On 09/08/2011 12:04 AM, Hyrum K Wright wrote: > At this stage in the release process, my primary concern is "are these > issues a regression from 1.6?" File externals in 1.6 were a hack; are > we making things worse in 1.7? If things are on parity with 1.6, I > don't see this as being anything close to a release blocker.
Comparing 1.6.17 to 1.7.x... - Issue #4000 is a regression. (I had nominated some revisions for that, but removed that nomination until Philip's find has also been fixed.) - Issue #4002 is a regression in that it shows an error message, fix is r1164027 (was included in #4000's nomination, now has its own). - file external in unversioned: 1.6: [[[ Fetching external item into 'UNVER/a' svn: warning: 'UNVER' is not a working copy <no external brought in> ]]] 1.7: [[[ Fetching external item into 'UNVER/a': A UNVER/a Updated external to revision 1. At revision 1. + svn status UNVER/a svn: E155010: The node '/tmp/1.7_unver_fileextl.vJb/wc/UNVER' was not found. + svn ci -mm UNVER/a svn: E155010: Commit failed (details follow): svn: E155010: The node '/tmp/1.7_unver_fileextl.vJb/wc/UNVER' was not found. + svn up ... Fetching external item into 'UNVER/a': U UNVER/a Updated external to revision 3. ]]] If we leave it as-is, updating FEs (by recursion) works as expected, not much else. Should we reinstate the bail, disallowing file externals in unversioned dirs, until we have a complete implementation? Otherwise I know of no regressions. (Re-ran all my recent externals test scripts comparing 1.6.17 to 1.7.x.) Workin' on #4000. ~Neels
signature.asc
Description: OpenPGP digital signature
