freaky deaky plus sign usage

metaperl has asked for the wisdom of the Perl Monks concerning the following question:

Replies are listed 'Best First'.
Re: freaky deaky plus sign usage by AnomalousMonk (Archbishop) on Jul 31, 2009 at 15:58 UTC
This is to get around Perl's 'indirect object syntax'. Otherwise `ok Mock::Basic->update(...);` looks like `Mock::Basic->ok->update(...);` which is probably not what is intended. `ok(Mock::Basic->update(...));` works as well.	[reply] [d/l] [select]
Re^2: freaky deaky plus sign usage by metaperl (Curate) on Jul 31, 2009 at 18:37 UTC
ok, and how does `+` "get around" it? What exactly is syntactically happening? Is "ok" being added to "Mock::Basic->update(...)"	[reply] [d/l]
Re^3: freaky deaky plus sign usage by Anonymous Monk on Jul 31, 2009 at 19:23 UTC
No, the return value from Mock::Basic->update() is numified (if it had a 0+ overload, it gets called). CSJewell, I just don't want to log in atm.	[reply]
Re^4: freaky deaky plus sign usage by AnomalousMonk (Archbishop) on Aug 01, 2009 at 02:31 UTC
Re^5: freaky deaky plus sign usage by JadeNB (Chaplain) on Aug 01, 2009 at 17:11 UTC
Some notes below your chosen depth have not been shown here
Re^4: freaky deaky plus sign usage by Anonymous Monk on Aug 02, 2009 at 15:36 UTC
Re^3: freaky deaky plus sign usage by Anonymous Monk on Aug 01, 2009 at 02:58 UTC
`perl -MCGI -MTest::More -e " ok 1; ok +CGI->new, +CGI->new; ok scalar +CGI->new; ok( CGI->new ); ok CGI->new; " ok 1 ok 2 - CGI=HASH(0x1a08a30) ok 3 ok 4 Undefined subroutine CGI::ok at -e line 1 # Tests were run but no plan was declared and done_testing() was not s +een.` [download]	[reply] [d/l]
Re: freaky deaky plus sign usage by pemungkah (Priest) on Jul 31, 2009 at 20:39 UTC
It tells the parser "Hi. Since I have a leading (unary) plus sign, I'm now a scalar expression, so I qualify to match the first item in `ok()`'s prototype; please do not get confused and interpret me as a class name you should call an `ok()` method on." This was done to avoid putting the parens around `ok()`'s argument, like this: `ok( Mock::Basic->update( 'mock_basic', { name => 'python' }, {id => 1} ), 'I should actually put a description on this test' );` [download] To me this is a lot clearer; it's less mental effort to just add the parens than it is to dork around trying to remember the parsing rules. Once the parens are there, we've unambiguously declared "these are arguments to a subtroutine called `ok()`. Please to not attempt to interpret it otherwise, thank you." I'll igonore the fact that there's no description on the test - I think that's more objectionable than the syntax trickery, frankly. Perhaps you should file a bug on that - feel free to snag this rewrite as a start on a patch.	[reply] [d/l]
Re: freaky deaky plus sign usage by Anonymous Monk on Aug 02, 2009 at 15:43 UTC
Apparently, there's ambiguity in the parser that does not get resolved as desired. `a b->c() desired: a( b->c() ) actual: ( b->a() )->c()` [download] unary-plus is a no-op, but it changes what the compiler sees from a bareword to an operator. That removes the ambiguity. Using parens around the parameters (`ok( ... )`) would also remove the ambiguity. - ikegami	[reply] [d/l] [select]