Merge markups: Extend features

Now that we have a new script to marge/remove markups, we want to extend its features.

1. Support multi-line markups.
   • Requires a different way to read the .tex files. Currently the script scans one line at a time. We need to either manually acknowledge line feeds by ourselves, which could cause cross-platform troubles, or use a queue to store the text being processed.
2. Support nested markups.
   • We need a stack (or stack array) to make the script stateful.
3. Mooar braces.
   • Include braces that are not changes markups.
   • Mix them with line feeds, markup braces, and other braces.
   • ¿When there's a lexical error in the input file, should we report an error, or just crash the program?
4. Sort markups in the order as they show up in the text, rather than in categories.
   • This helps with the queue/stack idea.
5. Properly treat % comments.

This is an expansion/continuation of Issue #47 (closed), and calls for an example covered by Issue #49.