| commit | 79fdc6e0a3a44642ac8d49225f14241872097fb7 | [log] |
|---|---|---|
| author | Caolán McNamara <caolanm@redhat.com> | Mon Dec 09 16:52:54 2019 +0000 |
| committer | Michael Stahl <michael.stahl@cib.de> | Tue Dec 10 09:44:40 2019 +0100 |
| tree | d916d8fb894e353dc10f69140eea0603e533d013 | |
| parent | bd0764d4d24f3a9f5fd279db888d12b55bd6b241 [diff] |
tdf#129210 don't assert with clear on already cleared mutex reset it after load completes, seeing as the comment suggests that's why the clear exists in the first place Change-Id: I3af633c5a1f99b5d0dd754c21e220c6c94b66634 Reviewed-on: https://gerrit.libreoffice.org/84789 Tested-by: Jenkins Reviewed-by: Michael Stahl <michael.stahl@cib.de>
diff --git a/dbaccess/source/core/dataaccess/databasedocument.cxx b/dbaccess/source/core/dataaccess/databasedocument.cxx index 9f72c08..9cd69e9 100644 --- a/dbaccess/source/core/dataaccess/databasedocument.cxx +++ b/dbaccess/source/core/dataaccess/databasedocument.cxx
@@ -685,6 +685,7 @@ void SAL_CALL ODatabaseDocument::recoverFromFile( const OUString& i_SourceLocati aGuard.clear(); // (load has an own guarding scheme) load( aMediaDescriptor.getPropertyValues() ); aGuard.reset(); // Without a controller, we are unable to recover the sub components, as they're always tied to a controller. // So, everything else is done when the first controller is connected.