Philip Martin <philip.mar...@wandisco.com> writes: > Daniel Shahaf <danie...@elego.de> writes: > >> Philip Martin wrote on Mon, Dec 10, 2012 at 17:49:44 +0000: >>> Daniel Shahaf <danie...@elego.de> writes: >>> >>> > Philip - perhaps you can move the relevant definitions to that header? >>> > Then I'll follow up with a transform_sql.py patch. >>> >>> OK, I've done that. I didn't annotate the maps or elements in any way. >> >> Implemented. Please note the question embedded in the r1419771 log >> message. (The code added in r1419817 might be relevant, too, in case >> both the node-kind and the depth maps have an "unknown" map member.) > > 'unknown' for depth corresponds to svn_depth_unknown while 'unknown' for > kind corresponds to svn_kind_unknown. As C enums these are different > values. > > You changed STMT_HAS_SPARSE_NODES to use the kind mapping for depth. > This will work since the numeric values are not involved, but it feels > wrong to me: we should introduce a svn_depth_t map.
The reason we don't currently have a depth map is that we are reusing svn_depth_from_word which converts command-line literals to an enum. I see no reason why we should not introduce a depth map and use svn_sqlite__column_token instead. -- Certified & Supported Apache Subversion Downloads: http://www.wandisco.com/subversion/download
