Fix jar loading issue under Windows.

Special case Windows.

Author: ghislainfourny

README.md

@@ -338,6 +338,9 @@ Even more queries can be found [here](https://colab.research.google.com/github/R
 
 # Latest updates
 
+## Version 2.0.9
+- Solve an issue under Windows that caused the error "Java gateway process exited before sending its port number.".
+
 ## Version 2.0.8
 - Decoupled the internal materialization cap (when a parallel sequence of items is materialized, e.g., into an array) from the outer result size cap (for printing to screen) with now two distinct configuration parameters. The default materialization cap is set to 100'000 items while the default outer result size is set to 10. They can be changed by the user through the Rumble configuration.
 - Fixed an issue in the implementation when a FLWOR gets executed locally with a return clause with an underlying RDD or DataFrame. 

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "jsoniq"
-version = "2.0.8"
+version = "2.0.9"
 description = "Python edition of RumbleDB, a JSONiq engine"
 requires-python = ">=3.11"
 dependencies = [

src/jsoniq/session.py

@@ -8,7 +8,11 @@
 import importlib.resources as pkg_resources
 
 with pkg_resources.path("jsoniq.jars", "rumbledb-2.0.8.jar") as jar_path:
-    jar_path_str = "file://" + str(jar_path)
+    if (os.name == 'nt'):
+        jar_path_str = str(jar_path)
+    else:
+        jar_path_str = "file://" + str(jar_path)
+    print(f"[Info] Using RumbleDB jar file at: {jar_path_str}")
 
 def get_spark_version():
     if os.environ.get('SPARK_HOME') != None: