On Thu, Apr 25, 2002 at 12:13:06AM +0200, Claus Hindsgaul wrote: > Or perhaps an optimized executable...
Still a valgrind bug ! Or binutils, or gcc ... anyway > I have recompiled LyX without optimizations, deleting all "-O" options > in the configure file (since 'configure --disable-optimizations' did not > seem to do what it is supposed to do!) IT doesn't remove -O from C files, but that doesn't matter. If it's not removing -O from c++ files I want to know :) > I think the new valgrind output looks more sane (attached) > ==20545== at 0x81044A6: Row::previous(void) const (lyxrow.C:112) > ==20545== by 0x813296B: LyXText::drawInset(LyXText::DrawRowParams &, int) >(text.C:494) > ==20545== by 0x8133427: LyXText::draw(LyXText::DrawRowParams &, int &) >(text.C:649) > ==20545== by 0x8141CCC: LyXText::paintRowText(LyXText::DrawRowParams &) >(text.C:3665) > ==20545== Address 0x42FBFA6C is 32 bytes inside a block of size 36 free'd > ==20545== at 0x4003D312: __builtin_delete (vg_clientfuncs.c:171) > ==20545== by 0x81439D9: LyXText::removeRow(Row *) const (text2.C:376) > ==20545== by 0x813797F: LyXText::breakAgain(BufferView *, Row *) const >(text.C:1653) > ==20545== by 0x8149FF5: LyXText::checkParagraph(BufferView *, Paragraph *, int) >(text2.C:1956) Jug ? > ==20545== Invalid read of size 4 > ==20545== at 0x81044B6: ??? (lyxrow.h:92) Hmm, valgrind is odd sometimes ... regards john