commit 9dda6c1f8aa75f324509a06cef826eb3239bf794
author scottb@google.com <scottb@google.com@8db76d5a-ed1c-0410-87a9-c151d255dfc7> Tue May 12 00:08:12 2009 +0000
committer scottb@google.com <scottb@google.com@8db76d5a-ed1c-0410-87a9-c151d255dfc7> Tue May 12 00:08:12 2009 +0000
tree d5b020c73f975613eeb1f0f0412a781ebd96d919
parent d241edb03ef29d5da143d7737cb332de90200bf5

During precompile, only do a draft optimize due to superlinear optimization performance.

This makes precompile faster at the expense of each sharded compile being slower.  This is a win if youc can properly parallelize the individual compile shards, but we should revisit this change for users who only have a single build machine, as it could be a net loss for them.

This also makes draft compiles (as a feature) faster by doing even fewer optimizations.

Review by: spoon

git-svn-id: https://google-web-toolkit.googlecode.com/svn/trunk@5338 8db76d5a-ed1c-0410-87a9-c151d255dfc7