Re: Any info about "Thread has exited with code 32772 (0x8004)"?

You cut out the important line I wrote:

As such, it would be useful to raise this issue with Microsoft, and
require a spec for a
conforming installer program that requires specific return codes, or
classes of return
codes.

"Classes of standards" could mean "0 for success, nonzero for error. If
the user-defined
bit (bit 29) of the return code is 1, then the installer shall provide
[description of
what this means, here] a MESSAGETABLE resource against which that error
code can be
interpreted".

That's one way to specify what is going on, which *could* be universal.

Then note that if two programmers agree on an exit code protocol of their
choice, there is
no problem, so the issue you were complaining about, that
installers-in-general do not
return meaningful error codes, goes away. The creator of the installer
program chooses to
do something and the creator of the program launching the installer
program agrees to use
that specification. If there is no agreement, there is no reason to
expect an arbitrary
program will have to return a code consistent with some other arbitrary
program, and
consequently you are back at the point where the issue of error codes
being meaningless is
again true.