what is umask

Lew Pitcher

Not exactly. IIRC, the GE system was a 635 runnoning GECOS-2 (later GECOS-3, still later after Honeywell bought the GE computer division, GCOS3). But I do not believe they ran the code on that machine. They just edited it all there, probably from punched cards (until enough UNIX was running that they sould support it themselves). Remember, the GE did not even do time sharing in those days. When all was done, they probably punched a paper tape or wrote a mag tape, depending on what the PDP-7 would take. It is remotely possible that they had a direct connection between the two machines, but to run it would take some software on the PDP-7. The machines (PDP-7 and GE-635) were in the same building (though on different floors, IIRC), but Dennis and Ken were in research and could have arranged a direct connection pretty easily if they wanted to.

True, and since I never programmed a PDP 7, and I do not remember the op-codes for the PDP-5, I would not know.

True. I know the PDP-11 C compiler generated an XOR and a BITC (or something like that) to do the AND operation in C. Of course, on other machines that actually had an AND operation, the compiler probably would generate that. But when the first UNIX OS was written, it was 6 (IIRC) files of buttembly code. 6 files because the file size limit was 65536 bytes and the kernel was larger than that.

But since the time spent for this is so small, I doubt anyone bothered to change the code. Just one of those silly optimizations (if it were done for that reason).

A am certainly NOT advocating a change in the behavior of umask. The confusion would never ever end.

-- .~. Jean-David Beyer Registered Linux User 85642. V PGP-Key: 9A2FC99A Registered Machine 241939. ^^-^^ 10:05:00 up 10 days, 20:35, 4 users, load average: 4.42, 4.27, 4.20