FvwmCommand error in opening message fifo (regression from fvwm2) about fvwm3 HOT 6 CLOSED

OK this is a trifle embarrassing but I think what's been happening is that when fvwm invokes FvwmCommand from inside itself, it gets the fvwm2 version in /usr/bin instead of the new one in /usr/local/bin. When I invoke it from my shell, it gets /usr/local/bin. A consequence of having both versions installed. So when I run my script from its function, it says FvwmCommand: invalid option -- 'd' but when I explicitly type "Exec pixelcapture.tcl plain | /usr/local/bin/FvwmCommand -cd" into the console, it all works: Trying to determine if /tmp/fvwm_mfl.sock exists as MFL socket ... Using /tmp/fvwm_mfl.sock as MFL socket. Connecting to /tmp/fvwm_mfl.sock Sending data to MFL socket /tmp/fvwm_mfl.sock After I changed my PATH in both .tcshrc (my normal login shell) and .bashrc so that /usr/local/bin comes before bin, it started working right. But that's a little sub-optimal because I may have broken fvwm2 on my machine, if not this time but if something similar happens in the future. The same thing must be happening to the person on the forum, so I'm not the only one who's being a little dumb. If the package install on my FreeBSD had succeeded, It would have replaced fvwm2 and all would be well, probably. I got the source code for the official 1.0.6a version, and I'll compile that. After all, the hard part was rounding up the libraries, so it's easy-peasy from now on.