Age | Commit message (Collapse) | Author |
|
|
|
|
|
readlink(2).
|
|
-Erik
|
|
|
|
There was just a never initialized pointer.
-Erik
|
|
|
|
uses it to avoid possible buffer underruns whn strlen is zero, and avoid the
possible space-hogging inline of strlen() in several cases.
-Erik
|
|
|
|
<manty@debian.org>
-Erik
|
|
|
|
for glibc's lchown support. Thanks Larry.
|
|
This way, we can new get rid of all that tedious #define rubbish we used to
need to enable specific messages. This way is enormously simpler, and as a
bonus also ends up saving us 96 bytes.
-Erik
|
|
|
|
adds 384 bytes to the overall size. But having each app be standalone
is the Right Thing(tm) so we will just have to live with it.
-Erik
|
|
|
|
|
|
a patch was provided by Vladimir Oleynik, and am improved patch commited
by me.
|
|
|
|
|
|
|
|
|
|
-Erik
|
|
|
|
|
|
|
|
|
|
-Erik
|
|
|
|
shadowed variables. Move (almost) all syscalls to libbb/syscalls.c, so I can
handle them sanely and all at once.
-Erik
|
|
not going to scale up as well as I would like, and Matt Kraai and I have
discussed a better long term solution. But for now this will at least make all
the human-readable apps give correct answers.
Please test the human readable/non-human readable options on your systems!!!
-Erik
|
|
-Erik
|
|
used by 'du' and 'df'. Now we just need to fix make_human_readable_str
so it behaves they way it used to before Matt fixed it.
|
|
file size. Should close bug #1140.
|
|
|
|
cleans up most of the now-revealed problems.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Problem was 512/1024 = 0 if you use longs, so I cast KILOBYTE to a double, then
cast the result back to an int after the math is done, letting C's type
promotion do its magic for the rest.
-Erik
|
|
|
|
namespace polluting things that really should be static.
|