>>>>> "Rich" == Richard Lowe <richlowe at richlowe.net> writes:
Rich> nightly.sh: Rich> 348: Rich> It'd be good to comment what the else branch is doing (support Rich> for copy_source'ing a subdirectory of usr/closed, I think). Okay. Rich> 309: Rich> It may be better to pass a pattern (or -I <dir>) to hg locate, Rich> rather than list the whole thing and grep for portions. I'll look into this. If I decide to keep the current grep structure, I'll comment why. Rich> 2322: Rich> It's kinda sad that we'll lose the version information if Rich> nightly is run from onnv-tools, rather than installed on the Rich> build machine, but I don't see a nice way around that other than Rich> a filter when 'building' it. We don't want to bake build information into the delivered script, to avoid false positives from wsdiff. Rich> nightly.1: Rich> 272: Rich> Internally, both $CODEMGR_WS and $CODEMGR_WS/usr/closed will Rich> contain .hg/ this should probably be explicit (as self evident Rich> as it may seem) I'll reword to clarify. Rich> 311: Rich> It may still (and I think should still) differ from the parent Rich> in the general case, since the clone will still exist (and as Rich> far as having nightly bring it over, the moving local tag Rich> solution doesn't work as you'd hope). Yeah, it occurred to me over the long weekend that the local tag approach would require additional changes to nightly.sh. So I'm happy to declare we're going to use a separate clone repo. Anyone object? Thanks for the comments. mike
