[
https://issues.apache.org/jira/browse/MATH-1026?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13768774#comment-13768774
]
Gilles commented on MATH-1026:
------------------------------
bq. But any patch on the test cases would have been large because of all the
code that was copied and pasted between the methods.
The patch _can_ be big if it's clear that it contains the same repetitive
change. but in this case, it changes too many things at once.
bq. think of it as fewer lines of code to review.
I'm not sure I understand what you mean.
The simplest for review is to change things iteratively. In this case, I wanted
to be sure that all tests are run and produce the same results. This is easy if
the number of tests stays the same, and their name stays the name.
Just look at it from this viewpoint: we had 23 unit tests for the old
"LevenbergMarquardtOptimizer", and now we have only 2! We thus have to read
code to look for an explanation? Where is the saving?
A test is also often a small use-case, and it is probably clearer that they are
relatively independent. This is not to say that they cannot share code (see
e.g. "MinPackTest").
bq. you would prefer JUnit to directly execute each test case
Indeed.
As I said, you can later raise an improvement request for the tests layout.
Alternately, you can try and raise your point now on the ML, and have someone
else review those changes.
> Separate Optimization Problem from Algorithm
> --------------------------------------------
>
> Key: MATH-1026
> URL: https://issues.apache.org/jira/browse/MATH-1026
> Project: Commons Math
> Issue Type: Improvement
> Reporter: Evan Ward
> Attachments: opt.tar.gz
>
>
> See discussion on the mailing list starting with:
> http://www.mail-archive.com/[email protected]/msg39681.html
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
