On Fri, Apr 04, 2014 at 10:51:24PM -0400, Eliot Moss wrote: > On 4/4/2014 10:29 PM, Duncan Roe wrote: > >I just found that gdb's "run" command doesn't action redirection (e.g. run > ></dev/pty2 >/dev/pty2 2>&1, where the shell on /dev/pty2 is doing a long > >sleep). > >Instead, the invoked program gets the redirections as command line arguments. > > > >Looking through the archives, I found > >https://sourceware.org/ml/cygwin/1999-04/msg00355.html documenting this > >behaviour. Chris Faylor commented at the time that fixing it was more trouble > >than it appeared. > > > >That was 15 years ago - has anything changed since? Anyone up for this or > >should > >I have a go? I *could* simply make my target do the redirection itself, but > >that > >doesn't help anyone else. OTOH if changing gdb really *is* that hard, maybe I > >should just change my program anyway. > > > >Any advice welcomed, > > I think this is the intended design (see: > https://sourceware.org/gdb/current/onlinedocs/gdb/Input_002fOutput.html#Input_002fOutput > ). If you want *gdb's* input and output redirected, I would think you want > to invoke > gdb with I/O redirection on the command line, as in: > > gdb foo < infile > outfile > > Regards -- Eliot Moss > Thanks for replying.
Sorry but your suggested command line doesn't help: I need gdb command i/o in one screen and program i/o in another. The set inferior-tty command documented in the link you posted *very nearly* gave me what I was after. Having updated my startup script to use it, I now find that I can't use Control-C to interrupt the session: it's ignored in the gdb window and causes catastrophic gdb failure in the application window (as well as waking the shell from sleep). It's better than nothing I guess. So, I'm back to needing i/o redirection in the run command (as also documented in your link). It works in Linux. Cheers ... Duncan. -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple