Re: [code-review] Re: Lingua::Phonology
"Martyn J. Pearce" <[email protected]> Thu, 13 Nov 2003 18:41:13 +0000
| Newsgroups | gmane.comp.lang.perl.code-review-ladder |
|---|---|
| Message-ID | <[email protected]> |
On Thu, Nov 13, 2003 at 11:42:44AM -0600, Hodges, Paul wrote: > > For a more annoying example, see > > http://blog.simon-cozens.org/bryar.cgi/id_6416 > > which caught me out in one of my modules. > > O Lord! And here I was thinking I knew what I was doing. How have I > consistently missed this topic on the boards for so long? Geez, this is too > scary! > > So from the above link, what is it that is different? How the hell can order > of definition change the behavior? This is nutz! Because in the first example, the 'new X::Y' is compiled before Perl even saw the existence of the X::Y package, and so had to guess at which side of the ambiguity to reckon on. In the second example, package X::Y had been seen, so Perl determines that 'new X::Y' is probably an indirect object method call. I provide this explanation by way of reasoning, not excuse :-) Mx.