A multi-phased trip seems reasonable, where you land the production units, and some robotics to deal with the unknowns.
Then before the manned trip even happens you know that there is viable base already operating. If something happens between the manned launch, and their landing, their spaceship could be setup to handle a shot around mars for a return trip as an emergency measure.
Still making a ship capable of holding 6 years of supplies isn't very attractive.
I agree, you're going to want a base on Mars before you land there. No point going all the way there just to spend three days there (like the Moon) so significant living quarters are a great idea. To my way of thinking the best way to build is using the local materials -- either digging underground or building out of rocks. If you were clever, and if you could find a decent source of water, you could probably build a pretty good structure by using dirt and ice. This, of course, requires a lot of fancy robots.
As far as the ship goes, you're definitely going to need a "mothership" capable of entering and leaving Mars orbit (so, a crapload of fuel onboard or else a new engine technology) as well as a lander, so abort capability shouldn't be that big an issue.
The more I think about it, the more difficult a Mars mission seems, and the more sensible it seems to keep on sending robots until we're really ready for a permanent Mars colony (i.e. the point at which it actually makes economic sense).
Still making a ship capable of holding 6 years of supplies isn't very attractive.