Pedantic errors (Was: Re: [PHP-DEV] [RFC][VOTE] Improve array to string conversion)
| From: | Patrick ALLAERT | Date: | Fri, 06 Mar 2015 15:18:09 +0000 |
| Subject: | Pedantic errors (Was: Re: [PHP-DEV] [RFC][VOTE] Improve array to string conversion) | ||
| References: | 1 | Groups: | php.internals |
| Request: | Send a blank email to internals+get-84383@lists.php.net to get a copy of this message | ||
Le lun. 23 févr. 2015 à 17:06, François Laupretre <francois@php.net> a
écrit :
> Hi,
>
> Starting the vote for https://wiki.php.net/rfc/array-to-string.
>
> Please note that, while the initial RFC proposed both options of either
> fully supporting the feature, or disabling it, the voting choices are now :
>
> - either definitely disabling array to string conversion,
> - or keep current behavior as-is.
>
> Vote ends March 6, 00:00 (UTC).
>
> Regards
>
> François
>
As said in the original thread [1], I'm not in favor of breaking the
execution in the case of "Array to string" conversion with an
E_RECOVERABLE_ERROR with no "opt-in" choice for that (which implies an
important BC break).
However, maybe there is a better option for those that want to be 'more
"strict" and "explicit"' (quoting Pascal Martin [2]).
Having the option to bail out at any kind of errors (E_NOTICE,
E_WARNING,...) might fit that as well.
gcc has a similar switch:
"-pedantic-errors: Like -Wpedantic, except that errors are produced
rather than warnings."
This could be easily developed thanks to a configuration switch (off by
default) or an "error_reporting"-like bit field instructing on which error
level to bail out additionally.
By doing so, we would let people opt-in for that kind of behaviour.
Thoughts?
[1] http://marc.info/?t=142470775700006
[2] http://marc.info/?l=php-internals&m=142559407125491