public final Class TokenFixer

Useful helper. Takes a "builder" that has been accumulating tokens. Groups all tokens into a CommandToken containing the accumulated tokens as a single argument and clears the builder.

An empty group is legal and will result in a "fake" empty token assuming the same Slice and LaTeXMode as parentToken.

Returns:CommandToken

null if the builder is empty, otherwise grouped CommandToken

This looks for an outermost pairs of \left and \right and converts them to a more tree-like fence token.

Mismatched pairs will cause an error, as they do in LaTeX.

See Also

inferParenthesisFences(Token, List)

Converts leading occurrences of '-' followed by a MathNumberInterpretation into a single token representing the negation of the given number.

Helper to fix the representation of a list environment. The original content should be of the form: \item ..... \item ..... \item ...... (and maybe include CorePackageDefinitions#CMD_LIST_ITEMs as well) We replace with a more tree-like version. If \item is used outside a suitable environment, then it will be left as-is. The DOMBuilder will cope with this in due course.

This handles the old-fashioned "... \over ..." by refactoring the tokens into a \frac{...}{...}. As with LaTeX, we only allow one \over in a single level.

Hunts through tokens for occurrences of primes used as f', which are converted to a superscript by binding to the preceding token

We'll look for any occurrences of T1_T2[^T3] or T1^T2[_T3], where [..] denotes "optional". If found, we'll replace with the more tree-like equivalents: \msub{T1}{T2} \msup{T1}{T2} \msubsup{T1}{T2}{T3} which will be easier to handle later on.

Helper to fix the content of tabular environments to make it more clearly demarcated as rows and columns. This kills off instances of CorePackageDefinitions#CMD_CHAR_BACKSLASH and TokenType#TAB_CHARACTER within tables/arrays and replaces then with zero or more CorePackageDefinitions#CMD_TABLE_ROW each containing zero or more CorePackageDefinitions#CMD_TABLE_COLUMN.

Infers explicit paragraphs by searching for the traditional LaTeX TokenType#NEW_PARAGRAPH and/or CorePackageDefinitions#CMD_PAR tokens, replacing with the more tree-like CorePackageDefinitions#CMD_PARAGRAPH.

This attempts to group (not necessarily matching) pairs of open and close brackets into a corresponding CorePackageDefinitions#ENV_BRACKETED environment that's easier to handle.

It also handles brackets which are not correctly nested, using fake empty opener/closer delimiters at the start/end of the group as required.

Note that this means that while $[1,2)$ will successfully be matched, the notation $[1,2[$ will *not* be considered matched, even though it is common in subjects like Mathematical Analysis. To get the correct semantics here, you must use $\left[1,2\right[$ and let fencePairedParentheses(Token, List) take care of this for you.

See Also

fencePairedParentheses(Token, List)

Replaces elements in the tokens List with the given replacement Token, starting at the given startIndex (inclusive) up to endIndex (exclusive).

The computed style of the replacement token is set to be that of the first token it replaces.

LR Mode doesn't let "block" stuff do its normal thing. We strip out any "new paragraph" markers and fail on any other kind of block tokens.

Strips out redundant whitespace text tokens at the start and end of a List of siblings and between a pair of "block" tokens.