cvs commit: src/usr.bin/truss Makefile amd64-fbsd.c extern.h i386-fbsd.c i386-linux.c ia64-fbsd.c main.c powerpc-fbsd.c setup.c sparc64-fbsd.c syscall.h syscalls.c truss.1 truss.h

[Alfred Perlstein]

* Ed Maste <ed.maste at gmail.com> [070411 13:46] wrote:
> On 11/04/07, Alfred Perlstein <alfred at freebsd.org> wrote:
> >* Ed Maste <ed.maste at gmail.com> [070410 12:47] wrote:
> >> This would make the -s option to gcore redundant (since the process
> >> will be stopped after attaching anyway).  I don't know how useful a
> >> core from a non-stopped process is, anyhow.
> >
> >Very, very useful.
> >
> >Imagine a running program that's having some form of abnormal
> >behavior but can't be stopped (for long) or allowed to core...
> 
> It's stopped for only as long as the core takes to be written.
> 
> Currently, without the -s option the core produced by gcore is
> inconsistent, and it's that behaviour that would be eliminated with my
> change.  Do you actually have a use for inconsistent core files?

Not so much as a need for an inconsistent core so much as a need
for a core without halting the program for the time it takes to
dump core.

-- 
- Alfred Perlstein