Em 2012-09-03 23:54, Stas Malyshev escreveu:
Hi!
indicating such mental model is widespread and (2) given that the
exceptions you claim to be surprising don't happen but for a
programming
error (i.e. they're not exceptions that need to be caught).
Following this logic, we'd have to convert all E_NOTICE and E_STRICT
to
fatal errors or exceptions - they are usually produced by programming
errors and aren't meant to be caught by surrounding code (actually,
can't). But I don't see anybody benefiting from this - as I don't see
anybody benefiting from generator that will explode your application
if
you touch it twice.
That does not follow at all.
First, E_NOTICE exists to permit a compromise between people like Zeev
that think its introduction was an error and that one should be able to
refer to undefined variables and people who prefer a stricter model. For
the programmers that follow this stricter model, an E_NOTICE ends ups
frequently being a fatal (because, for instance, you invoke a method on
the resulting null); some other times you're lucky and you just missed
an "else $var = null" branch, other times the error is far more serious,
like an unintended null -> 0 conversion.
Second, E_STRICT exists to encourage good OOP practices. There is a lot
of code that generates E_STRICT that does what it was intended to do
(ask Lester). By definition, code that's not supposed to generate
E_STRICT messages but does is buggy, but the fact that some code raises
an E_STRICT does not signal, by itself, a bug.
--
Gustavo Lopes
--
PHP Internals - PHP Runtime Development Mailing List
To unsubscribe, visit: http://www.php.net/unsub.php
