Community
Participate
Working Groups
For tracking the change on 3.7.2 maintenance branch, Cloned from: 350014: [transport] intermittent timeout during downloading artifacts should be forgivable https://bugs.eclipse.org/bugs/show_bug.cgi?id=350014 P2 and ECF treats the socket timeout as the fatal error, it never tries to reconnect the server and get the data again. However the intermittent timeout is probably happened when downloading a lot of artifacts from remote server, the p2 installation(collect phase) will fail if the timeout occurs only one time. Comparing to other famous download utility, say as 'wget', it will try again(the number is configurable) after the read timeout happens. I'm not sure which component(p2/ecf) should take the responsibility to retry the downloading after timeout occurs. I would like to own this one if you guys think it's a p2's problem.
Hi Pascal, Do we have special rules to control defect fixes for maintenance build? I would like to merge this one into 3.7.2. The fix is safe, and it's disable by default.
released and tagged on R3_7_maintenance branch http://git.eclipse.org/c/equinox/rt.equinox.p2.git/log/?h=R3_7_maintenance