WIP

client/stats/correlate-coverage.py

@@ -53,45 +53,82 @@ def getFileCoverage(baseURL, filename):
   try:
     urlObj = mech.open(covURL)
   except:
-    return ("N/A", "N/A")
+    return ("N/A", "N/A", "N/A", "N/A")
 
   parsedUrlObj = lxml.html.parse(urlObj)
 
-  # Xpath to the coverage, see below
-  lineCoveragePercent = float(parsedUrlObj.xpath("/html/body/table[1]/tr[3]/td/table/tr[2]/td[7]")[0].text.replace(" %", ""));
-  # ------------------------------------------------------------------------------^
-  #   2 - line coverage
-  #   3 - function coverage
-  #   4 - branch coverage
-  # ------------------------------------------------------------------------------------^
-  #   5 - # hit
-  #   6 - # total
-  #   7 - % hit
-
 
-  lineCoverageHit = int(parsedUrlObj.xpath("/html/body/table[1]/tr[3]/td/table/tr[2]/td[5]")[0].text)
-  lineCoverageTotal = int(parsedUrlObj.xpath("/html/body/table[1]/tr[3]/td/table/tr[2]/td[6]")[0].text)
+  branchCoveragePercent = "N/A"
+  branchCoverageMissing = "N/A"
+  lineCoveragePercent = "N/A"
+  lineCoverageMissing = "N/A"
 
-  lineCoverageMissing = lineCoverageTotal - lineCoverageHit
-
+  try:
+    # Xpath to the coverage, see below
+    lineCoveragePercent = float(parsedUrlObj.xpath("/html/body/table[1]/tr[3]/td/table/tr[2]/td[7]")[0].text.replace(" %", ""));
+    # ------------------------------------------------------------------------------------^
+    #   2 - line coverage
+    #   3 - function coverage
+    #   4 - branch coverage
+    # ------------------------------------------------------------------------------------------^
+    #   5 - # hit
+    #   6 - # total
+    #   7 - % hit
+
+    lineCoverageHit = int(parsedUrlObj.xpath("/html/body/table[1]/tr[3]/td/table/tr[2]/td[5]")[0].text)
+    lineCoverageTotal = int(parsedUrlObj.xpath("/html/body/table[1]/tr[3]/td/table/tr[2]/td[6]")[0].text)
+    lineCoverageMissing = lineCoverageTotal - lineCoverageHit
+  except ValueError:
+    pass
+
+  try:
+    branchCoveragePercent = float(parsedUrlObj.xpath("/html/body/table[1]/tr[3]/td/table/tr[4]/td[7]")[0].text.replace(" %", ""));
+    branchCoverageHit = int(parsedUrlObj.xpath("/html/body/table[1]/tr[3]/td/table/tr[4]/td[5]")[0].text)
+    branchCoverageTotal = int(parsedUrlObj.xpath("/html/body/table[1]/tr[3]/td/table/tr[4]/td[6]")[0].text)
+    branchCoverageMissing = branchCoverageTotal - branchCoverageHit
+  except ValueError:
+    pass
+
   mech.close()
 
-  return (lineCoveragePercent, lineCoverageMissing)
+  return (lineCoveragePercent, lineCoverageMissing, branchCoveragePercent, branchCoverageMissing)
 
 def main():
   currentCovURL = getCurrentCoverageDirectory("http://people.mozilla.org/~choller/firefox/coverage/")
 
   for record in fileinput.input():
     record = record.lstrip().rstrip()
+
+    # Preserve comments, possibly append new data column description though
+    dataDescriptorFollows = False
+    if record.startswith('#'):
+      if "DATA DESCRIPTOR FOLLOWS" in record:
+        dataDescriptorFollows = True
+        print record
+      elif dataDescriptorFollows:
+        print record + " | % Line Coverage | Number of lines missing | % Branch Coverage | Number of branches missing"
+        dataDescriptorFollows = False
+      else:
+        print record
+
+      continue
+
     recordElements = record.split()
 
     if (len(recordElements) != 2):
       print "Warning: Ignoring malformed record: " + record
       continue
 
     cov = getFileCoverage(currentCovURL, recordElements[0])
+
+    covMetric = "N/A"
+
+    try:
+      covMetric = (float(recordElements[1]) / (float(cov[0]) / float(100))) * float(cov[1])
+    except:
+      pass
 
-    print record + " " + str(cov[0]) + " " + str(cov[1])
+    print record + " " + str(cov[0]) + " " + str(cov[1]) + " " + str(cov[2]) + " " + str(cov[3]) #+ " " + str(covMetric)
 
 if __name__ == '__main__':
     main()

client/stats/crithighfixes.py

@@ -36,12 +36,12 @@
     'resolution':       'FIXED',
     # Ignore old bugs, should be fixed at most 180 days ago
     'chfieldto':        'Now',
-    'chfieldfrom':      '-60d',
+    'chfieldfrom':      '-180d',
     'chfield':          'resolution',
     'chfieldvalue':     'FIXED',
-     # Advanced search criteria
+    # Advanced search criteria
     'query_format':     'advanced',
-     # Should have an [sg:crit/high tag in whiteboard
+    # Should have an [sg:crit/high tag in whiteboard
     'type0-0-0':        'regexp',
     'field0-0-0':       'status_whiteboard',
     'value0-0-0':       '\[sg:(critical|high)',
@@ -52,6 +52,16 @@
     'include_fields':   '_default',
 }
 
+debug = False
+
+def printOptionsComment(opts):
+  commentStr = str(opts)
+  commentStrParts = commentStr.strip('{}').split(', ')
+
+  print "## Bugzilla Search Criteria ##"
+  for commentStrPart in commentStrParts:
+    print "# " + commentStrPart
+
 def fetchBug(bug_id):
   """Fetch the bug for the given bug_id and apply necessary workarounds."""
   bug = bz.get_bug(bug_id)
@@ -78,6 +88,25 @@ def extractRevisionsFromURL(text):
 
   return None
 
+def hgCheckLog(repoDir, rev, bugnum):
+  """Given a HG repository, a revision and a bug number, return true if the bug number is in the commit message."""
+  origCwd = os.getcwd()
+  os.chdir(repoDir)
+
+  if debug:
+    print "Running hg log..."
+
+  logDesc = subprocess.check_output(['hg', 'log', '-r', rev, '--template', '{desc}'])
+
+  os.chdir(origCwd)
+
+  bugInLog = (str(bugnum) in logDesc)
+
+  if not bugInLog:
+    print >> sys.stderr, "Warning: Ignoring possible false positive (bug %s, revision %s)" % (str(bugnum), rev)
+
+  return bugInLog
+
 def hgDiffStat(repoDir, rev):
   """Given a HG repository and revision, return the files touched by that revision."""
   origCwd = os.getcwd()
@@ -114,7 +143,7 @@ def scanBug(bugnum, repoDir):
   """
   # Fetch the bug
   bug = fetchBug(bugnum)
-  
+
   for comment in bug.comments:
     revs = extractRevisionsFromURL(comment.text)
 
@@ -127,6 +156,9 @@ def scanBug(bugnum, repoDir):
     totalFiles = []
 
     for rev in revs:
+      if not hgCheckLog(repoDir, rev, bugnum):
+        continue
+
       files = hgDiffStat(repoDir, rev)
       totalFiles.extend(files)
 
@@ -136,17 +168,23 @@ def scanBug(bugnum, repoDir):
     # a file twice, then this should not be counted as two bugs.
     return list(set(totalFiles))
 
-  print "Warning: No fix found in bug " + str(bugnum)
+  print >> sys.stderr, "Warning: No fix found in bug %s" % str(bugnum)
 
 def main():
+  printOptionsComment(options)
+
+  print "# DATA DESCRIPTOR FOLLOWS"
+  print "# Filename | Number of Changes from Security Bugs"
+
   # Get the bugs from the api
   buglist = bz.get_bug_list(options)
 
   if len(buglist) == 0:
     print "No bugs found."
     sys.exit(0)
 
-  print "Found %s bugs:" % (len(buglist))
+  if debug:
+    print "Found %s bugs:" % (len(buglist))
 
   dirs = {}
   cnt = 0;