Changes made, will submit an updated patchset after finalizing the rest of this patch. On 6/13/11 11:06 PM, Eric Blake wrote: > I'm debating whether this line belongs here, or whether it belongs > better in virsh.c. That is, I don't know whether this is a > virsh-specific feature, or whether all of the libvirt library should > behave like this. This very well could be only be needed in virsh, it just seemed like util/util.c was where it could have the least duplication of code. > I'm not even sure if stdin being a tty is the right check. isatty() produces the correct behavior in all cases but the first one you mentioned, ie running "virsh define", which will not wait for input with the patch as it is. I see the ctl->imode parameter used in main() to determine the interactivity of the session, though that is only accessible at the cmdOPERATION level, not in virFileReadAll(). Maybe (yet) another read wrapper local to virsh.c that would check the ctl->imode if no path is offered, and make a decision from there. Thoughts? -- ~Michael Williams -- libvir-list mailing list libvir-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/libvir-list