RE: Enh Req: Make SIGUSR1 stop tracing

[Jim A. <ji...@ca...>]

Hmm, good idea, I will try that.

But it would still be convenient to tell strace after the fact to
detach.
In my case, there are several concurrent processes running the same
executable, and it is a bit difficult to determine which one was=20
started by my test script after the fact (the actual fork is done by
a large program I can't change easily, and the PID isn't logged).

Anyway, thanks for the suggestion.=20

-Jim

P.S. FYI my initial email to str...@li... got a
strangely
truncated bounce message (which I will attach).  Something may be wrong
with sourceforge.

>-----Original Message-----
>From: Daniel Jacobowitz [mailto:dr...@fa...]=20
>Sent: Wednesday, January 31, 2007 11:54 AM
>To: Jim Avera
>Cc: str...@li...
>Subject: Re: Enh Req: Make SIGUSR1 stop tracing
>
>On Wed, Jan 31, 2007 at 11:51:25AM -0800, Jim Avera wrote:
>> PROPOSED ENHANCEMENT:
>>   1. Make SIGUSR1 tell strace to stop tracing the child process,
>> allowing it to run free.
>>       (presumably strace would thereupon wait() for the child to exit
>> before exiting itself).
>
>Have you considered using strace -p to attach, instead?  Then you can
>send strace a C-c, and it will detach before it exits.
>
>
>--=20
>Daniel Jacobowitz
>CodeSourcery
>