Fix: Weigh programming language tallies by file size (bytes) instead … Fixes #3868

src/scancode/interrupt.py

@@ -86,8 +86,15 @@ def handler(signum, frame):
             raise TimeoutError
 
         try:
-            create_signal(SIGALRM, handler)
-            setitimer(ITIMER_REAL, timeout)
+            # We try to setup the signal. If we are not in the main thread
+            # this will raise a ValueError. In this case we just run the
+            # function without timeout.
+            try:
+                create_signal(SIGALRM, handler)
+                setitimer(ITIMER_REAL, timeout)
+            except ValueError:
+                pass
+
             return NO_ERROR, func(*(args or ()), **(kwargs or {}))
 
         except TimeoutError:
@@ -97,7 +104,10 @@ def handler(signum, frame):
             return ERROR_MSG + traceback_format_exc(), NO_VALUE
 
         finally:
-            setitimer(ITIMER_REAL, 0)
+            try:
+                setitimer(ITIMER_REAL, 0)
+            except ValueError:
+                pass
 
 elif on_windows:
     """
@@ -191,4 +201,4 @@ def fake_interruptible(func, args=None, kwargs=None, timeout=DEFAULT_TIMEOUT):
     try:
         return NO_ERROR, func(*(args or ()), **(kwargs or {}))
     except Exception:
-        return ERROR_MSG + traceback_format_exc(), NO_VALUE
+        return ERROR_MSG + traceback_format_exc(), NO_VALUE

src/summarycode/summarizer.py

@@ -176,8 +176,12 @@ def get_declared_holders(codebase, holders_tallies):
         if entry['holder']
     }
     unique_key_file_holders = unique(entry_by_key_file_holders.keys())
+
+    # FIX: Added 'if holder in entry_by_holders' to prevent crash if a holder
+    # exists in key files but was filtered out of the main tallies (0 byte count)
     unique_key_file_holders_entries = [
         entry_by_holders[holder] for holder in unique_key_file_holders
+        if holder in entry_by_holders
     ]
 
     holder_by_counts = defaultdict(list)
@@ -195,7 +199,10 @@ def get_declared_holders(codebase, holders_tallies):
     # If we could not determine a holder, then we return a list of all the
     # unique key file holders
     if not declared_holders:
-        declared_holders = [entry['value'] for entry in unique_key_file_holders_entries]
+        # We must also filter here to avoid crashing on missing entries
+        declared_holders = [
+            entry['value'] for entry in unique_key_file_holders_entries
+        ]
 
     return declared_holders