[imp] PATCH: IMP_ACL_rfc2086 - things I learnt yesterday!

[Chris Hastie]

On Wed, 5 Feb 2003, Jan Schneider <jan at horde.org> wrote
>Zitat von Chris Hastie <lists at oak-wood.co.uk>:
>
>> Attached is a patch to imp/lib/ACL/rfc2086.php that deals with various
>> issues that came up yesterday, along with a small patch to imp/acl.php.
>
>The patch to acl.php isn't necessary, as a variable that's equal to
>'no_support' is _always_ a string.
>
I have to confess to not entirely understanding why, but empirically the 
patch _is_ necessary, here any way. If I try to change permissions on a 
folder that I had no 'a' right on, I was getting thrown back to the 
Options page. Investigating this it seemed that

0 == 'no_support'
but
!(0 === 'no_support')

The patch sorted out that problem, so I submitted it.

>> *   Warnings include more detailed information about what the problem
>>     was. Hopefully this will make further debugging easier.
>
>Great, but please replace double with single quotes where possible (see
>Eric's addition to CODING_STANDARDS).

OK, I'll clean that up this evening.
>
>I already committed his fix. If you didn't yet subscribe to the cvs mailing
>list you should as soon as possible.

I do. The message arrived whilst I was composing that email and I saw it 
within about 1 sec of pressing 'Post'!

>> 2   Why is it preferable to get the values of variables in a class by
>>     calling functions that return them rather than simply looking at the
>>     variable? Is it a question of preferred style or is there some
>>     intrinsic problem with just looking at variables?
>
>Because this is OOP. At least what PHP allows of it.
>
OOP?
-- 
Chris Hastie