On 7/9/14 12:28 PM, Monnier, Gary wrote:
Yes, very curious indeed. Makes you wonder if it crept into V7 as
well. I did see an APAR regarding threads and the profile handle
APIs. Maybe their connected to your issue?
I finally ran this thing under the proverbial microscope, with the ILE
debugger on the server job and the Eclipse debugger on the client.
And as usual, it turned out to be something stupid. Something between
the chair and the screen. Something between MY chair and MY screen.
It seems that my server code was, at that point, misinterpreting a
closed socket as a valid password change request, and attempting to
service that request with what, by then, was a buffer full of encrypted
Not quite sure why it would increment the invalid sign-on counter under
V6, but not under V5, but it was probably either (a) V6 counts bad
passwords in QSYCHGPW calls, but V5 doesn't, or (b) something changed
from V5 to V6 in what a read() on a socket returns if the socket gets
closed from the other end. I don't know, and I don't care; I just want
to put the fix in, and put this whole episode behind me!