On Tue, Oct 18, 2011 at 7:20 PM, sebb <seb...@gmail.com> wrote: > On 18 October 2011 17:53, Philip Martin <philip.mar...@wandisco.com> wrote: >> sebb <seb...@gmail.com> writes: >> >>> In that case, either that is an insufficient check, or the upgrade >>> fails to act correctly on the results of the check. >> >> In what way? The upgrade detected the problem, stopped the upgrade and >> left the 1.6 working copy unchanged apart from some files in .svn/tmp. > > I've obviously misunderstood some of the postings then. > > I thought there were some reports (in other threads) of issues after > the upgrade completed.
Yes, there were also some reports about problems with using 1.7 *after a successful upgrade* (the one with "checksum != NULL" [1], and the one with "sha1_checksum != NULL" [2]). AFAICT, those are currently being worked on, to improve the situation in 1.7.1 (either by handling this situation gracefully, or by detecting the problem during the upgrade process, and aborting the upgrade). But if a problem happens *during upgrade* (like the one reported here by the OP), and the upgrade gets aborted (or if you abort it manually), you will normally still have a working 1.6 working copy (the removal of the old .svn directories is the last step in the upgrade -- as long as those are still present, you still have a valid 1.6 WC). So it's possible to try and repair some things (be careful though not to lose any uncommitted stuff while doing so -- maybe best to make a backup copy), and retry the upgrade. Also, AFAICT 1.7.1 will give more information about which files exactly in your 1.6 WC are related to the corruption. So you might still be able to upgrade after repairing the corruption (or making it "disappear" by executing "svn up -r0 /the/corrupted/file" with your 1.6 client before retrying the upgrade). [1] http://subversion.tigris.org/issues/show_bug.cgi?id=4035 [2] http://subversion.tigris.org/issues/show_bug.cgi?id=4038 -- Johan