rof File Format

STATUS: Likely trashed

I couldn't get the regex-only parsing to work. I needed a config format for bash, but that was kind of... complex... I thought I could find a really simple, straightforward way to handle key:value pairs in multiple programming languages, but... Idunno. I'm very likely giving up on this project.

a.key: A value

Another_key-whoo: A
multiline
Value

A.Key.again: \   <- Keep leading whitespace
Escape the next line, because it looks like a key
\not.really-a_key: Needs escaping
\\ <- To have an actual backslash

The next key has an empty value, but that's okay.
This key wants to keep trailing whitespace      \

last.key:

Under Development

This idea is kind of a flop atm. But I have a crude version that kind of works, detailed just below.

What's working:

The Regex

^([a-zA-Z0-9_\-\.]+):(?:\s|\r|\n)*((?:(?:.|\n|\r)(?!^[a-zA-Z0-9_\-\.]+:))+)
Require multi-line & global flags/modifiers

Rules:

keys contain a-z, A-Z, dash (-), underscore (_), 0-9, and dot (.) and
Keys terminated by :
values contain any characters
values can be multi-line
A value line, if matching keypattern:, will be parsed as a key & a new value will start
White-space is trimmed form beginning of value
White-space is NOT trimmed from end of value
- I desperately want to fix this

Decompressed version of the regex

^([a-zA-Z0-9_\-\.]+) # key
 :
(?:\s|\r|\n)*
 ((?:
    (?:.|\n|\r) # characters we want
    (?!^[a-zA-Z0-9_\-\.]+:) # But NOT if those characters make up a key
 )+)

TODO

Refine the regex so all keys are $1 & all values are $2
Trim whitespace surrounding values
Expand the keys to allow for ([a-zA-Z\-\_0-9\.]+) & possibly other characters.
- I'm testing with just a-z keys because that's wayyyy simpler
Create a multi-lingual examples, at least: bash, PHP, javascript (because I know how to use those lol)
And never use (.|\r|\n)*, use (?s).*

Notes & Wishful thinking

The target format is

key: value 1
nightmare:DELIM:
notakey:
    obviously not a key
notakey:
:DELIM:
abc: value 2
new line
anotherkey:: value 
nostring: on this one
::

Which would yield These key/value pairs

key

value1

nightmare

notakey:
    obviously not a key
notakey:

abc

value 2
new line

anotherkey

value 
nostring: on this one

What is working

Correctly matches non-delimited keys & values
- ([a-z]+):((?:(?:.|\n|\r)(?!^[a-z]+:))+)
- ([a-z]+):((?:(?:.|\n|\r)(?!^[a-z]+:(?![A-Z]*:)))+)
Correctly matches delimited keys & values
- ([a-z]+):([A-Z]*:)((.|\r|\n)*)^:\2
Matches everything correctly, BUT references are not right
- (?:(?:([a-z]+):([A-Z]*:)((.|\r|\n)*)^:\2)|([a-z]+):((?:(?:.|\n|\r)(?!^[a-z]+:))+))
- The 'list' feature on regexr, using ($1|$5) = --$3 || $6 --\n shows everything with clear differentiation between delim & non-delim
  - (?:(?:([a-z]+):([A-Z]:)((.|\r|\n))^:\2)|([a-z]+):()((?:(?:.|\n|\r)(?!^[a-z]+:))+)) to make it 1/5 & 3/7
- # $1$5\n$3$6\n - shows everything cleanly
Matches everything and gives me victory:
- (?|(?:([a-z]+):([A-Z]*:)((.|\r|\n)*)^:\2)|([a-z]+):()((?:(?:.|\n|\r)(?!^[a-z]+:))+))
- $1 = $3\n\n does a nice print of it

Taeluf.com