converse@EXPO.LCS.MIT.EDU (Donna Converse) (05/30/90)
> First of all, the very definition of XtNameToWidget seems > sorta bizarre. It requires a name starting *below* the > reference widget, not *from* the reference widget. If it started at the reference widget in trying to match the name, and the reference widget matched, it would simply return its first argument. That functionality isn't useful, so it starts with the children of the reference widget. > Nevertheless, it has the following bug: it always checks > the children of a widget, and this does not work when the > children are *Gadgets*... It will seg fault if you give it a non-widget argument in the first parameter. The spec stipulates that it will take a non-widget argument, so this is a bug in the code. Since Objects can not and will not ever have children, XtNameToWidget should return NULL if passed a non-widget argument as the reference widget. Test that it is not a widget, rather than using XmIsGadget. Thanks for bringing this up. I filed a bug report. Donna Converse converse@expo.lcs.mit.edu
marbru@auto-trol.UUCP (Martin Brunecky) (06/01/90)
In article <9005301627.AA26071@expo.lcs.mit.edu> converse@EXPO.LCS.MIT.EDU (Donna Converse) writes: > >It will seg fault if you give it a non-widget argument in the first >parameter. The spec stipulates that it will take a non-widget argument, >so this is a bug in the code. Since Objects can not and will not ever >have children.... STOP here ! Why ? Motif "flavor" of R3 Xt had a concept of compositeObject. A very usefull feature - at least when you want to REALLY start using Xt geometry management. By that I mean far more intelligent geometry managers than what's currently around ( attached boxes and row columns ). Geometry managers used in nested combinations, so that the geometry negotiation really takes place. In such configurations you may end up having several layers of pure geometry management objects - with no need for a window at each of them. Thus, PLEASE, don't rule out composite objects - not yet ! -- =*= Opinions presented here are solely of my own and not those of Auto-trol =*= Martin Brunecky marbru@auto-trol.COM (303) 252-2499 {...}ncar!ico!auto-trol!marbru Auto-trol Technology Corp. 12500 North Washington St., Denver, CO 80241-2404