Front page | perl.perl5.porters |
Postings from May 2004
On Mon, May 24, 2004 at 05:19:41PM +0100, Nicholas Clark wrote:
> 3: we need another way of specifying that we'd like uninit warnings to be
> reported in the new format.
This also needs to be somethiong that makes it clear that it's a temporary
expedient for 5.8.x only and that it's on in 5.10.x automatically;
ie if we introduce the general public to (say)
use warnings 'undef_var_names';
then in 5.10.0 they'll have to do this instead to maintain the status quo:
no warnings 'undef_var_names';
which leads one to wonder whether the name needs to be more like:
use warnings 'undef_var_names_in_58x_only';
which leads one to wonder whether there should be a more general
mechanism for early enabling of new features - then we could include
the // operator say; so maybe something like
use feature qw(undef_var_names defined_or);
where it's clearly documented that at some point in a future release
a particular feature will become enabled by default.
Of course this runs into the lexical scoping of hints problem, which
I've never been sure whether its been fixed yet.
One final Cassanda-ism - the uninit warnings code is still bound
to tuen up some false positives - eg printing the wrong variable name.
These will gradually be ironed out over time, but it needs time...
Dave.
--
The warp engines start playing up a bit, but seem to sort themselves out
after a while without any intervention from boy genius Wesley Crusher.
-- Things That Never Happen in "Star Trek" #17
Thread Previous
|
Thread Next