issue 27: parse x3 as x operator plus a number

[moregan]

Fix for Issue #27, plus tests.

This change includes the elimination of existing code that would separate the 'x' operator from a number immediately following, but only when preceded by a small handful of token types, e.g: "a"x3 This change eliminates paying attention to the preceding token.

[moregan]

I put tests in t/ppi_token_operator.t, rather than in t/data/, since I find the former easier to work with. Are new tests still being written for t/data/ ?

[adamkennedy]

The t/data tests should be parsing tests that do not or can not specify
where in the implementation they are being implemented.

I would recommend them where possible (or at least put a copy of the
failing parse case in there for future regression testing).

The use of ppi_token_operator.t implies your test is related specifically
to that single class, which these tests are not necesarily going to.

However, I'd rather someone fix it, so doing it your way should be fine.

Just please don't convert the existing tests over, they represent a
hard-earned store of regression tests.

On 3 February 2014 19:07, moregan notifications@github.com wrote:

I put tests in t/ppi_token_operator.t, rather than in t/data/, since I
find the former easier to work with. Are new tests still being written for
t/data/ ?

Reply to this email directly or view it on GitHubhttps://github.com//pull/33#issuecomment-34026723
.

[moregan]

A/B testing with Perl::Critic against a sampling of CPAN discovered that x=>$x was being parsed as x= > $x on master. Fix with tests added to x-operator-not-recognized branch with 8cb2974

[wchristian]

Looks good and passes tests, merged. :)