>>>>> "Alejandro" == Alejandro Aguilar Sierra <[EMAIL PROTECTED]> writes:

Alejandro> On Fri, 18 Dec 1998, Jean-Marc Lasgouttes wrote:
>> This happens when you copy (whole inset) an equation array with
>> numbering turned on (M-m n). I just tried it again. It seems that
>> everything works well when numbering is off.

Alejandro> Sorry but I can't reproduce the error even when numbering
Alejandro> is on with labels or not. Maybe it's a bug that doesn't
Alejandro> appear in all platforms, like the one in AIX. Any clue?

Sorry, but I have no reasonable hint yet :(

A question though: MathedRowSst constructor is often called as 
foo = new MathedRowSt(ncols+1), while the constructor already does:

MathedRowSt(int n) {
        w = new int[n+1]; 
[...]

Isn't there an additional '+1' there? I mention this because purify
gives me errors like

      UMR: Uninitialized memory read
      This is occurring while in:
            MathMatrixInset::Metrics() [math_inset.C:442]
                       
                       cxrow->y = (cxrow==row) ? cxrow->asc:
                                  cxrow->asc + cprow->desc + MATH_ROWSEP + cprow->y;
            =>         h += cxrow->asc + cxrow->desc + MATH_ROWSEP;    
                       cprow = cxrow;
                       cxrow = cxrow->next;
                   }
            MathedCursor::Insert(unsigned char,MathedTextCodes) [math_cursor.C:354]
            InsetFormula::LocalDispatch(int,const char*) [formula.C:775]
            LyXFunc::Dispatch(int,const char*) [lyxfunc.C:495]
            LyXFunc::processKeyEvent(_XEvent*) [lyxfunc.C:319]
            LyXView::KeyPressMask_raw_callback(forms_*,void*) [LyXView.C:337]
      Reading 4 bytes from 0x4b64dc in the heap.
      Address 0x4b64dc is 4 bytes into a malloc'd block at 0x4b64d8 of 28 bytes.
      This block was allocated from:
            malloc         [rtlib.o]
            __bUiLtIn_nEw  [libgcc.a]
            __builtin_new  [rtlib.o]
            MathedXIter::adjustVerticalSt() [math_iter.C:940]
            MathMatrixInset::Metrics() [math_inset.C:414]
            MathedCursor::Insert(unsigned char,MathedTextCodes) [math_cursor.C:354]


Could this be related? Should the int array of MathedRowSt be cleared
by the constructor [it might be that your malloc already does that]?
Is this struct related to the display of an inset?

JMarc

Reply via email to