Refactor length parsing and make offset a length

[bynect]

Move sanitization from the length parser to the settings file. Also make offset a length with the (N,N) syntax.
offset still accepts the old syntax for backward compatibility.

This will lay the foundations for implementing #991 and dynamic height.

This is a partial reversal of commit acf0a0f, @fwsmit what do you think?

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 78.00000% with 11 lines in your changes are missing coverage. Please review.

Project coverage is 65.95%. Comparing base (07b68f0) to head (9de5769).
Report is 35 commits behind head on master.

Files	Patch %	Lines
src/settings.c	42.85%	8 Missing ⚠️
src/option_parser.c	85.00%	3 Missing ⚠️

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1330      +/-   ##
==========================================
+ Coverage   65.71%   65.95%   +0.23%     
==========================================
  Files          50       50              
  Lines        8041     8209     +168     
  Branches      925      962      +37     
==========================================
+ Hits         5284     5414     +130     
- Misses       2757     2795      +38     

Flag	Coverage Δ
unittests	65.95% <78.00%> (+0.23%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[fwsmit]

Looks fine to me. Don't have time to review the whole thing though

[bynect]

Can I go ahead?

[zappolowski]

I haven't looked into the code in detail yet, but treating offset as a length of (min, max) seems odd as the values stored therein have completely different meaning, though they have the same structure. What is puzzling me is that min from it is just used during parsing for validation and otherwise ignored. If this is the case, why bother having it at all?

Maybe I am missing the bigger picture here.

[bynect]

I haven't looked into the code in detail yet, but treating offset as a length of (min, max) seems odd as the values stored therein have completely different meaning, though they have the same structure. What is puzzling me is that min from it is just used during parsing for validation and otherwise ignored. If this is the case, why bother having it at all?

Maybe I am missing the bigger picture here.

My idea was making in a future refactor of the parser a unified "tuple"-like type. Also it is odd for me that we have the syntax for couple of values but offset is the only one using 1x2.
Also min and max are not really used, it's just a parsing artifact, because it is cast to a struct position. (for the offset)

For the height I wanted to add a min and max because dynamic heights was one of the low hanging fruit I am gonna add soon (aka one of the next prs)