Post

1 follower Follow
0

Understanding the impact of warnings in analysis logs

While errors and fatal errors will stop the analysis, since the analyzer is exiting with an error code, warning messages do not and so need to be evaluated in the analysis log.

The golden rule is that not all warnings are the same! The current logic of CAST AIP is to catch any exceptions thrown and post them as warnings in analysis log. This means that some warnings may in fact indicate a crash of an analyzer or an extension that is running on top pf an analyzer

Some hints of how to understand the impact are

  • Check for information  in official documentation, troubleshooting guides or Zendesk tickets, so as to exploit information that already exists
  • If the warning involves a specific file and after the warning the analysis continues with usual information messages, then the warning will be limited to the specific file
  • If the warning has the form of an exception stack trace and involves the main analyzer running, then this may indicate that the analysis has stopped. For example if this happens during parsing files and no file is parsed after that, then this is a sign that the analysis has failed. In this case validate results using Enlighten
  • If the warning has the form of an exception stack trace and involves the an extension running on top of the main analyzer, then this impacts just the extension. The impact for th extension can be evaluated according to where the exception has been produced. During parsing phase it typically involves just the parsed files. During object creation phase or link creation phase this may mean lost objects/links or no objects/links for the particular extension.In this case validate results using Enlighten and referring to official documentation regarding the extension

Please sign in to leave a comment.