"Bert Huijben" <b...@qqmail.nl> writes: >> -----Original Message----- >> From: MARTIN PHILIP [mailto:codematt...@ntlworld.com] On Behalf Of Philip >> Martin >> Sent: woensdag 11 april 2012 16:33 >> To: Julian Foad >> Cc: SubversionDevelopment >> Subject: Re: docco improvement for svn_wc__db_scan_deletion()
>> To be honest I'm not sure exactly what this function is supposed to do >> with multi-layer-move, or whether what it does at the moment is correct. >> Consider: >> >> svn mv A/f A/g >> svn mv A B >> >> select op_depth, local_relpath, presence, moved_to, moved_here >> from nodes order by op_depth, local_relpath" >> >> 0||normal|| >> 0|A|normal|| >> 0|A/f|normal|| >> 1|A|base-deleted|B| >> 1|A/f|base-deleted|| >> 1|B|normal||1 >> 1|B/f|normal||1 >> 2|B/f|base-deleted|B/g|0 >> 2|B/g|normal||1 >> >> What are MOVED_TO_ABSPATH and MOVED_TO_OP_ROOT_ABSPATH for A/f? > > "The world was much simpler when we stored moved_from only at op_depth > 0/BASE." I assume you mean moved_to rather than moved_from. > The original WC-NG model only allowed moving files once. What is "original"? BASE+WORKING? NODES with op_depth=0 for moved_to? > The whole moving of files that aren't committed yet model is not well > defined (yet). And I'm not entirely sure if it can be well-defined > with cases like this. I see problems with the scan_deletion API. I suppose that might indicate a problem with the data model, but it may simply be that the query API is wrong. -- uberSVN: Apache Subversion Made Easy http://www.uberSVN.com
