We have 3 options here.
1. Use the proposed fix
2. Use another fix (revert the bug fix that revealed this)
3. Do nothing.
I really think #3 is terrible option. With this bug, users will complete a successful install only to have an unusable Eclipse install. That leaves 1 or 2. The Eclipse and Equinox teams have decided that 1 is better than 2. Pascal outlined this too .
The reasons are
- #1 is already done and tested
- The code change required for #1 has very high automated test coverage
- This fixes the underlying problem, resulting in a better 'Eclipse'
I realize it's late in Europe and most North American's have been busy today with the hockey game, but it would be great to hear from everyone with a +/- 1.