@Newify table name pattern support #686
Conversation
…ression.isSynthetic()
…[reversed] is declared final but is reassigned :-)
…into bytecode)
// 2) shouldNotCompile checks that the Groovy compiler responds in the expected way to an attempt at assigning a value to a method parameter
… paramName, final List<String> classBodyTerms)
… extends AbstractASTTransformation works
…d if(node instanceof ClassNode)
# Conflicts: # src/main/groovy/transform/AutoFinal.java # src/main/org/codehaus/groovy/transform/AutoFinalASTTransformation.java
# Conflicts: # src/test/org/codehaus/groovy/transform/AutoFinalTransformBlackBoxTest.groovy
| @@ -196,39 +322,10 @@ private boolean shouldTransform(DeclarationExpression exp) { | |||
| } | |||
|
|
|||
| private boolean hasClassesToNewify() { | |||
| return classesToNewify != null && !classesToNewify.getExpressions().isEmpty(); | |||
| return (classesToNewify != null && !classesToNewify.getExpressions().isEmpty()) || (classNamePattern != null); // 2018-04-07 TODO: Check if classNamePattern != null can be made more "early out" | |||
There was a problem hiding this comment.
I see TODO here, are you going to refine it later?
There was a problem hiding this comment.
This TODO refers to e.g. an edge case optimization to defer the class name Pattern compilation to the point where is actually accessed. I don't expect this to make sense (also I am now checking against the class name pattern before checking against the class name list, since I expect this to be faster in practice), so I have not implemented it. Should probaly not be a TODO any more, but at most a regular comment, I can change this and redo the PR if you want.
| if(classType.getOuterClass() != null && ((classType.getModifiers() & org.objectweb.asm.Opcodes.ACC_STATIC) == 0)) { | ||
| if(!(argsExp instanceof ArgumentListExpression)) { | ||
| addError("Non-static inner constructor arguments must be an argument list expression; pass 'this' pointer explicitely as first constructor argument otherwise.", mce); | ||
| return mce; // TODO: Expected behavior to return the untransformed method call here ? |
There was a problem hiding this comment.
Similarly, I see TODO here, are you going to refine it later? ;-)
There was a problem hiding this comment.
This is the first Groovy extension I am doing on my own, and I took this code part from existing Groovy code, and wanted to make sure somebody more experienced had a look at this. If you say this code is fine, then the TODO can be removed (see above comment).
|
Hi Daniel, will do (I also added two more tests (for the import alias case) and removed the TODOs). |
If this set already contains the element, the call of `add` leaves the set unchanged
…gBuilder.append()' call
…or try/catch/finally
This reverts commit 2850721
mgroovy
force-pushed
the
mgroovy_newify_pattern
branch
from
9f3c841
to
2035a79
Compare
|
As the PR #689 supersedes the PR. Please close it. Thanks. |
|
(obsolete since superseded by #689 => closing) |
Added class name pattern parameter support to
@Newifyannotation (e.g.@Newify(pattern=/[A-Z].*/)) that allows users to define the classes with newify support through the pattern in addition to the existing class list parameter.