The MergeCitations batch tool had a cut-and-pasted code
replicating MergeCitationQuery logic. Naturally, it diverged :-)
Removed MergeCitations.Merge and use MergeCitationQuery instead.
svn: r22969
Refactor common base out of Citation, EventRef, and Repository
class IndirectCitationBase(object):
Citation management logic for objects that don't have citations
for the primary objects, but only for the child (secondary) ones.
The derived class must implement get_citation_child_list method
to return the list of child secondary objects that may refer
citations.
Note: for most objects, this functionality is inherited from
CitationBase, which checks both the object and the child objects.
Includes:
* has_citation_reference
* replace_citation_references
* remove_citation_references
svn: r22960
Cut-n-pastery of Repository functionality into Citation
to fix the bug pinpointed by merge_ref_test.SourceSourceCheck.
Refactoring to come next...
svn: r22953
0006848: citationtreeview corrupts when search bar open and edit occurs
0006829: Searching into source/citation selector disallows to select all related citations, only matching rows
by fixing treebasemodel so it properly deals with secondary objects, and also citationtreemodel so it adds secondary odes even when the parent is not present.
svn: r22941
Error merging citations containing objects with citations.
reapply from gramps40 the fix to MergeCitationQuery
merge_ref_test suite still fails :(
svn: r22938
Add to -h output new text about -y and -q
impex.sh switched to use --yes and --quiet
Refactor ArgHandler to reuse User object
ArgHandler now uses user.prompt
No longer custom code duplicating user.prompt functionality
This dropped support for English yes/no and prefixes in the
"OK to overwrite?", as User.prompt allows pressing "Enter"
to accept by default, and everything else except
verbatim accept choice will be treated as reject.
cli.user.User.prompt now supports treating EOF as a reject
prompt message reformatted: added newline after title
Previously, code
'-q' in ('--qml')
returned True, which was not what ArgParser meant.
Changed the rhs of in to [] from () to avoid this for every case
in ArgParser.parse in the future as well.
Tests run: the new UT added and impex.sh
svn: r22916
