Previously, Steve Reid wrote in qdn.public.porting:
Alain Magloire <> alain@qnx.com> > wrote:
: Tony Mantler <> tony@astra.mb.ca> > wrote:
: : Yeah, those ones. The documentation guys seem to have been glossing over
: : a lot of them, and not covering some of them at all.
Actually, the developers have a say in what’s documented and in what detail.
Well, whoever’s making the documentation decisions is deciding to not document some of the stuff I need.
: : Not to mention there’s a few miscategorized and undocumented calls like
: : PtGetWindowConfig(), which you only need when you don’t use Pt calls to
: : make a window (it gets the window manager defaults for stuff like titlebar
: : colour etc).
If something is in the wrong category, please let us know.
PtGetWindowConfig() is the only one that comes to mind atm.
: : It’s an interesting adventure to say the least. > 
: When you find this you should consider them as bugs and report them to
: doc group.
If you find something in a header file and it isn’t documented, you should
usually assume that it’s for internal use only, deprecated, or unsupported.
Details of internal functions can change, so you use them at your own risk.
Ok, how about this one: QNX4 Photon docs, bottom of “PhWindowOpen()” page (which is not present at all in the NTO docs), clearly documents the PhWindowInfo_t structure by saying “See PhWm.h for a description of PhWindowInfo_t”. This structure is rather vital to the functioning of the PhWindowOpen() call.
Also, the same page doesn’t at all touch on the importance of setting the data_len element of the PhRegion_t argument, without which the call will fail with a descriptive error of “0”. Instead, it states that the ‘parent’ member must be filled in, and the rest can be left to the default values.
I mean, I’m not trying to be picky or anything, but…
\
Tony Mantler | Proud ---- Days since the last
QNX Consulting | of our | 27 |
tony@astra.mb.ca | Record ---- “Gerbil Incident”