the constant 
> >> PATH_MAX the include file from <limits.h> to <dirent.h>.
> >> After this change: no compile errors.
> >
> > I wonder how this affect as PATH_MAX is defined in MiNTLib in limits.h.
> 
> If I do a search in the mintlib source directory: find . -iname "*.h" 
> -exec grep -H PATH_MAX {} ";"
> The only matching result is: in <bits/local_lim.h>
> 
> This file is included by <bits/posix1_lim.h> which in turn is included 
> by <limits.h> and <bits/dirent.h>.
> The difference is: in limits.h the include "statement" needs a defined 
> __USE_POSIX. This is not the case for <bits/dirent.h>.
> So I compiled it wrong?
> 
> I just did a "make CROSS=yes".
> 
> 
> 
> >> But: Is it normal that GCC spits out sooooo much warnings? Or is this 
> >> perhaps a compiler version problem?
> >
> > Maybe, you know the system compiler we use is 2.95.3 (and this version 
> > don't report warnings (except one or two)).
> 
> Ups... so there are still some things I'll have to check before I 
> commit any changes!
> Further I need to recompile everything using a GCC 2.95.3 setup to make 
> sure I didn't break compatibility with the system compiler. (Perhaps on 
> my CT60 Falcon. But I think this will still take ages to compile.)
> 
> 
> regards
> Philipp
> 

I think you might be surprised about how long this stuff takes on your CT60 falcon.  Ages probably
isn't a very good description compared to 10 minutes.  I think you need to consider how long it
takes to compile the linux kernel or glibc on current linux hardware in order to quantatively
say it takes ages.  I stand by and say we have competitive compile times on real hardware compared
to linux ;-)  Just my $0.02 as a Gentoo user.  Gcc 2.95.3 is faster than gcc3 too (compile time wise).

Thanks,
Mark