Silence was the answer.

Nobody possibly felt that the problem is worth to solve.
Note that there is a rationale behind the tries to combine the architectures.

We have a development environment, both hardware and software, for x86 32-bit.

We lack the hardware for doing regular development on x86_64
nor are we inclined to duplicate the efforts for maintaining another
development environment.
Keeping userland 32-bit does indeed spare a lot of resources.

Making the same Coda client binaries work on both platforms
makes the maintenance a lot easier. The barrier seems to be relatively low
(Coda still is 32-bit inside and even when it will become 64-bit,
32-bit clients ought to be capable of supporting that).
The goal is essentially to make Coda even less expensive to develop
and maintain. Each extra platform adds to the cost.

Would somebody give me a hand and either point in the right direction
for the fix or explain that the solution is unavailable or too expensive?

Regards,
Rune