diffoscope/comparators/utils/compare.py

@@ -93,8 +93,9 @@ def compare_files(file1, file2, source=None, diff_content_only=False):
 
     # Specialize the files first so "has_same_content_as" can be overridden
     # by subclasses
-    specialize(file1)
-    specialize(file2)
+    with profile("specialize", "specialize"):
+        specialize(file1)
+        specialize(file2)
 
     force_details = Config().force_details
     with profile("has_same_content_as", file1):

diffoscope/diff.py

@@ -32,6 +32,7 @@ from multiprocessing.dummy import Queue
 
 from .tools import get_tool_name, tool_required
 from .config import Config
+from .profiling import profile
 from .tempfiles import get_temporary_directory
 
 DIFF_CHUNK = 4096
@@ -386,7 +387,7 @@ def color_unified_diff(diff):
 DIFFON = "\x01"
 DIFFOFF = "\x02"
 MAX_WAGNER_FISCHER_SIZE = (
-    1024  # any higher, and linediff takes >1 second and >200MB RAM
+    256  # any higher, and linediff takes >1 second and >200MB RAM
 )
 
 
@@ -450,6 +451,7 @@ def linediff_wagnerfischer(s, t):
     for j in range(1, n + 1):
         d[0][j] = (j, (0, j - 1))
 
+    # NB. This loop is O(len(s) * len(t))
     for i in range(1, m + 1):
         for j in range(1, n + 1):
             if s[i - 1] == t[j - 1]:
@@ -611,7 +613,8 @@ class SideBySideDiff:
             type_name = "unmodified"
         else:
             type_name = "changed"
-            s1, s2 = linediff(s1, s2, self.diffon, self.diffoff)
+            with profile("diff", "linediff"):
+                s1, s2 = linediff(s1, s2, self.diffon, self.diffoff)
 
         yield "L", (type_name, s1, self.line1, s2, self.line2)
 

diffoscope/presenters/formats.py

@@ -16,9 +16,11 @@
 # You should have received a copy of the GNU General Public License
 # along with diffoscope.  If not, see <https://www.gnu.org/licenses/>.
 
+import os
 import logging
 
 from ..profiling import profile
+from ..utils import format_bytes
 
 from .text import TextPresenter
 from .json import JSONPresenter
@@ -102,6 +104,17 @@ class PresenterManager:
             with profile("output", name):
                 data["klass"].run(data, difference, parsed_args)
 
+            size = "n/a"
+            if os.path.isfile(data["target"]):
+                size = format_bytes(os.path.getsize(data["target"]))
+
+            logger.debug(
+                "Generated %r output at %r (size: %s)",
+                name,
+                data["target"],
+                size,
+            )
+
     def compute_visual_diffs(self):
         """
         Don't waste time computing visual differences if we won't use them.