tests: Strip before garbage check in test_parser_noncontiguous_leading

Hypothesis brought up the following example:

    @example(subs=[srt.Subtitle(index=0, start=timedelta(0), end=timedelta(days=21186, seconds=33261), content='0', proprietary='')],
        garbage='\x1f0',
    )

This is a bit confusing, the real characters are "\x1f" and "0". Since
we will strip inside due to using \s* in the regex, this just appears as
"0" to the parser.

tests/test_srt.py

@@ -476,8 +476,8 @@ def test_parser_noncontiguous_leading(subs, garbage):
     # index...
     assume(garbage.strip()[0] != ".")
     assume(garbage.strip()[0] != "-")
-    assume(not _parseable_as_int(garbage))
-    assume(not _parseable_as_float(garbage))
+    assume(not _parseable_as_int(garbage.strip()))
+    assume(not _parseable_as_float(garbage.strip()))
 
     # Put some garbage at the beginning that should trigger our noncontiguity
     # checks