http://gwt-code-reviews.appspot.com/1045801/diff/11001/12003 File dev/core/src/com/google/gwt/core/ext/linker/PrecompilationMetricsArtifact.java (right):
http://gwt-code-reviews.appspot.com/1045801/diff/11001/12003#newcode40 dev/core/src/com/google/gwt/core/ext/linker/PrecompilationMetricsArtifact.java:40: this.permutationId = permutationId; On 2010/10/25 23:55:47, scottb wrote:
Based on the call sites, I think this is permutation base. Sharded
precompiles
can handle more than 1 permutation per precompile shard.
renamed to permutation base and created a list of permutation ids as a second attribute http://gwt-code-reviews.appspot.com/1045801/diff/11001/12004 File dev/core/src/com/google/gwt/core/linker/CompilerMetricsLinker.java (right): http://gwt-code-reviews.appspot.com/1045801/diff/11001/12004#newcode43 dev/core/src/com/google/gwt/core/linker/CompilerMetricsLinker.java:43: @LinkerOrder(Order.POST) On 2010/10/25 23:55:47, scottb wrote:
I think it doesn't touch <pre> stuff unless it recognizes it as Java
code, in
which case there's an option to format the embedded code.
Done. (See new CompilerMetricsXmlFormatter.java) http://gwt-code-reviews.appspot.com/1045801/diff/11001/12012 File dev/core/src/com/google/gwt/dev/jjs/JavaToJavaScriptCompiler.java (right): http://gwt-code-reviews.appspot.com/1045801/diff/11001/12012#newcode637 dev/core/src/com/google/gwt/dev/jjs/JavaToJavaScriptCompiler.java:637: precompilationMetrics.setAstTypes(result.getReferencedJavaClasses()); On 2010/10/25 23:55:47, scottb wrote:
Sorry, the point I should have stated more clearly is that you
actually have the
AST already, in scope, in "jprogram". So there's really no reason to
loop this
through UnifiedAst at all, you can just collect the referenced classes
inline. Done. http://gwt-code-reviews.appspot.com/1045801/diff/33001/34002 File dev/core/src/com/google/gwt/core/ext/linker/ModuleMetricsArtifact.java (right): http://gwt-code-reviews.appspot.com/1045801/diff/33001/34002#newcode44 dev/core/src/com/google/gwt/core/ext/linker/ModuleMetricsArtifact.java:44: this(SoycReportLinker.class, nextInstance.getAndAdd(1)); On 2010/10/26 00:09:54, scottb wrote:
Nit: there's a "getAndIncrement()" for the +1 case.
Done. http://gwt-code-reviews.appspot.com/1045801/diff/33001/34004 File dev/core/src/com/google/gwt/core/linker/SoycReportLinker.java (right): http://gwt-code-reviews.appspot.com/1045801/diff/33001/34004#newcode117 dev/core/src/com/google/gwt/core/linker/SoycReportLinker.java:117: artifacts = new ArtifactSet(artifacts); On 2010/10/26 00:09:54, scottb wrote:
Nit, but the old formulation was okay because it conditionally created
a new
artifact set only when there was something to update, relying on
returning the
new artifact set. If you would rather unconditionally replace the old
set
(unlikely to be much of a perf hit), then all the helpers can just
return void
and you return the original set at the bottom; makes the data flow
easier to
eyeball.
Done. http://gwt-code-reviews.appspot.com/1045801/show -- http://groups.google.com/group/Google-Web-Toolkit-Contributors
