> 2) Unfortunately, I hadn't taken enough care to note the disposition of > AXIS2-4751. Now that Jarek's current changes are on the branch, the built > code will have regressions for several earlier issues (search for "axis2 > close_wait" - AXIS2-2883 is a good example) - just confirmed this on Windows. > We have to fix this before the release, since this was the whole reason for > 1.5.1 in the first place. My proposal for now would be to revert Jarek's > change and re-test the branch with Rampart. If the 1.5.1-style code works, > then we'll need to figure another way to fix AXIS2-4751. If not, though, we > go from there and try re-applying r951718, finding another way to fix the > Rampart issue.
Glen, I disabled connection reuse because it was causing weird problems with Rampart. If you enable connection caching with my changes (one line change) or revert my changes all together I believe the results will be the same. I think there are other problems in Axis2 or Rampart that cause Rampart failures when connection caching is done. The problem with the old connection reuse code that it wasn't quite thread safe. Jarek --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
