Nicolas, Thanks for persisting with your point.
> > > @@ -146,7 +146,7 @@ > > > in the child's memory. > > > As above, the two requests are currently equivalent. > > > .TP > > > -.B PTRACE_POKEUSR > > > +.B PTRACE_POKEUSER > > > > This is not correct. PTRACE_POKEUSR is right. > > After checking, it is quite strange because: > /usr/include/linux/ptrace.h defines PTRACE_POKEUSR > (from the Debian package linux-libc-dev 2.6.22-4) > /usr/include/sys/ptrace.h defines PTRACE_POKEUSER > (from the Debian package libc6-dev 2.6.1-5) > > Also, the man2 ptrace page mentions both PTRACE_POKEUSR and > PTRACE_POKEUSER (in the description of PTRACE_SETREGS and > PTRACE_SETFPREGS). It looks to me like I spoke to soon. I checked the kernel source but not the glibc source, and it looks as though you are right: glibc uses PTRACE_POKEUSER. > Are these ptrace requests standardized in SVr4 and 4.3BSD? Looks like PTRACE_POKEUSER is on AIX and Solaris, but I couldn't find PTRACE_POKEUSR on any of the other systems I checked. > I guess the glibc shall define PTRACE_POKEUSR (and redefine > PTRACE_POKEUSER as an alias for backward compatibility). > (and the same for he PEEK request) I think glibc is trying to do the "right thing" -- employing the names that are used on other implementations. I changed PTRACE_POKEUSR to PTRACE_POKEUSER, and PTRACE_PEEKUSR to PTRACE_PEEKUSER on this page. Seem okay? Cheers, Michael -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
