Imagine you have a system where a program is running and somehow an abnormality occurs. (it can be a crash, or an abnormal screen or any other thing)
Imagine reproducing the problem is next to impossible but you have some logs files that record what has happened till that point.
Usually you debug by reproducing the problem and checking where does it go wrong but if you just try to find the causes of the abnormality by going through the log files, is there a specific name for this?
-
My coworkers have called it log-based debugging, but I've no idea if that's official.Ed Grimm– Ed Grimm2019年01月29日 04:32:06 +00:00Commented Jan 29, 2019 at 4:32
-
It may be nit-picking, but I would just call that troubleshooting since there's no debugger attached.Dan Wilson– Dan Wilson2019年01月29日 04:48:18 +00:00Commented Jan 29, 2019 at 4:48
-
It’s just a variant of printf debugging.Telastyn– Telastyn2019年01月29日 06:22:38 +00:00Commented Jan 29, 2019 at 6:22
2 Answers 2
Yes, there is a well-known term for this, it is called
Post-mortem debugging
(See Wikipedia, Debugging, Techniques).
This is typically done using something like an automatically generated crash-dump file, and a debugger which can read that file, but I don't think the term is (or should be) restricted to that solution. IMHO it fits to all debugging techniques which can be applied to analyse a program after "death".
If you don't mind a somewhat humorous answer, this could be considered a case of "examining the entrails " according to the Jargon File.