Age | Commit message (Collapse) | Author |
|
naming along the way to get USE() macros for everything.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
process are exhausted.
|
|
Cherry-picked from Devin Bayer's big MacOS X patch.
|
|
|
|
|
|
md5_sha1_sum.c to convert some #ifdef CONFIG to if(ENABLE).)
|
|
|
|
area. Probably broke something...
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
display or extract. (Only shows up when no gnu longname extension, either
because it's disabled or because tarball hasn't got it.)
|
|
|
|
|
|
truncated for length. SVN 14135 made sure that the truncated version would
always be null terminated. SVN 14144 broke this for no readily apparent
reason, and I have no idea what it was even trying to accomplish. Reverted.
|
|
|
|
xfunc/BB_STRLEN_IMPLEMENTATION
|
|
|
|
|
|
code: we can do a null dereference if one of our huffman tables has all zero
length codes. This fixes it. (Thanks solar.)
|
|
gethostname
|
|
|
|
|
|
|
|
it, breaking compile on systems lacking this function.
|
|
|
|
|
|
|
|
|
|
|
|
pathological archives.
(Unlikely to have security implications, the only way to trigger it basically
wound up doing memset(dbuf,x,2^31) and triggering an immediate segfault. The
test basically gives us a more polite error message.)
Thanks to Ned Ludd and the Gentoo security guys for finding this.
|
|
|
|
|
|
|
|
|
|
removed depend loop: busybox.h depend with BB_BT, and all sources depend with busybox.h
|
|
(only\!)
|
|
|
|
options.
|