Hi Ales -- Quoting Ales Hvezda <[EMAIL PROTECTED]>: > >* In libgeda, o_complex_basic modified to place a new type -- > >OBJ_PLACEHOLDER -- into the object list when a complex is found with > >no symbol file. This happens when the RC files are misconfigured, and > >o_read can't find the sym file. The old behavior was to just place > >nothing into object_list, which meant that the component was deleted > >when the object was saved out. > > Oh great! I'm glad that this is finally fixed. Thanks Stuart. > > I looked over your code and it looks fine, however, I took it > one step farther. I added a graphical representation to OBJ_PLACEHOLDER > and you can now (to a limited degree) manipulate (move, copy, edit > attributes) these objects in gschem. Please test it out and make sure > I didn't break gattrib. The biggest change was to move the logic out > of o_complex_read and put it into o_complex_add (this makes copying > possible).
I grabbed the stuff out of CVS and tried out your graphic. It does look obnoxious, which is the desired effect! :-) There is a bug, however: Most pages with the placeholder graphic are zoomed waaaaay out. When I do a keyboard "ve" to zoom to the extent of the schematic page, the schematic remains zoomed way out. What's strange is this: One page (with two placeholder graphics on it) will zoom normally. Most others (with one or two placeholders) won't zoom in using "ve". Therefore, it looks like some kind of bounding box problem. Any ideas? If desired, I can supply you with the troublesome schematics. > You may want to just put one dialog box instead of one per missing > component. Sorta like my symbol version mismatch dialog box, where I display > one dialog box but with all the symbols listed. Yeah, I considered doing this when I made the change. Here were my thoughts: Pro one dialog box: * It probably looks more professional * It annoys the user less. Pro multiple dialog boxes: * Each dialog box can specify exactly which component is missing, which might help with debug. * The proper response to this problem is to abort the program & fix the problem. This takes only one button click. If the user wants to continue with a misconfigured design, I punish him since he must click the multiple boxes. * I was lazy, and this is how I started the implementation, so I just stuck with it. That having been said, I agree that only one box is more professional, so I'll probably change the code as you suggested. But first, however, I wonder if anybody else wants to respond or make a comment about my pro/con points above? Thanks for your testing, support, and suggestions! Stuart
