Bugzilla – Full Text Bug Listing
| Summary: | build.stamp should be propagated to sub-builds | ||
|---|---|---|---|
| Product: | [RT] Virgo | Reporter: | Glyn Normington <glyn.normington> |
| Component: | virgo-build | Assignee: | Steve Powell <zteve.powell> |
| Status: | CLOSED FIXED | QA Contact: | |
| Severity: | normal | ||
| Priority: | P3 | CC: | eclipse, zteve.powell |
| Version: | 2.2.0.M01 | ||
| Target Milestone: | 3.0.0.M01 | ||
| Hardware: | PC | ||
| OS: | Mac OS X - Carbon (unsup.) | ||
| Whiteboard: | |||
|
Description
Glyn Normington
Embarrassing problem here:
When full process-isolated sub builds were initiated (to overcome ivy permgen leaks) we had to explicitly pass properties to the sub build (cannot use the subant task mechanisms).
We failed to pass enough parameters -- in particular we (originally) omitted to pass ${timestamp} so the sub-project builds got a different bundle version (including a timestamp) from each other and from the top-level build.
This was fixed, but then we discovered that 'ripplor' builds failed to generate the correct build.stamp for the sub-project builds -- this was "fixed" by passing ${build.stamp} but unfortunately propagated before testing (hard to do with a ripple) and it wasn't a fix at all!
Finally we passed ${bundle.version} which turned out to be the correct parameter for a ripple. I ought to check that ordinary builds work properly, too.
. . .
. . . OK local builds still OK. virgo-build/ version 1.38 propagated to master branches, and to 2.1.x and 1.1.x branches. Ripple on master (from util) is running OK now. Fixed |