Re: DDD

[Russell Shaw <rjshaw@netspace.net.au>]

On 12/03/15 18:19, shaunak saha wrote:
> Hi,
>
> We are working on porting GTK port version of DDD. hopefully the first GTK
> official release will be done end of this year. Will make a last motif release
> before that with few fixes.

I found the bug is in XmTextXYToPos(), and is caused by three unicode copyright 
(c) chars in my file being displayed in DDD. Each (c) consists of two bytes: 
0xc2 0xa9 present in the text buffer of the text_w widget.

XmTextXYToPos() only counts the (c) as one char, even though it's displayed as 
two chars on the screen.

The same applies for any other non 7-bit-clean unicode bytes in the file.

ddd-3.3.12/ddd/SourceView.C" line 5071

void SourceView::startSelectWordAct (Widget text_w, XEvent* e,
                                      String *params, Cardinal *num_params)
{
#if XtSpecificationRelease < 6
     selection_event = *e;
#endif

     XtCallActionProc(text_w, "grab-focus", e, params, *num_params);

     if (e->type != ButtonPress && e->type != ButtonRelease)
         return;

     XButtonEvent *event = &e->xbutton;

     XmTextPosition pos = XmTextXYToPos (text_w, event->x, event->y);

     XmTextPosition startpos, endpos;

     if (app_data.source_editing)
	startpos = endpos = pos;
     else
	find_word_bounds(text_w, pos, startpos, endpos);

     ...


XmTextXYToPos() is in the debian source from:
   apt-get source libmotif-common

The deb source adds quite a few patches, which could have an effect on the 
problem, but i couldn't trace into XmTextXYToPos() because of debian library 
dependencies involved in installing more things from source.

The easy fix was to banish all chars not 7-bit clean from the source being debugged.
-- 
regards,
Russell Shaw, B.Eng, M.Eng(Research)