On 18 October 2011 17:10, Daniel Shahaf <d...@daniel.shahaf.name> wrote: > sebb wrote on Tue, Oct 18, 2011 at 16:57:14 +0100: >> On 18 October 2011 16:33, Daniel Shahaf <d...@daniel.shahaf.name> wrote: >> > Les Mikesell wrote on Tue, Oct 18, 2011 at 10:18:24 -0500: >> >> I think there is a bigger question regarding why there are so many >> >> corrupt 1.6 workspaces around that nothing so far had noticed. Should >> >> people be concerned about that even if they aren't upgrading yet? >> >> How can you tell if a workspace is corrupt? >> > (you mean "working copy") >> > >> > Make a copy of it and upgrade that. >> > >> > Or, failing that, check all md5's in the working copy. But I expect the >> > former option is far simpler for most people. >> >> By checking MD5, do you mean calculate and compare the hashes for the >> working copy and corresponding base files? > > Yes, svn_wc_entry_t->checksum. > >> If so, is that something that the upgrade process could (should) do >> before making any changes? >> > > Upgrade already does that...
In that case, either that is an insufficient check, or the upgrade fails to act correctly on the results of the check. >> That would also be useful as a stand-alone tool. > > Could be. If I were to hack it I'd base it either on libsvn_wc's > entry-reading APIs or on change-svn-wc-format.py.
