*But* I did not manage to have this crash the Writer, since there is a
fixed mempool for SwTxtNode object, which basically reserves some memory
for the paragraphs. Therefore most probably you won't get any access
violations.

As I said, it's very dependent on what happens to the SwIndexReg's
memory between the call to ~SwIndexReg and ~SwIndex. Of the 4 machines
I've tested this on only one crashes reliably (a machine with only
512MB RAM), and that crash only happens for a very specific usage
pattern (opening an external letterhead first and then opening the
Abtretungserklärung).

Is this a problem that is easily fixed? In what timeframe can we expect a fix?

And another question: Are you interested in other bug reports which we
can only reproduce with our letterhead system, even if we can't
provide information as detailed as for this one. I posted this on the
mailing list rather than opening an issue in the issue tracker,
because I feared that no one would look at the problem if reproducing
it required installing an extra module and the issue would hang around
in unconfirmed state forever.

If you are interested, we have other problems related to the dynamic
document construction performed by our letterhead system.
We have one strange problem for instance where a document is
constructed by our letterhead system and when you save it right away
as DOC, the resulting document crashes Word. But if you save as ODT
first, then terminate OOo, start it again and load the ODT and then
save it as DOC, the resulting document is okay. Note that just closing
the document and reopening it is not sufficient. OOo needs to be
terminated and restarted completely to recover. This again lets me
suspect that some internal data structure gets corrupted during the
dynamic construction of the document.  This is a grave issue for us,
because we need Word exchangeability and the workaround of "save as
ODT, terminate OOo completely, restart OOo, reopen document, export as
Word" is very hard to communicate to end users.

Unfortunately I am at a complete loss how to debug the above, because
it doesn't cause any visible symptoms in OOo. So I can't provide any
stack traces or other details except steps to reproduce using our
letterhead system. Is there a point in opening an issue for this or
posting it here on the mailing list? Or will this just be
closed/ignored with the remark "Please provide a minimal testcase that
does not require external modules" ?

Matthias

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

  • [sw-discussion... Matthias B.
    • Re: [sw-d... Frank Meies - Sun Germany - Development - Software Engineer
      • Re: [... Frank Meies - Sun Germany - Development - Software Engineer
        • R... Matthias B.
          • ... Andreas Martens
          • ... Frank Meies - Sun Germany - Development - Software Engineer

Reply via email to