1. We should test if EUPHORIA is handling bad code well
- Posted by SDPringle Sep 02, 2008
- 852 views
A good idea was to put a test suite with the new EUPHORIA. We see many tests that anyone can run to see if the new interpreter interprets good EUPHORIA code well. And by good I mean good syntax and am not talking of coding style. However, it is also important that EUPHORIA handle syntax errors correctly. For example, if I declare a namespace in one include file and the same namespace in another include file and include both of these in my main program will the interpreter tell me the what is wrong with my code? If it reports the wrong syntax errror the t_c*.* tests would return true, good it closed with a syntax error. If euphoria just crashes without generating an ex.err at all the test would return true. This is a big problem. Let me and some of you developers design a test framework for how testing that bad code is handled correctly. What do you say?
I should have posted this in the developers list on sourceforge but I left my sourceforge password at home. EUPHORIA is a program written in EUPHORIA so I can ask for help here.
Shawn
2. Re: We should test if EUPHORIA is handling bad code well
- Posted by ne1uno Sep 03, 2008
- 877 views
- Last edited Sep 04, 2008
If euphoria just crashes without generating an ex.err at all the test would return true. This is a big problem. Let me and some of you developers design a test framework for how testing that bad code is handled correctly.
what about write a 0 to a log file on start, just before exit write a 1. the test calling program will know the program under test at least got to write either a zero or a 1.
is there another obvious way for a program to detect a crash? a new instance of ex.err might also show a machine exception that could be counted as a passed test if it wasn't anticipated.
