optimize mapper builder

[mawen12]

Hi, I have found the XMLMapperBuilder contains two points can be optimized.

First: use method instead of parent variable

private void cacheElement(XNode context) {
  ...
  Class<? extends Cache> typeClass = typeAliasRegistry.resolveAlias(type);
  ...
  Class<? extends Cache> evictionClass = typeAliasRegistry.resolveAlias(eviction);
  ...
}

Its superclass is BaseBuilder which provide method resolveAlias(alias) to do this, may be can use method directly.

Second: use lambda to lazy compute

private Discriminator processDiscriminatorElement(XNode context, Class<?> resultType, List<ResultMapping> resultMappings) {
  ...
  String resultMap = caseChild.getStringAttribute("resultMap", processNestedResultMappings(caseChild, resultMappings, resultType));
  ...
}

The method getStringAttribute(value, default) return default only when value is null.
So when value is not null, default will be ignored,
By use lambda, default will be computed only when value is null.

[coveralls]

coverage: 87.18% (+0.009%) from 87.171%
when pulling 85bd9ba on mawen12:optimize-mapper-builder
into 266542a on mybatis:master.

[hazendaz]

@mawen12 What benefit does the first portion of this provide? The method doesn't actually provide any other value so it could be that we should get rid of that extra method instead or we make the fields private in all regard. I'm just not seeing the value of that by itself. The others, ok with. I did fix up the test as we want to avoid star imports.

[mawen12]

@mawen12 What benefit does the first portion of this provide? The method doesn't actually provide any other value so it could be that we should get rid of that extra method instead or we make the fields private in all regard. I'm just not seeing the value of that by itself. The others, ok with. I did fix up the test as we want to avoid star imports.

I agree with you. The first portion is due to my personal coding habit.
I will undo the changes to the first portion.

[hazendaz]

thanks. merging this.