I feel your pain. It is a decision you need to make and a risk you decide on.
I think you have already discussed this stuff internally. Still, I like to debate this topic because it intrigues me so much.
The phrase "need IE6 support" cannot be correct, since it would mean that installing a second browser would let the office building collapse or something. Maybe some customers "want IE6 support". This way they can ignore the web evolvement of the pas decade for yet another few years.
I was thinking: when will the software that requires IE6 lift their restrictions? The answer I came up with: never. That's right: ie6-websites will require IE6 forever.
Why do I think so? Because that is the essence of legacy applications. Nothing new here. Look at the old mainframes that are still in use. "It works, don't touch it". Companies have wrapped these old applications in expensive layers that allows them to interface to a more modern infrastructure.
Some companies who are trapped in an antique browser have bet on the wrong horses and if they did improve their it-decision making in the meantime they will know that their company needs to work around the legacy applications. IE6 is bound to windows xp. In 2014 xp support will die.
Companies will then be forced to upgrade to a new os. And their 50+ ie6 applications? They will survive in virtual machines running xp.
If extjs decides to pursue the difficult path of continuing with the ie6 based code, they will make the ie6-customers happy, because they can rest for another 2½ years. From then they have only 1½ year to migrate *everything* to a new os [win7/vista]/..]. They also need to setup and configure virtual machines running legacy apps that integrate well with the new os (think about datasharing).
To put it mildly: supporting ie6 is a ridiculous waste of time. Within a few years all ie6 instances lives only in special sanctuaries.
If you want to develop for IE6, why not use extjs 3? All in all: what's wrong with being behind?