Question

Satisfactory that is to not go into functions that eat much time.

Format of the data:
It is basically similar to the HTML tag parameters, but not quite.

<name1=value1 name2=value2 [here can be any whitespace characters]
33=55 "name with spaces and ="="value with spaces and '" '"title"'="can't" "in
names"="and
values can be tabs and new lines"
double"and_single'quotes=are_allowed_in_names_and_values_if_there_is_no_whitespace_in_it
"escaping is ""sql-like"""='the same with ''single quotes'''
`third option`=`are "backticks" that also can be ``escaped```>

Is this format called somehow ? I ‘invented’ it myself but I highly doubt that no one came up with this idea before, maybe it (or very similar format, for example with slash escaping) already has a parsing function in php ?

If not, how to write my own, using regexes, string operations, some kind of buffers, accumulators ? I’ve never written any parser myself yet.

Answer 1

Try something like this:

$in = <<<BLOCK
name1=value1 

name2=value2 [here can be any whitespace characters]

33=55 

"name with spaces and ="="value with spaces and '" 

'"title"'="can't" 

"in names"="and
  values can be tabs and new lines"

double"and_single'quotes=are_allowed_in_names_and_values_if_there_is_no_whitespace_in_it

"escaping is ""sql-like"""='the same with ''single quotes'''

`third option`=`are "backticks" that also can be ``escaped```
BLOCK;

$string = "(?:                    # open non-capture group
             \"(?:\"\"|[^\"])+\"  # match a double quoted string
           | '(?:''|[^'])+'       # OR match a single quoted string
           | `(?:``|[^`])+`       # OR match a back-ticked string
           )                      # close non-capture group
          ";

$key    = "(                      # start capture group 1
             $string              # match any string
           | [^=\s]+              # OR match one or more chars other than '=' and space-chars
           )                      # end capture group 1
          ";

$value  = "(                      # start capture group 2
             $string              # match any string
           | [^\r\n]+             # OR match one or more chars other than '=' and space-chars
           )                      # end capture group 2
          ";

$regex = "/$key=$value/x"; // combine the patterns

preg_match_all($regex, $in, $matches);

$key_value_pairs = sizeof($matches[0]);

for($index = 0; $index < $key_value_pairs; $index++) {
  echo "===============================================\n" . 
       "KEY:\n" . $matches[1][$index] . "\n" .
       "VALUE:\n" . $matches[2][$index] . "\n";
}

which will print the following:

===============================================
KEY:
name1
VALUE:
value1 
===============================================
KEY:
name2
VALUE:
value2 [here can be any whitespace characters]
===============================================
KEY:
33
VALUE:
55 
===============================================
KEY:
"name with spaces and ="
VALUE:
"value with spaces and '"
===============================================
KEY:
'"title"'
VALUE:
"can't"
===============================================
KEY:
"in names"
VALUE:
"and
  values can be tabs and new lines"
===============================================
KEY:
double"and_single'quotes
VALUE:
are_allowed_in_names_and_values_if_there_is_no_whitespace_in_it
===============================================
KEY:
"escaping is ""sql-like"""
VALUE:
'the same with ''single quotes'''
===============================================
KEY:
`third option`
VALUE:
`are "backticks" that also can be ``escaped```

Use, and maintain, at your own risk!