[Tutor] Unit testing

[Tino Dai]

On 6/27/06, Tino Dai <tinoloc at gmail.com> wrote:
>
> On 6/27/06, Kent Johnson <kent37 at tds.net> wrote:
>
> > Tino Dai wrote:
> > > And there is one caveat, I
> > > will have to make a bunch of semaphores global instead of local to the
> > > classes. While I know that there is no hard and fast rule about using
> > > global variables, where can I find or can somebody tell me where I can
> >
> > > find some guidelines about them (or is this a use common sense rule)?
> >
> > Main guideline - don't use globals. Fallback rule - don't use globals.
> > Third rule - OK, if you really can't think of any other way, make it a
> > global. :-)
> >
> > Can you pass the semaphore to the class constructor (__init__() method)?
> > What is it about the unit test that pushes you to make it global?
>
>
> Man, spoil all my fun :) What I'm doing is I have a
> set.......actually...hang on...are you talking about the program or the unit
> test? The unit test doesn't have an globals. The program itself does. What
> the program is doing launching a bunch of threads that are linked together
> via queues. The reading of the queues by the next stage in the program is
> controlled by a semaphore. The semaphore will release on one side and
> acquire on the other side. The data is passed along the different threads
> until the data is indexed. The semaphores are global so that the unit test
> can bring in only one class at a time. How I had it before was: the
> semaphore would be local to the class and subsequent class would all that
> local semaphore. I think it might be easier if I just shown you.
>
> How I had it before:
>
> class nameA:
>     sema = threading.semaphore()
>     def __init__(self):
>         <do some stuff>
>
>     def run(self):
>        <do some stuff>
>        nameA.sema.release ()
>
> class nameB:
>      def __init__(self):
>         <do some stuff>
>
>      def run(self):
>          nameA.sema.acquire()
>          <do some stuff>
>
>
> How I have it now:
>
> semaA = threading.semaphore()
>
> class nameA:
>    def __init__(self):
>         <do some stuff>
>
>    def run(self):
>          <do some stuff>
>          semaA.release()
>
> class nameB:
>    def __init__(self):
>         <do some stuff>
>
>    def run(self):
>          semaA.acquire()
>          <do some stuff>
>
>
> Does that make sense. Or is there a better way?
>
> -Tino
>
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/tutor/attachments/20060627/3b29a8b8/attachment.html