Age | Commit message (Collapse) | Author |
|
function old new delta
bb_die_memory_exhausted - 10 +10
xstrdup 28 23 -5
xsetenv 27 22 -5
xrealloc 32 27 -5
xputenv 22 17 -5
xmalloc 30 25 -5
xfdopen_helper 40 35 -5
xasprintf 44 39 -5
wget_main 2387 2382 -5
open_socket 54 49 -5
glob_brace 419 414 -5
bb_get_chunk_from_file 146 141 -5
------------------------------------------------------------------------------
(add/remove: 1/0 grow/shrink: 0/11 up/down: 10/-55) Total: -45 bytes
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
|
|
This fragment did not work right:
temp = bb_get_chunk_from_file(fp, &len);
if (temp) {
/* len > 0 here, it's ok to do temp[len-1] */
char c = temp[len-1];
With "int len" _sign-extending_, temp[len-1] can refer to a wrong location
if len > 0x7fffffff.
Signed-off-by: Quentin Rameau <quinq@fifth.space>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
|
|
Signed-off-by: Quentin Rameau <quinq@fifth.space>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
|
|
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
|
|
This also moves bb_get_chunk_with_continuation into its sole user,
parse_config.c.
This allows to optimize both functions separately,
they need to be optimized for speed.
(this need was highlighted by slow modprobe caused in part by slow
bb_get_chunk_with_continuation in config parser).
function old new delta
bb_get_chunk_from_file 7 130 +123
config_read 457 558 +101
bb_get_chunk_with_continuation 194 - -194
------------------------------------------------------------------------------
(add/remove: 0/1 grow/shrink: 2/0 up/down: 224/-194) Total: 30 bytes
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
|
|
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
|
|
This change retains "or later" state! No licensing _changes_ here,
only form is adjusted (article, space between "GPL" and "v2" and so on).
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
|
|
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
|
|
function old new delta
bb_get_chunk_with_continuation - 176 +176
find_pair 169 187 +18
...
process_stdin 443 433 -10
config_read 549 456 -93
bb_get_chunk_from_file 139 7 -132
------------------------------------------------------------------------------
(add/remove: 1/0 grow/shrink: 7/7 up/down: 215/-254) Total: -39 bytes
|
|
This fixes problems with NULs in files being scanned, but
costs +800 bytes. The same can be done to sed (TODO).
|
|
|
|
|
|
function old new delta
xrealloc_vector_helper - 51 +51
create_list 84 99 +15
getopt_main 690 695 +5
passwd_main 1049 1053 +4
get_cached 85 89 +4
msh_main 1377 1380 +3
add_match 42 41 -1
read_lines 720 718 -2
grave 1068 1066 -2
fill_match_lines 143 141 -2
add_to_dirlist 67 65 -2
add_input_file 49 47 -2
act 252 250 -2
fsck_main 2252 2246 -6
man_main 765 757 -8
bb_internal_initgroups 228 220 -8
cut_main 1052 1041 -11
add_edge_to_node 55 43 -12
dpkg_main 3851 3835 -16
ifupdown_main 2202 2178 -24
sort_main 838 812 -26
------------------------------------------------------------------------------
(add/remove: 1/0 grow/shrink: 5/15 up/down: 82/-124) Total: -42 bytes
|
|
text data bss dec hex filename
808035 611 6868 815514 c719a busybox_old
804472 611 6868 811951 c63af busybox_unstripped
|
|
|
|
Add xmalloc_fgetline_str which does strip terminator,
and use it in dpkg instead of xmalloc_fgets_str.
netstat: use xmalloc_fgets_str - allows to eat strings with NULs
(this fixes bug with some weird /proc/net/unix input)
function old new delta
xmalloc_fgets_internal - 191 +191
xmalloc_fgetline_str - 18 +18
do_info 116 120 +4
unix_do_one 451 447 -4
tcp_do_one 423 419 -4
send_tree 369 365 -4
xmalloc_fgets_str 178 15 -163
------------------------------------------------------------------------------
(add/remove: 2/0 grow/shrink: 1/4 up/down: 213/-175) Total: 38 bytes
text data bss dec hex filename
778445 832 7344 786621 c00bd busybox_old
778483 832 7344 786659 c00e3 busybox_unstripped
|
|
|
|
|
|
|
|
xstrdup'ing and memory consumption. Made linewrap saner.
regex matching code was awful - still buggy, but not as
leaky as before. Made buffer size configurable. Killed
several static and on-stack buffers. Hopefully eliminated
staircase effect on Ctrl-C (unable to reproduce).
|
|
|
|
* do not make backup copy by copying (just retain old file)
* correctly fall back to /etc/passwd if user is not in shadow
* fix bug with overlong passwd entries
* be permissive on some kinds of failures
* reduce stack usage
* code size: -500 bytes
|
|
Renaming...
|
|
|
|
|
|
unconditionally, even if it was not a newline.
This was apparently broken by r14254
- whitespace while at it.
|
|
definitions. (That should only be on prototypes.)
|
|
Checking in to reduce the diff between my tree and svn...
|
|
a couple of infrastructure bits.
|
|
|
|
or the posix standard.
Put the cleanup code back the way it was.
|
|
|
|
|
|
|
|
-Erik
|
|
files. Clean up the resulting damage and fix up the makefile.
-Erik
|