db-derby-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jack Klebanoff <klebanoff-de...@sbcglobal.net>
Subject Re: All of derby_all fails when environment corresponding derbyLocale_**.jar exists in CLASSPATH
Date Tue, 31 May 2005 14:45:51 GMT
Øystein Grøvlen wrote:

>>>>>>"DJD" == Daniel John Debrunner <djd@debrunners.com> writes:
>>>>>>            
>>>>>>
>
>    DJD> I think the test suite has to have a known environment to allow it to
>    DJD> run successfully on widely different machines. Examples of this known
>    DJD> environment are the expected locale, character set encoding for the
>    DJD> scripts etc. Ideally the test harness should set this environment itself.
>
>Asa long as you have a test harness that is based on evaluating the
>output to file from each test program, I think you are right.
>However, I think it should be possible to build a locale independent
>test harness in JUnit by instead evaluating the return values from API
>calls (e.g., message ids instead of message strings).
>
>  
>
Error messages are important. We should test them.

Remember that the messages are generated by substituting argument values 
into strings fetched from the messages files. There may be bugs in the 
message generation system, more often Derby programmers make mistakes in 
messages and arguments. So if a test just looks at the message IDs and 
not at the message strings we will miss some bugs.

Jack Klebanoff

Mime
View raw message