[Bug 2097264] Re: FTBFS on LP buildd infrastructure

[Simon Chopin]

Jammy isn't affected either.

** Description changed:

- This might be only in PPAs and/or a recent change in our build
- environment, but when trying to rebuild the package for an upcoming SRU
- I got a nice FTBFS.
+ [ Impact ]
+ 
+ The package FTBFS on the LP builders due to the cloud build name
+ containing the 'buildd' string, thus failing the anonymization test.
+ This makes it difficult to do a Noble SRU :)
+ 
+ It's a false positive, that particular field isn't relevant when looking
+ for PII, it's just a side-effect of the extremely artificial build
+ environment.
+ 
+ The fix is to skip that test when such a case is detected.
+ 
+ [ Test plan ]
+ 
+ Check the build logs to see that the test has been skipped.
+ Also build the package in a local schroot and check that the test has NOT been skipped and has passed.
+ 
+ [ Where problems could occur ]
+ 
+ Since we're skipping the tests we could have a regression on that very
+ feature, which is the reason for the local build in the test plan.
+ 
+ [ Original report ]
+ This might be only in PPAs and/or a recent change in our build environment, but when trying to rebuild the package for an upcoming SRU I got a nice FTBFS.
  
  https://launchpadlibrarian.net/773820283/buildlog_ubuntu-noble-
  amd64.apport_2.28.1-0ubuntu3.4~ppa1_BUILDING.txt.gz
  
  The relevant log parts:
  
  self = <tests.integration.test_ui.T testMethod=test_run_crash_anonymity>
  
-     def test_run_crash_anonymity(self):
-         """run_crash() anonymization"""
-         r = self._gen_test_crash()
-         utf8_val = b"\xc3\xa4 " + os.uname()[1].encode("UTF-8") + b" \xe2\x99\xa5 "
-         r["ProcUnicodeValue"] = utf8_val.decode("UTF-8")
-         r["ProcByteArrayValue"] = utf8_val
-         report_file = os.path.join(apport.fileutils.report_dir, "test.crash")
-         with open(report_file, "wb") as f:
-             r.write(f)
-         self.ui = UserInterfaceMock()
-         self.ui.present_details_response = apport.ui.Action(report=True)
-         self.ui.run_crash(report_file)
-         self.assertIsNone(self.ui.msg_severity, self.ui.msg_text)
-     
-         assert self.ui.report
-         self.assertNotIn("ProcCwd", self.ui.report)
-     
-         dump = io.BytesIO()
-         # this contains more or less random characters which might contain the
-         # user name
-         del self.ui.report["CoreDump"]
-         self.ui.report.write(dump)
-         report = dump.getvalue().decode("UTF-8")
-     
-         for s in self._get_sensitive_strings():
+     def test_run_crash_anonymity(self):
+         """run_crash() anonymization"""
+         r = self._gen_test_crash()
+         utf8_val = b"\xc3\xa4 " + os.uname()[1].encode("UTF-8") + b" \xe2\x99\xa5 "
+         r["ProcUnicodeValue"] = utf8_val.decode("UTF-8")
+         r["ProcByteArrayValue"] = utf8_val
+         report_file = os.path.join(apport.fileutils.report_dir, "test.crash")
+         with open(report_file, "wb") as f:
+             r.write(f)
+         self.ui = UserInterfaceMock()
+         self.ui.present_details_response = apport.ui.Action(report=True)
+         self.ui.run_crash(report_file)
+         self.assertIsNone(self.ui.msg_severity, self.ui.msg_text)
+ 
+         assert self.ui.report
+         self.assertNotIn("ProcCwd", self.ui.report)
+ 
+         dump = io.BytesIO()
+         # this contains more or less random characters which might contain the
+         # user name
+         del self.ui.report["CoreDump"]
+         self.ui.report.write(dump)
+         report = dump.getvalue().decode("UTF-8")
+ 
+         for s in self._get_sensitive_strings():
  >           self.assertIsNone(
-                 re.search(rf"\b{re.escape(s)}\b", report),
-                 f"dump contains sensitive word '{s}':\n{report}",
-             )
+                 re.search(rf"\b{re.escape(s)}\b", report),
+                 f"dump contains sensitive word '{s}':\n{report}",
+             )
  E           AssertionError: <re.Match object; span=(97, 103), match='buildd'> is not None : dump contains sensitive word 'buildd':
  E           ProblemType: Crash
  E           Architecture: amd64
  E           CasperMD5CheckResult: unknown
  E           CloudBuildName: ubuntu-base:buildd

-- 
You received this bug notification because you are a member of Ubuntu
Foundations Bugs, which is subscribed to apport in Ubuntu.
https://bugs.launchpad.net/bugs/2097264

Title:
  FTBFS on LP buildd infrastructure

Status in apport package in Ubuntu:
  Invalid
Status in apport source package in Noble:
  In Progress

Bug description:
  [ Impact ]

  The package FTBFS on the LP builders due to the cloud build name
  containing the 'buildd' string, thus failing the anonymization test.
  This makes it difficult to do a Noble SRU :)

  It's a false positive, that particular field isn't relevant when
  looking for PII, it's just a side-effect of the extremely artificial
  build environment.

  The fix is to skip that test when such a case is detected.

  [ Test plan ]

  Check the build logs to see that the test has been skipped.
  Also build the package in a local schroot and check that the test has NOT been skipped and has passed.

  [ Where problems could occur ]

  Since we're skipping the tests we could have a regression on that very
  feature, which is the reason for the local build in the test plan.

  [ Original report ]
  This might be only in PPAs and/or a recent change in our build environment, but when trying to rebuild the package for an upcoming SRU I got a nice FTBFS.

  https://launchpadlibrarian.net/773820283/buildlog_ubuntu-noble-
  amd64.apport_2.28.1-0ubuntu3.4~ppa1_BUILDING.txt.gz

  The relevant log parts:

  self = <tests.integration.test_ui.T
  testMethod=test_run_crash_anonymity>

      def test_run_crash_anonymity(self):
          """run_crash() anonymization"""
          r = self._gen_test_crash()
          utf8_val = b"\xc3\xa4 " + os.uname()[1].encode("UTF-8") + b" \xe2\x99\xa5 "
          r["ProcUnicodeValue"] = utf8_val.decode("UTF-8")
          r["ProcByteArrayValue"] = utf8_val
          report_file = os.path.join(apport.fileutils.report_dir, "test.crash")
          with open(report_file, "wb") as f:
              r.write(f)
          self.ui = UserInterfaceMock()
          self.ui.present_details_response = apport.ui.Action(report=True)
          self.ui.run_crash(report_file)
          self.assertIsNone(self.ui.msg_severity, self.ui.msg_text)

          assert self.ui.report
          self.assertNotIn("ProcCwd", self.ui.report)

          dump = io.BytesIO()
          # this contains more or less random characters which might contain the
          # user name
          del self.ui.report["CoreDump"]
          self.ui.report.write(dump)
          report = dump.getvalue().decode("UTF-8")

          for s in self._get_sensitive_strings():
  >           self.assertIsNone(
                  re.search(rf"\b{re.escape(s)}\b", report),
                  f"dump contains sensitive word '{s}':\n{report}",
              )
  E           AssertionError: <re.Match object; span=(97, 103), match='buildd'> is not None : dump contains sensitive word 'buildd':
  E           ProblemType: Crash
  E           Architecture: amd64
  E           CasperMD5CheckResult: unknown
  E           CloudBuildName: ubuntu-base:buildd

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/apport/+bug/2097264/+subscriptions