Ted Husted wrote:
On 9/14/05, Laurie Harper <[EMAIL PROTECTED]> wrote: James Mitchell setup some automatic tests for the taglib, though a lot of the behavior is still tested through the tag-exercises application. Niall is correct. When I've applied taglib patches, the bullwork is always setting up tests that first prove that there's a problem, and then prove that the patch fixes it.
Ah, yes, I see the test-cactus stuff in svn. I'll see if I can get that going before I start making any changes.
With tags, a tricky thing is deprecating side effects, like, as Nial mentions whether content is trimmed to an empty string. But, if we had a process for doing that, such behavior could be deprecated and removed like any other. We just need to give people fair notice that change is afoot.
In the absense of such a process, how about adding prominent notes in the documentation? Perhaps along with a temporary option somewhere to select which behaviour to use, the default for which can flip when the deprecated behaviour is dropped?
L. --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
