Adjust MethodNameCasing to work with C#
#397
Conversation
![github-actions[bot]](https://avatars.githubusercontent.com/in/15368?s=60&v=4){kind=small}
src/main/java/org/openrewrite/staticanalysis/nameconvention/CsharpNameConvention.java
Outdated
Show resolved
Hide resolved
src/main/java/org/openrewrite/staticanalysis/nameconvention/JavaNameConvention.java
Outdated
Show resolved
Hide resolved
src/main/java/org/openrewrite/staticanalysis/nameconvention/NameConvention.java
Outdated
Show resolved
Hide resolved
src/main/java/org/openrewrite/staticanalysis/nameconvention/NameConventionFactory.java
Outdated
Show resolved
Hide resolved
src/main/java/org/openrewrite/staticanalysis/nameconvention/NameConventionFactory.java
Outdated
Show resolved
Hide resolved
|
Possibly we should leverage the service mechanism we have in JavaSourceFile. We should pull that up API to SourceFile anyway one of these days. Just a thought. |
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
It actually already is, so the generic interface could be called something like |
src/main/java/org/openrewrite/staticanalysis/MethodNameCasing.java
Outdated
Show resolved
Hide resolved
| } | ||
|
|
||
| String toName = standardized.toString(); | ||
| NamingService service = getCursor().firstEnclosing(SourceFile.class).service(NamingService.class); |
There was a problem hiding this comment.
The JavaVisitor in fact has a service() method which is much easier to use. We might want to pull that up to TreeVisitor at some point.
We might also have to cache the services a bit more as "breadcrumbs" along the cursor path, so that the service lookups don't get too expensive for deeply nested elements. But that is also something for another day.
src/main/java/org/openrewrite/staticanalysis/MethodNameCasing.java
Outdated
Show resolved
Hide resolved
Laurens-W
changed the title
MethodNameCasing to work with C#
Also move checks to NamingService
| String toName = service.standardizeMethodName(normalized); | ||
| if (!StringUtils.isBlank(toName) && !StringUtils.isNumeric(toName) && | ||
| !methodExists(method.getMethodType(), toName)) { |
There was a problem hiding this comment.
standardizeMethodName is not guaranteed to give back a different method name; do we already have a test to check expecting no changes? Or should we add a comparison here between the old and new name?
What's changed?
Introduce interfacing and factory pattern to dynamically retrieve a
NamingConventionto be applied for given sourcefileWhat's your motivation?
Introduction of C# as a supported language
Anyone you would like to review specifically?
@timtebeek @OlegDokuka @knutwannheden
Checklist