On Thu, Mar 3, 2011 at 3:27 PM, <[email protected]> wrote:
This works out great for TypeOracleMediatorFromByteCodeTest, but when
we
compile from source in TypeOracleMediatorFromSourceTest, the place we need to get in to modify the Type Oracle instance is nested inside of CompilationState which is also nested inside of
CompilationStateBuilder.
That means disruption to these other two classes. Seems like the
test
would be even more intrusive if I did that - what do you think?
I have a solution for you! In TypeOracleMediatorFromSourceTest.buildTypeOracle(), you have a chance to mess with the type oracle before you run the checks. So here is where you'd inject your call to typeOracle.setupHackyAlternateJsoClassForTesting(...). Have the implementation of setupHackyAlternateJsoClassForTesting() do http://gwt-code-reviews.appspot.com/1369805/diff/3006/dev/core/test/com/google/gwt/dev/javac/TypeOracleMediatorTestBase.java File dev/core/test/com/google/gwt/dev/javac/TypeOracleMediatorTestBase.java (right): http://gwt-code-reviews.appspot.com/1369805/diff/3006/dev/core/test/com/google/gwt/dev/javac/TypeOracleMediatorTestBase.java#newcode219 dev/core/test/com/google/gwt/dev/javac/TypeOracleMediatorTestBase.java:219: // caching strategy that involves the TypeOracle. Use multiline comment. http://gwt-code-reviews.appspot.com/1369805/diff/3006/dev/core/test/com/google/gwt/dev/javac/mediatortest/JavaScriptObjectCommonInterfaceImpl.java File dev/core/test/com/google/gwt/dev/javac/mediatortest/JavaScriptObjectCommonInterfaceImpl.java (right): http://gwt-code-reviews.appspot.com/1369805/diff/3006/dev/core/test/com/google/gwt/dev/javac/mediatortest/JavaScriptObjectCommonInterfaceImpl.java#newcode1 dev/core/test/com/google/gwt/dev/javac/mediatortest/JavaScriptObjectCommonInterfaceImpl.java:1: package com.google.gwt.dev.javac.mediatortest; Oh yeah, don't forget copy headers. http://gwt-code-reviews.appspot.com/1369805/ -- http://groups.google.com/group/Google-Web-Toolkit-Contributors
