On Thu, 29 Nov 2001, John Levon wrote:
> Start a new doc, insert a footnote, go to Documents and select the same doc -> crash.
>
> Because colapsable::edit() doesn't consider that the inset may be already locked, it
>ends
> up trying to lock an inset at pos 0 *inside* the footnote, thinking that
>bv->theLockingInset()
> exists inside the footnote.
>
> Patch to "fix" the crash is below, but first some questions.
This patch also "fixes" the crash I reported the other day triggered
by being inside an inset and opening/closing another document. An
interesting point here is that the cursor is always placed at the
start of the inset irrespective of where it was when I changed buffers
(even if it was inside an inset in an inset).
> 1) shouldn't ::buffer() check to see if we are really switching buffers :
>
> if (buffer == b) {
> return;
> }
Isn't this what it used to do at some stage?
> 2) how does the slept inset stay locked between the insetSleep and the insetWakeup ?
>I can't
> see why the inset doesn't unlock properly in insetSleep ...
>
> 3) insettext and insettabular also have this problem (turn on -dbg insets to see !),
>but the
> locking_inset stuff there is FAR too confusing for me to understand. I think they
>survive by
> "the skin of their teeth".
Good questions. I know less about this stuff than you do.
Allan. (ARRae)
