Steve Anderson has most of the story right, but I have yet to see an recent official document regarding the Oka-Hudson routing. While the idea pops once in a while here and there
(from this engineeing bulletin in 1974 to some chamber of commerce in 2012), the only official allusion I can find is a
1966 press release regarding the inception of the primitive autoroute numbering system ― and it's still very vague about the schedule and the routing, citing only two towns as terminuses.
Beyond 1970, official maps call for a
western terminus of A-640 at Saint-Placide. The choice of Mirabel as a site for the new airport called for a revision of the highway planification northwest of Montreal; if a new east-west expressway was to serve Mirabel, it would run farther north than A-640 and double as a Montréal bypass towards Ontario and western Québec, with A-13 connecting both at the western tip of the metropolitan area (
interview with De Belleval, Transport minister in 1981). 1981 is before the standoff between aboriginal peoples and the army at Oka, and right after a national park was created at the very terminus of autoroute 640.
The most recent (2003) proposal about A-640 completion is the ¾-loop (A-640 and A-30, with a bridge in Repentigny), that motivated the completion of A-30. I give more details about A-640
in a two-post series in another topic.
tl;dr
A full loop looks good on a map, but a bridge over Outaouais river would be as useless as socially explosive.