I must admit that the problem of how to handle the orientation has been an on-going struggle. Up until this latest 1241.3 patch, A-Shell pretty much ignored the ORIENTATION parameter in the printer init file until after selecting the printer. And even then, it only worked to force LANDSCAPE.
In the process of investigating the original issue, it seemed annoying to me that the printer dialog didn't get initialized to match the requested ORIENTATION from the start. So I changed that, but the change was again more complicated than it sounds, and as a side effect caused the ORIENTATION in the printer init file to not only initialize the printer dialog, but then to override it later, even if you changed it within the printer dialog.
But as you say, that is arguably correct, since presumably if you went to the trouble of specifying an ORIENTATION in the printer init file, then you really wanted that orientation, and didn't want the user to change it (accidentally or deliberately).
So, unless or until someone posts a strong counter-argument, I'm inclined to leave it this way. In other words, if you specify an ORIENTATION in the printer init file, you're going to get it, regardless of whether you attempt to change it within the print dialog. Previously that was true only for ORIENTATION=LANDSCAPE, but it's now true for any explicit ORIENTATION. (Unfortunately there is no way to disable the orientation option in the printer dialog to make it clear to the user that they don't have power over it, but at least now the orientation in the dialog gets initialized properly.) If you want to allow the user to have control over the orientation, then you cannot specify an ORIENTATION in the print request or printer init file.
P.S. I think I'm going to need to make a few adjustments to the
ORIENTATION topic in the documentation.