Parsing the filepath: regex vs custom parser

[yjmantilla]

So for parsing the filepath I originally had in mind something like this functionality from mp3tag:

Basically the user inputs the pattern using percentages to delimitate the fields.

This is also possible with regex like this:

Each strategy has their strength and weaknesses:

• custom parser is more intuitive
• regex can parse more complex patterns, for example when the pattern is the number of characters instead of a separator
• custom parser cannot distinguish if there is no separator between the fields (ie in 'i1s001rest' it cannot separate 001 from rest)
• regex needs the user to input in another variable the fields (in order) as they are parsed on the string

As of now I'm supporting the custom parser but I think supporting regex will allow more flexibility for power users.

[stebo85]

I vote for using an already established technique like regex because people already know this and creating your own parser always looks easier than it is in the end

[yjmantilla]

@stebo85 I had already implemented the custom parser; it works well enough, but the point of regex being the standard is a good argument. I do find the custom parser more intuitive for non-technical users.

I will try to implement the regex one and maybe put the whole parsing backend through regex, that is , translating the percentage notation of the custom parser to a regex pattern instead of being a parser itself. That way all is done through regex and the percentage notation is optional and only for non-technical users.

[yjmantilla]

closing this since it is already implemented in regex and the percentage notation translation to regex is also already implemented.