OSX uiwriter stacktrace needed

[Bjoern Michaelsen]

Hi,

On Wed, Jun 10, 2015 at 09:12:32AM +0100, Caolán McNamara wrote:
> On Wed, 2015-06-10 at 08:40 +0200, Stephan Bergmann wrote:
> > > Assertion failed: (!m_pFirst && !m_pLast && "There are still indices registered"), function ~SwIndexReg, file sw/source/core/bastyp/index.cxx, line 226.
> 
> I imagine that reset on the pCursor before dispose would clear that
> assert. But then at that point the test presumably becomes pointless.

I checked with the attached patch against
master at a861a234b0f5f39b06fd6180655182c167c3a94d and get:
> CPPUNITTRACE="gdb --args" make CppunitTest_sw_uiwriter debug=T
> ...
> debug:4359:1: DidTestCleanup: Document disposing
> debug:4359:1: NodeIndex at 0
> debug:4359:1: Position at 0
> debug:4359:1: UnoCursorPointer is disposed.
> ...

assuming it to be the same on OSX, pCursor should be innocent -- unless we leak
an SwIndex (on OSX only?!?).

Best,

Bjoern
-------------- next part --------------
A non-text attachment was scrubbed...
Name: didtestcleanup.diff
Type: text/x-diff
Size: 3242 bytes
Desc: not available
URL: <http://lists.freedesktop.org/archives/libreoffice/attachments/20150610/4505c919/attachment.diff>