| 1 |
* Hans-Werner Hilse <hilse@×××.de> [04/05/06 21:00]: |
| 2 |
> Hi, |
| 3 |
> |
| 4 |
> On Thu, 4 May 2006 18:55:28 +0300 |
| 5 |
> Moshe Kaminsky <kaminsky@××××××××××××.il> wrote: |
| 6 |
> |
| 7 |
> > Funny, I just tried the same, and it worked. It also didn't print any |
| 8 |
> > "after" (appropriately, since the sig handler includes 'exit'), and I |
| 9 |
> > didn't find any sleep process. Maybe it was from some different |
| 10 |
> > experiment? |
| 11 |
> |
| 12 |
> I don't think so. You probably sent the signal to the child process |
| 13 |
> (CTRL-C, perhaps?) and the script at once, not to the parent only. |
| 14 |
> |
| 15 |
> proper way would be to try it like this: |
| 16 |
> |
| 17 |
> ./test.sh & |
| 18 |
> # wait some short time |
| 19 |
> kill -TERM $! |
| 20 |
> |
| 21 |
> After that I see that the child process is still running. |
| 22 |
|
| 23 |
You are right. I was using 'kill -TERM %1', which sends the signal to |
| 24 |
the whole group, I guess. Anyway, I just tried the analogous thing with |
| 25 |
perl, and it does work. I wonder what is the reasoning in the bash case. |
| 26 |
|
| 27 |
Moshe |
Archives