Optimum grammar for expressions in strings

[Lachlan McDonald]

Hi everyone,

    I'm trying to find the formal\acceptable method for finding patterns within a group of otherwise unimportant information. For example to find ${i} within a string, where n is some integer.

The grammar looks quite verbose:

Pattern

    = (Plain Integer Plain)*

    

Plain

    = $(!'${'.)* { return "<String>" }

   

Integer

    = '${' v:$([0-9]+) '}' { return parseInt(v, 10); }

 

Which for the input:

abc ${1}${2} xyz ${3}

Will output:

[

   [

      "<String>",

      1,

      "<String>"

   ],

   [

      "<String>",

      2,

      "<String>"

   ],

   [

      "<String>",

      3,

      "<String>"

   ]

]

Where as this would be far more preferable:

[

   "<String>",

   1,

   "<String>",

   2,

   "<String>",

   3,

   "<String>"

]

Any and all suggestions are welcome.

[jas...@snmpstack.com]

this is about a close as your going to get I think

Pattern

    = (Plain / Integer)*

    

Plain

    = $(!'${'.)+ { return "<String>" }

   

Integer

    = '${' v:$([0-9]+) '}' { return parseInt(v, 10); }

given the provided input 

gives 

[
   "<String>",
   1,
   2,
   "<String>",
   3
]

If you change Plain to : $(!'${'.)* { return "<String>" } you will get a potential infinite loop detected due to the fact it could match nothing forever.....