[GROOVY-5159] Grammar could accept nls! before SPREAD_DOT

• Type: Improvement
• Status: Closed
• Priority: Major
• Resolution: Duplicate
• Affects Version/s: 1.8.4
• Fix Version/s: 2.3.0-beta-1
• Component/s: parser, parser-antlr2
• Assignee: Paul King
• Reporter: Tim Yates
• Created: 02/Dec/11 04:27
• Updated: 05/Apr/14 00:00
• Resolved: 06/Mar/14 20:57

Description

When laying our code to be readable, the ability to do:

'name score\ntim 1\ndave 4\ntim 3'.split( /\n/ )               // split on newline
                                  .drop( 1 )                   // drop header
                                  .collect { it.split( / / ) } // split each by space
                                  .groupBy { it[ 0 ] }         // group by name

Makes code fantastically readable when explaining it to people.

However, if I want to replace the line

                                  .collect { it.split( / / ) } // split each by space

with

                                  *.split( / / )               // split each by space

Then I can't as the parser complains that

unexpected token: *. at line: 3, column: 35

The only workaround is to place all of the operators at the ends of the lines, which to my eyes looks a bit of a mess (and in my experience is confusing to people)

ie:

'name score\ntim 1\ndave 4\ntim 3'.split( /\n/ ).      // split on newline
                                   drop( 1 )*.         // drop header
                                   split( / / ).       // split each by space
                                   groupBy { it[ 0 ] } // group by name

Obviously, I could go for a style of appending the single '.' with appending the '*.', but I feel this makes it more confusing again...

I've had a look at groovy.g, but can't seem to spot the trick to allow this, without causing ambiguity