tools/test-crash-trace: do not use fixed signal numbers

The numbers of signals vary by arch. On the common arches the signals
listed here all use the same numbers, but people are likely to use
this on more fringe architectures too, so let's use symbolic names
instead.

Also the comment about gdb "hitting the same kill" didn't make sense.

The syntax is a bit baroque, but using a helper variable does not work.
Also shellcheck complains about $[ ] which would have made this more
legible.

diff --git a/tools/test-crash-trace.sh b/tools/test-crash-trace.sh
index d4a8a60f82b8029eb038d04075a2a791403980d4..fc925d42eeb85d075947507b9d5d96a0831a3168 100755 (executable)
--- a/tools/test-crash-trace.sh
+++ b/tools/test-crash-trace.sh
@@ -17,12 +17,12 @@ fi
 rc=0
 "$@" || rc=$?
 
-# Replay only on actual crash signals (128 + signal): SIGILL=132, SIGABRT=134,
-# SIGBUS=135, SIGFPE=136, SIGSEGV=139. SIGTERM/SIGKILL/SIGPIPE/SIGALRM mean the
-# test was killed by the environment (timeout, etc.), not that it crashed —
-# replaying those just makes gdb hit the same kill.
+# Replay only on actual crash signals (128 + signal).
+# SIGTERM/SIGKILL/SIGPIPE/SIGALRM mean the test was killed by the environment
+# (timeout, etc.), and not that it really crashed, so it is not useful to
+# replay it under gdb.
 case "$rc" in
-    132|134|135|136|139)
+    $((128 + $(kill -l ILL)))|$((128 + $(kill -l ABRT)))|$((128 + $(kill -l BUS)))|$((128 + $(kill -l FPE)))|$((128 + $(kill -l SEGV))))
         if command -v gdb >/dev/null 2>&1; then
             echo "===== exit $rc — replaying under gdb =====" >&2
             style_args=()