Parsing binary data

[romanlehnert]

Hey there,

i'm trying to figure out how to parse string that may contain binary data, which is prefixed with the byte-length inside the string and may contain unescaped control characters.

Regularly, the elements in the string are separated by a :. And when the : appears inside a string, it is escaped by a ?.

Example:

elem?:ent_1:element_2:element_3

Should parse to

[ "elem?:ent_1", "element_2", "element_3"]

But such an element may also contain binary data. This binary data

• Contains unescaped controlcharacters (:)
• Is prefixed with its length: @length@data

Within a row, it may look like this:

elem?:ent_1:@18@my_binary:string!!:element_3

Should parse to

[ "elem?:ent_1", "my_binary:string!!, "element_3"]

What is the best way to handle this with parslet? I'm really thankful for any advise.

[ghost]

1. Have you read Capturing input section in this article? Something like this should work:

   require 'parslet'
   include Parslet
   ele =
    str('@') >>
    match('\\d').repeat.capture(:size) >>
    str('@') >>
    dynamic { |s, c| any.repeat(c.captures[:size].to_i, c.captures[:size].to_i) }
   ele.parse('@4@abcd')

2. If the string contains any binary data, try to use ASCII-8BIT encode, it may work.