On 11/26/09 3:58 PM, Erik Huelsmann wrote:
On Thu, Nov 26, 2009 at 12:14 AM, Erik Huelsmannehuels@gmail.com wrote:
Hi Thomas,
Just a minute ago, I committed a fix by replacing the current algorithm with a new one along the lines I described to you before. Your example works, hopefully any other uses too.
By the way, if someone could write a number of test cases to verify that the algorithm actually does what it should, that would be great! I tested it manually, but adding it to the automated tests is a guarantee breakage doesn't go unnoticed.
I definitely owe some explanation on how to add both Lisp and Java based tests to ABCL, so until I discharge my documentation debt, if anyone posts tests, I promise that I will make sure they make it into the tree in a reasonably timely fashion.
In general, as we fix things it would be nice to add tests so that things "stay fixed" in the future. But I am as guilty as anyone else at not sticking to this philosophy.
Mark