busybox-w32/libbb/u_signal_names.c, branch zstd

win32: enhanced support for SIGPIPE

2025-02-08T12:49:06+00:00

Commit 68ddd4ec3c (win32: emulate SIGPIPE) allowed broken pipes to be detected so that processes could either return EPIPE or terminate. Later, commit 1b2ee3667 (win32: add fake HUP and QUIT signals) added support for two additional fake signals, thus allowing the 'kill' applet to refer to them by name and the shell to define (though not execute) traps for them. Add PIPE to the list of signals known to 'kill' and the shell, though still without support for calling the trap function. (GitHub issue #482)

libbb: code shrink u_signal_names.c

2021-09-03T12:49:24+00:00

Reduce amount of wasted space in the signals[][] array. Saves 48 bytes.

win32: emulate SIGPIPE

2018-12-11T15:52:31+00:00

The code to check whether a write error is due to a broken pipe can now either: - return with error EPIPE; - cause the process to exit with code 128+SIGPIPE. The default is the latter but the behaviour can be changed by issuing signal(SIGPIPE, SIG_IGN) and signal(SIGPIPE, SIG_DFL) calls. No actual signal is involved so kill can't send SIGPIPE and handlers other than SIG_IGN and SIG_DFL aren't supported. This does, however, avoid unsightly 'broken pipe' errors from commands like the example in GitHub issue #99: dd if=/dev/urandom | tr -dc _A-Z-a-z-0-9 | head -c${1:-32};echo;

libbb/u_signal_names.c: don't check errno after bb_strtou

2018-10-31T10:28:37+00:00

Since we're comparing the return value to a smallish integer anyway, we might as well use that bb_strtou() returns UINT_MAX for malformed input. Referencing errno is kinda bloaty on glibc. While NSIG is not in POSIX, we do already rely on it being defined, compile-time const and smallish, since arrays in struct globals_misc are defined in terms of it. function old new delta get_signum 312 286 -26 Signed-off-by: Rasmus Villemoes Signed-off-by: Denys Vlasenko

libbb: optionally honour libc provided SIGRTMIN/SIGRTMAX in get_signum()

2018-10-31T10:13:55+00:00

When an application documents that it responds such and such to SIGRTMIN+n, that almost always means with respect to the libc-provided SIGRTMIN. Hence I disagree with the "more correct" in commit 7b276fc17594. In any case, this is rather unfortunate: 36 34 (the first shell is bash). We probably can't change default behaviour after 7 years, but at least we can provide a config option. We avoid a little code generation (repeated calls to __libc_current_sigrtmin) by stashing SIGRTMIN/SIGRTMAX in local variables, but it does cost ~50 bytes. The next patch serves as penance for that. Signed-off-by: Rasmus Villemoes Signed-off-by: Denys Vlasenko

regularize format of source file headers, no code changes

2017-09-18T14:28:43+00:00

Signed-off-by: Denys Vlasenko

ash: allow "trap NUM [SIG]..." syntax

2017-07-25T18:06:17+00:00

While at it, make get_signum() return -1 for numeric strings >= NSIG. function old new delta trapcmd 292 306 +14 get_signum 295 300 +5 builtin_trap 413 412 -1 ------------------------------------------------------------------------------ (add/remove: 0/0 grow/shrink: 2/1 up/down: 19/-1) Total: 18 bytes Signed-off-by: Denys Vlasenko

config: deindent all help texts

2017-07-21T07:50:55+00:00

Those two spaces after tab have no effect, and always a nuisance when editing. Signed-off-by: Denys Vlasenko

*: slap on a few ALIGN1/2s where appropriate

2016-04-22T16:09:21+00:00

The result of looking at "grep -F -B2 '*fill*' busybox_unstripped.map" text data bss dec hex filename 829901 4086 1904 835891 cc133 busybox_before 829665 4086 1904 835655 cc047 busybox Signed-off-by: Denys Vlasenko

kill[all[5]],pkill: make signal list show signal numbers, and show RTMIN/MAX

2011-01-03T12:08:58+00:00

function old new delta print_signames 31 64 +33 signals 224 231 +7 Signed-off-by: Denys Vlasenko