[Bug 3239] Regression with patch id=19048

[Wine Bugs]

http://bugs.winehq.org/show_bug.cgi?id=3239

------- Additional Comments From [e-mail]  2005-02-10 11:20 -------
For duplicating using built a built in design file.
Start programa open File 
  Examples/educational/1563.asc
Hit the running man icon and the ASSERT should trigger.

I took a look and added some debugging traces. It is an off by one error caused
by the patch.
        Added emulation of RichEdit 1.0 behaviour when the 1.0 window class is
    being used. This emulation (introduced in M$ RichEdit 3.0) counts
    paragraph endings as 2 characters (CR+LF) instead of 1 (CR).

Here is a subset of a trace with the assert converted to a TRACE. and a trace
added to ME_SetSelectionParaFormat where it decrement nCharOfs.

trace:richedit:RichEditANSIWndProc hWnd 0x10046 msg 0437 (EM_EXSETSEL) 00000000
7baee27c
trace:richedit:RichEditANSIWndProc EM_EXSETSEL (0,-1)
trace:richedit:ME_RunOfsFromCharOfs At Assert nCharOfs = 124 >= nParaOfs = 0
trace:richedit:ME_RunOfsFromCharOfs At Assert nCharOfs = 124 >= nParaOfs = 32
trace:richedit:ME_RunOfsFromCharOfs At Assert nCharOfs = 124 >= nParaOfs = 34
trace:richedit:ME_RunOfsFromCharOfs At Assert nCharOfs = 124 >= nParaOfs = 59
trace:richedit:ME_RunOfsFromCharOfs At Assert nCharOfs = 124 >= nParaOfs = 91
trace:richedit:ME_RunOfsFromCharOfs At Assert nCharOfs = 124 >= nParaOfs = 124
trace:richedit:ME_RunOfsFromCharOfs At Assert nCharOfs = 0 >= nParaOfs = 0
trace:richedit:ME_RunOfsFromCharOfs At Assert nCharOfs = 124 >= nParaOfs = 0
trace:richedit:ME_RunOfsFromCharOfs At Assert nCharOfs = 124 >= nParaOfs = 32
trace:richedit:ME_RunOfsFromCharOfs At Assert nCharOfs = 124 >= nParaOfs = 34
trace:richedit:ME_RunOfsFromCharOfs At Assert nCharOfs = 124 >= nParaOfs = 59
trace:richedit:ME_RunOfsFromCharOfs At Assert nCharOfs = 124 >= nParaOfs = 91
trace:richedit:ME_RunOfsFromCharOfs At Assert nCharOfs = 124 >= nParaOfs = 124
trace:richedit:ME_RunOfsFromCharOfs At Assert nCharOfs = 124 >= nParaOfs = 0
trace:richedit:ME_RunOfsFromCharOfs At Assert nCharOfs = 124 >= nParaOfs = 32
trace:richedit:ME_RunOfsFromCharOfs At Assert nCharOfs = 124 >= nParaOfs = 34
trace:richedit:ME_RunOfsFromCharOfs At Assert nCharOfs = 124 >= nParaOfs = 59
trace:richedit:ME_RunOfsFromCharOfs At Assert nCharOfs = 124 >= nParaOfs = 91
trace:richedit:ME_RunOfsFromCharOfs At Assert nCharOfs = 124 >= nParaOfs = 124
trace:richedit:RichEditANSIWndProc hWnd 0x10046 msg 0447 (EM_SETPARAFORMAT)
00000000 7baee2e0
trace:richedit:ME_SetSelectionParaFormat nTo>nForm 124 0
trace:richedit:ME_RunOfsFromCharOfs At Assert nCharOfs = 0 >= nParaOfs = 0
trace:richedit:ME_RunOfsFromCharOfs At Assert nCharOfs = 123 >= nParaOfs = 0
trace:richedit:ME_RunOfsFromCharOfs At Assert nCharOfs = 123 >= nParaOfs = 32
trace:richedit:ME_RunOfsFromCharOfs At Assert nCharOfs = 123 >= nParaOfs = 34
trace:richedit:ME_RunOfsFromCharOfs At Assert nCharOfs = 123 >= nParaOfs = 59
trace:richedit:ME_RunOfsFromCharOfs At Assert nCharOfs = 123 >= nParaOfs = 91
trace:richedit:ME_RunOfsFromCharOfs At Assert nCharOfs = 123 <  nParaOfs = 124
trace:richedit:ME_AddUndoItem Pushing id=diUndoSetParagraphFormat to undo stack,
deleting

The ME_SetSelectionParaFormat decrements nCharOfs from 124 to 123 and may not be
needed.

I don't know any of the richtext formats to determine where to correct the
problem. So I will be leaving the correction to somebody with more knowledge on
the subject.