Re: eth0: Via Rhine
On Tuesday, 22. July 2003 20:06, Jochen wrote:
> it worked fine for me untill it came to this point:
>
> make[1]: Leaving directory `/home/jochen/Downloads/gnumach-1.3/i386'
> [...]
> nm -u kernel.o | sed 's/^_*//' | sort -u > kernel-undef
> sed '/^memcpy$/d; /^memmove$/d; /^memset$/d; /^bcopy$/d; /^bzero$/d;
> /^strchr$/d; /^strstr$/d; /^strsep$/d; /^strpbrk$/d; /^strtok$/d;
> /^htonl$/d; /^htons$/d; /^ntohl$/d; /^ntohs$/d; /^etext$/d; /^edata$/d;
> /^end$/d;' kernel-undef > kernel-undef-bad if test -s kernel-undef-bad;
> \ then cat kernel-undef-bad; exit 2; else true; fi
> U bcopy
> U bzero
> U _edata
> U end
> U _end
> U etext
> U htonl
> U memcpy
> U memset
> U ntohl
> U ntohs
> U strchr
> U strsep
> make: *** [clib-routines.o] Fehler 2
I do not know anything about gnumach, but I recognize that sed tries to
delete all these functions (/^bcopy$/d;) (exceptions: _edata, _end) from
kernel-undef.
So I'm sure sed expects "bcopy" instead of "U bcopy". You can try to edit
this script to change this, and do not forget to add _edata, _end. Maybe
you can use another version of nm, which produces the expected output??
Jens
Reply to: