Path: utzoo!censor!geac!torsqnt!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!elroy.jpl.nasa.gov!jpl-devvax!lwall
From: lwall@jpl-devvax.JPL.NASA.GOV (Larry Wall)
Newsgroups: comp.lang.perl
Subject: Re: Perl Lint?
Keywords: detab
Message-ID: <10751@jpl-devvax.JPL.NASA.GOV>
Date: 14 Dec 90 20:01:19 GMT
References: <1990Dec3.035541.24184@agora.uucp> <452@stephsf.stephsf.com> <111028@convex.convex.com> <454@stephsf.stephsf.com>
Reply-To: lwall@jpl-devvax.JPL.NASA.GOV (Larry Wall)
Organization: Jet Propulsion Laboratory, Pasadena, CA
Lines: 30

In article <454@stephsf.stephsf.com> wengland@stephsf.stephsf.com (Bill England) writes:
: In article <111028@convex.convex.com> tchrist@convex.COM (Tom Christiansen) writes:
: 
: >:	return ();
: >:}
: 
: >Hmm... shouldn't that last one be:
: >
: >	return '';  # or return undef
: >
: 
:   Good point  ...  "perl lint" with return type checking anyone ?

Remember that Perl subroutines can be context sensitive.  A Perl lint
would probably have to punt on any routine containing the wantarray
function, or do some rather hairy analysis.  And this still wouldn't solve
the "problem" that a routine can return () in a scalar context even when
the return type is consistent.  And it's not always possible to look
at a subroutine call and tell easily if it's in a scalar or array context,
since it can vary dynamically (if the call is embedded in another subroutine).
You'd have to construct the entire call graph.  Probably not worth it, given
that () in a scalar context Does the Right Thing anyway.

It might possibly be worth putting in a warning about, if -w is set and
it can determine at runtime that there's an inconsistency.  (I can tell
I've been writing the Book too long\*-I wanted to write
.P \-w
instead.)

Larry