Details
-
Task
-
Resolution: Unresolved
-
P3: Somewhat important
-
None
-
5.12
-
None
Description
The generate_expected_output.py script replaces parts of a test's output that vary from run to run (and don't implicate a test issue) with "standard" values that save us noise in git when updating output. For example, test durations and line numbers are set to 0, file names are pruned of path prefixes outside the source tree. The tst_selftests.cpp code to compare actual output with expected knows to do matching blanking. It also knows to ignore the details of gdb output in tests that (by design) crash, trigger signals or time out. However, generate_expected_output.py doesn't canonicalise thise gdb output to match. It should do so, reducing the whole mess to just the minimal content that tst_selftests.cpp's code will recognise as gdb output and ignore (as long as the actual output also contains similar).