On Sun, May 9, 2010 at 18:21, Bert Huijben <b...@qqmail.nl> wrote: >... >> From: Stefan Sperling [mailto:s...@elego.de] >... >> > As Bert explained, we need to remove them when the user runs "svn >> > resolved". He also noted that (somtimes) it is possible manually >> > resolve a conflict by removing all the conflict files (a potentially >> > debatable feature). >> >> I see. Then let's just add another field to the skel. >> I guess we can store this within the conflict-type-specific data?
I believe so. It doesn't seem to apply to the OPERATION sub-skel. Each type of conflict has a different set of in-working-copy files that it may need to record. I'll update conflict-storage "soon". Not really there right now. >> Storing the basename should be enough since we can assume the file >> will be put into the same directory as the conflicted file, right? > > For text conflicts this would be the case, but for a property conflict on a > directory it would be harder to tell where the file is located. (What if the > directory is missing?). Greg suggested adding a wcroot-relative path on IRC. Right. The wcroot is definitely known, and any path is reachable/computable from there. We don't need to worry about "basename relative to WHAT directory?" I've written two functions: svn_wc__db_to_relpath() and svn_wc__db_from_relpath(). These can/should be used for all abspath/relpath conversions within libsvn_wc when a path needs to be persisted. Using them for *other* purposes is verboten. I'll migrate the workqueue path persistence over to these RSN. Cheers, -g