Front page | perl.perl5.porters |
Postings from June 2013
On 06/27/2013 10:52 AM, Leon Timmermans via RT wrote:
> On Thu, Jun 27, 2013 at 3:41 PM, Reini Urban <reini@cpanel.net> wrote:
> 4>
>>> I've attached a version of the patch updated to work with blead.
>>
>>
>> Great, Thanks for the improvements.
>> I've looked over it, LGTM
>>
>> Can someone please apply it now.
>
> I think having macros called CHECK_<whatever> that return from the
> function is unnecessarily obfuscating the code.
Have you got a better idea?
if (CHECK_SAFESYSCALL(pv)) return NULL;
if (CHECK_SAFEPATHNAME(pv)) return undef;
were the alternatives.
> I also think we shouldn't turn this warning on by default without some
> discussion about that.
We return with undef from syscalling functions with a wrong name,
set $^E to ENOENT
I think turning on this warning by default makes sense:
It only warns on wrong user data, not on code errors.
It's surprising, I want to see that attack on the console and especially
in log files, but have no strong opinion on it.
With require I die immediately.
~ grep DEFAULT_ON regen/warnings.pl
'inplace' => [ 5.008, DEFAULT_ON],
'debugging' => [ 5.008, DEFAULT_ON],
'malloc' => [ 5.008, DEFAULT_ON],
'syscalls' => [ 5.017, DEFAULT_ON],
'deprecated' => [ 5.008, DEFAULT_ON],
'glob' => [ 5.008, DEFAULT_ON],
'experimental::lexical_subs' => [ 5.017, DEFAULT_ON ],
'experimental::regex_sets' => [ 5.017, DEFAULT_ON ],
'experimental::lexical_topic' => [ 5.017, DEFAULT_ON ],
'experimental::smartmatch' => [ 5.017, DEFAULT_ON ],
--
Reini
Thread Previous
|
Thread Next