ok yesterday i tried to build mdadm with dietlibc (i want to use mdadm
in initrd instead of that raidautorun junk)
i found this is impossible, due to some issues

in order of difficulty

strings.h/string.h
for gcc >= 3 dietlibc disables prototypes for
strncpy
strncmp
strncat
because gcc 3.something already defined those, this is not true with gcc
3.2 (at least), so those prototypes will not be defined.

glibc defines strcasecmp in both string.h and strings.h, dietlibc only
in strings.h

nftw
dietlibc has a prototype for nftw, but does not implement it, neither
defines FTW_PHYS, FTW_MOUNT, FTW_DEPTH or FTW_CHDIR (used in th 4th arg
to nftw

-Os
invoking diet -Os gcc ...
yelds
cc1: warning: -malign-loops is obsolete, use -falign-loops
cc1: warning: -malign-jumps is obsolete, use -falign-jumps
cc1: warning: -malign-functions is obsolete, use -falign-functions

getopt:
there is an obscure feature of getopt which dietlibc does not support.
If the first character of optstring is `-', then each non-option
argv-element is handled as if it were the argument of an option with
character code 1.

should i file bugs on dietlibc, or go bother the original dietlibc
authors?

Regards,
L.

for the record i wrote mdassemble which is only 40k, and could be
included in mkinitrd, full mdadm would be 88k static (vs. 80 dynamic),
if only getopt worked :(


-- Luca Berra -- [EMAIL PROTECTED] Communication Media & Services S.r.l. /"\ \ / ASCII RIBBON CAMPAIGN X AGAINST HTML MAIL / \



Reply via email to