A solution whereby we run boost tests separately from non-boost tests.

tools/run_desktop_tests.py

@@ -61,21 +61,24 @@ class TestRunner:
         parser.add_option("-u", "--user_resource_path", dest="resource_path", help="Path to resources, styles and classificators (passed to the test executables as --user_resource_path=<value>)")
         parser.add_option("-i", "--include", dest="runlist", action="append", default=[], help="Include test into execution, comma separated list with no spaces or individual tests, or both. E.g.: -i one -i two -i three,four,five")
         parser.add_option("-e", "--exclude", dest="skiplist", action="append", default=[], help="Exclude test from execution, comma separated list with no spaces or individual tests, or both. E.g.: -i one -i two -i three,four,five")
-        
+        parser.add_option("-b", "--boost_tests", dest="boost_tests", action="store_true", default=False, help="Treat all the tests as boost tests (their output is different and it must be processed differently.")
+
         (options, args) = parser.parse_args()
 
         self.skiplist = set()
         self.runlist = list()
-        
+
         for tests in options.skiplist:
             for test in tests.split(","):
                 self.skiplist.add(test)
         
         for tests in options.runlist:
             self.runlist.extend(tests.split(","))
-            
+
+        self.boost_tests = options.boost_tests
+
         if self.runlist:
-            logging.warn("-i option found, the -e option will be ignored")
+            logging.warn("-i or -b options found, the -e option will be ignored")
         
         self.workspace_path = options.folder
         self.logfile = options.output
@@ -127,6 +130,18 @@ class TestRunner:
         return {TO_RUN:tests_to_run, SKIP:local_skiplist, NOT_FOUND:not_found, WITH_SERVER:tests_with_server}
         
 
+    def test_file_with_keys(self, test_file):
+        # omim-build-debug/out/debug/opening_hours_tests --report_format=xml --report_level=detailed --log_level=test_suite --log_format=xml 1> boostlog.xml 2> boosterr.xml
+
+        if self.boost_tests:
+            return "{test_file}  --report_format=xml --report_level=detailed --log_level=test_suite --log_format=xml {data}{resources}".format(test_file=test_file, data=self.data_path, resources=self.user_resource_path)
+
+        return "{test_file}{data}{resources}".format(test_file=test_file, data=self.data_path, resources=self.user_resource_path)
+
+
+        pass
+
+
     def run_tests(self, tests_to_run):
         failed = list()
         passed = list()
@@ -135,7 +150,9 @@ class TestRunner:
             
             self.log_exec_file(test_file)
 
-            test_file_with_keys = "{test_file}{data}{resources}".format(test_file=test_file, data=self.data_path, resources=self.user_resource_path)
+
+
+            test_file_with_keys = self.test_file_with_keys(test_file)
         
             logging.info(test_file_with_keys)
             process = subprocess.Popen("{tests_path}/{test_file} 2>> {logfile}".
@@ -156,6 +173,9 @@ class TestRunner:
 
 
     def log_exec_file(self, filename, result=None):
+        if self.boost_tests:
+            return
+
         logstring = "BEGIN" if result is None else "END"  #can be 0 or None. If we omit the explicit check for None, we get wrong result
         resstring = (" | result: {returncode}".format(returncode=result) if result is not None else "")
         with open(self.logfile, "a") as logf: