Enh Req: Make SIGUSR1 stop tracing
jima at cadence.com
Wed Jan 31 20:17:19 UTC 2007
Hmm, good idea, I will try that.
But it would still be convenient to tell strace after the fact to
In my case, there are several concurrent processes running the same
executable, and it is a bit difficult to determine which one was
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.
P.S. FYI my initial email to strace-devel at lists.sourceforge.net got a
truncated bounce message (which I will attach). Something may be wrong
>From: Daniel Jacobowitz [mailto:drow at false.org]
>Sent: Wednesday, January 31, 2007 11:54 AM
>To: Jim Avera
>Cc: strace-devel at lists.sourceforge.net
>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.
-------------- next part --------------
An embedded message was scrubbed...
From: <postmaster at paran.com>
Subject: [ERR] Enh Req: Make SIGUSR1 stop tracing
Date: Wed, 31 Jan 2007 11:52:14 -0800
More information about the Strace-devel