Add support for filter callback

README.md

@@ -18,7 +18,7 @@ Runs test items. This will re-use client processes from a previous call to `run_
 
 Args:
 - `path`: Filesystem path to a folder.
-- `filter`: Not implemented right now.
+- `filter`: A filter callback function that will be called for each identified test item. If the filter callback returns `true` that test item will be run, if `false` it will not run. Each call to the provided filter callback passes a named tuple argument with a number of fields that contain metadata about the specific test item. The provided information is `filename`, `name`, `tags` and `package_name`.
 - `verbose` Not implemented right now.
 - `max_workers`: Max number of child processes per identified project.
 - `timeout`: Timeout in seconds.

src/TestItemRunner2.jl

@@ -251,10 +251,10 @@ function run_tests(path; filter=nothing, verbose=false, max_workers::Int=Sys.CPU
     # Flat list of @testitems
     testitems = [(detail=i, code=get_text(jw._text_documents[i.uri])[i.code_range]) for i in Iterators.flatten(values(jw._testitems))]   
 
-    # # Filter @testitems
-    # if filter !== nothing
-    #     filter!(i->filter((filename=uri2filepath(i.detail.uri), name=i.detail.name, tags=i.detail.option_tags)), testitems)
-    # end
+    # Filter @testitems
+    if filter !== nothing
+        filter!(i->filter((filename=uri2filepath(i.detail.uri), name=i.detail.name, tags=i.detail.option_tags, package_name=i.detail.package_name)), testitems)
+    end
 
     executed_testitems = []