A repo with thousands of commits can take many hours to analyze (repo_refresh) after the initial import. What can we do to make it faster?
@@ -1,36 +1,2 @@
+A repo with thousands of commits can take many hours to analyze (repo_refresh) after the initial import. What can we do to make it faster?
-(engr) Name of Engineer pushing: Dave Brondsema
-(engr) Which code tree(s): Allura, forge-classic, sftheme
-(engr) Which release/revision is going to be synced?: allura_20121026
-(engr) Itemized list of changes to be launched with sync:
- * 8ea9518 make merge requests listing page not completely broken
- * 1eb2efb add exact search option for /p/upgrade for admins
- * Fix [#4942] (ux,projectnav) Change the "Downloads" tool to "Files" for consistency NEEDS CMD
- * Fix [#5048] (42cc) pagination links moved
- * Fix [#5077] (support,p3,42cc) Add confirmation when removing developer
- * Fix [#5081] (42cc) Mass edit should store a ticket revision
- * Fix [#5186] (migration) Deal with "Connection reset by peer" upgrade failures NEEDS MONGO CMD
-(engr) Pre-launch dependencies:
-Please summarize these upgrades and possible production.ini changes:
-(engr) Post-launch dependencies:
-(engr) Approved for release (Dave/Patrick): None
-(sog) Outcome of sync:
reverting content; not sure how it was changed
Many parts of an SVN refresh are skipped now (and fetched lazily on demand), so SVN refreshes of thousands of commits aren't a problem.