Jim Fulton wrote:

Your enhancements to check run time look very interesting,
but I think they ought to be proposed and discussed.

Please revert this change and then lets discuss what we want.


So, the main idea of the work we did last week on this topic
was to provide an easy tool to measure the performance of "low-level" code.

By "low-level" i mean: this is not intented to be some benchmarking tool or whatsoever, it's just a anti-regression tool, that can be used to mark *critical* parts of
a zope application in the tests.

When someone changes a zope app code, he is pretty confident on the result
because there's a full load of unit/functionnal tests backing him. But he *almost* never asks himself if the code that he added didn't lower down the performance of the app.

Maybe because it's not the front matter in web apps needs that in can be industrial apps sometime,
but when scalability is needed it can become quite important.

anyway, the idea we had was to provide a decorator than can be used in tests to mark a particular test,
saying :

"hey, that's a hot spot, let's mark it to prevent speed regression in the future,
it takes 5 kPystone right now, it shouldn't go over 6kPs"


Zope3-dev mailing list
Unsub: http://mail.zope.org/mailman/options/zope3-dev/archive%40mail-archive.com

Reply via email to