Collapse all sequences of contiguous space into a single space.
Group boundaries break contiguity.
This is a transformation instead of part of the parser since we may want to
output a tree that is relatively faithful to the input.

Parse a TeX 'control sequence', which can be a 'control symbol' or a 'control word'.
Returns a Keyword.

Parse a TeX 'control symbol', i.e., the escape character + another non-letter character.
Subsequent whitespace is not consumed. Returns a string.

Parse a TeX 'control word', i.e., the escape character + a sequence of letter characters.
All whitespace after the control word is immediately absorbed. Returns a string.

Flatten the list of nodes, removing commands and blocks throughout;
flatten this node into a string, or nil.
Flatten the node at loc, removing it if it's empty, replacing it with its
children if there are any, and removing any commands outright.

Unescape all accents in TeX node tree.
If loc points to an accent command, remove it and insert the appropriate
combining character after the following (non-whitespace) token
Take a node in the TeX tree and if it's an accent command (1-character keyword)
that has a corresponding combining character, return that character.

Replace each control sequence in a TeX tree that represents a specific
character with that character.

Parse and simplify the TeX reader into a string of TeX

Parse and simplify a TeX string into a simplified string of TeX

Replace fancy characters, then unescape accents, then flatten

Parse a TeX comment. For now, immediately discard the contents, returning only the empty string.

Handle operations on TeX nodes

Set of the characters:
Horizontal Tab (U+0009)
Line Feed (U+000A)
Vertical Tab (U+000B)
Form Feed (U+000C)
Carriage Return (U+000D)
Space (U+0020)

Write TeX-formatted output to a java.io.Writer.

Convert TeX tree into TeX-formatted string