On: Thursday, November 11, 1999 9:14 PM William B. Klein wrote:
>a FrmGetActiveForm() call from within that callback returns a NULL pointer.
This "behaviour" appears to have been introduced with OS3.3 We raised it
with Palm
some weeks back so maybe 3.5 has a fix - I know not. In our case we were
tracking
form ids in globals for other purposes so a simple work round was:
frmP = FrmGetFormPtr(gDATA.CurrentFrmID[gDATA.CurrentFrmIDx]);
More generally - what's the view on backward compatibility? The seeding
program is
of course excellent for finding "new os breaks my app" cases in advance. For
example
I note a new report in our defect db along the lines that Build 20 3.5 roms
now kill
cases where in a 3.0 build app WinCreateOffscreenWindow is called prior to
FrmDrawForm
- I'm guessing some change such that window initialisation in FrmInitForm
conflicts
with WinCreateOffscreenWindow attempting to clone bit depth or whatever?
The pain when backward compatibility breaks is not too great for ESD
product.
It rises somewhat when you have physical goods in the channel and further
still in
the case of say a masked ROM module ...
Rgds
Chris
