regression-hotspots: improve output, exclusions, comments and names

Change-Id: If548e67b7f39b24ccfc78c5d76d18f8fdc90fd80
Reviewed-on: https://gerrit.libreoffice.org/c/dev-tools/+/166021
Tested-by: Ilmari Lauhakangas <ilmari.lauhakangas@libreoffice.org>
Reviewed-by: Ilmari Lauhakangas <ilmari.lauhakangas@libreoffice.org>

1 files changed, 9 insertions, 5 deletions

diff --git a/scripts/regression-hotspots.py b/scripts/regression-hotspots.py
index 8a4fb76d..27a684fe 100755
--- a/scripts/regression-hotspots.py
+++ b/scripts/regression-hotspots.py
@@ -8,7 +8,7 @@
 #
 # Uses https://github.com/gitpython-developers/GitPython
 # Results published in https://wiki.documentfoundation.org/Development/RegressionHotspots
-# Run in LibreOffice core directory. Shouldn't take more than a minute.
+# Run in LibreOffice core directory directing output to a text file. Shouldn't take more than a minute.
 
 import sys
 import re
@@ -49,16 +49,20 @@ def get_dir_counts(file_counts, level):
 
 def print_counts(counts):
     printorder = reversed(sorted((count, name) for (name, count) in counts.items()))
+    # wiki page uses a widget to clamp the output while offering a button to expand
+    print('<pre class="clamped">')
     for count in printorder:
         print('%5d %s' % (count[0], count[1]))
+    print('</pre>')
 
 if __name__ == '__main__':
     file_counts = {}
-    excluderegex = re.compile(r'qa/|icon-themes/|extras/source/gallery/|extras/source/palettes/|extras/source/templates/|extras/source/truetype/|helpcontent2|dictionaries|translations|download\.lst|\.png|\.patch')
+    excluderegex = re.compile(r'qa/|qadevOOo/|icon-themes/|extras/source/gallery/|extras/source/palettes/|extras/source/templates/|extras/source/truetype/|\.git-hooks|helpcontent2|dictionaries|translations|download\.lst|\.png|\.patch')
     fixed_regression_ids = get_fixed_regression_bugs()
     sys.stderr.write('found %d fixed regressions: %s\n' % (len(fixed_regression_ids), fixed_regression_ids))
 
     # build a dictionary of hashes and bug IDs from all commits targeting a report in FDO/TDF Bugzilla
+    # (first commit with fdo# aka freedesktop.org is from 1 Oct 2010)
     gitbugs = {}
     buglog = git.Git('.').execute(['git', 'log', '--grep=(fdo|tdf)#', '-E', '--oneline', '--since=1.10.2010'])
     if buglog:
@@ -70,10 +74,10 @@ if __name__ == '__main__':
             if bugid:
                 gitbugs[githash] = int(bugid.group(1))
 
-    # filter by the bug IDs we got from the Bugzilla query
-    regression_hashes = [key for key, value in gitbugs.items() if value in fixed_regression_ids]
+    # create a list of bug fix hashes by filtering with the bug IDs we got from the Bugzilla query
+    fix_hashes = [key for key, value in gitbugs.items() if value in fixed_regression_ids]
 
-    for githash in regression_hashes:
+    for githash in fix_hashes:
         lognames = git.Git('.').execute(['git', 'show', githash, '--pretty=tformat:', '--name-only'])
         if lognames:
             for filename in lognames.split('\n'):