Shortening the original mail to something shorter.

The reason I am starting here, in -dev, rather than as an issue directly, is because I would like some direction/recommendation from concerned individuals before I take a "outsider" approach. Too often I have learned that I guessed wrong how the Python community "sees the world".

It seems so - however, Is there something such as PyUnsignedLong and is
that large enough for a "long long"? and if it exists, would that make
the value positive (for the first test).

posix.major and os.major will need to mask away the MSB and
posix.makedev and os.makedev will need to add it back.

OR - do I need to make the PyStat values "the same" in both 32-bit and
64-bit?

I think I already have the answer to this - follow the platform (as far as size goes).

On 32-bit max(1) == 256 (2^^8) and on 64-bit 65536 (2^^16).

However, I am still puzzeled re: the idiosyncratic MSB addition - that the value for

st.st_dev

on AIX 64-bit has always (as far back as I could look at least) had the MSB set so that, as a signed number st.st_dev is always negative. And MSB is not part of the major device number. It's just there.

So, it has been a few days since I studied this (waiting for reactions) - but is the return suppossed to always be a value returned by the posixmodule.c (if so, can strip the MSB bit and make everything positive - even for st.st_dev (but that would break under a direct comparision with a C program that does not strip the MSB).

Currently, the macros on AIX for 64-bit are broken, so I cannot guess how they will decide to implement major() and mkdevice() as far as striping (and adding) the MSB when "constructing" the st_dev number from two "regular" major and minor numbers.

Parked for now,

Best wishes,

Michael

Puzzled on what you think is the correct approach.

Michael