* Remove double encoding in MathML via Unicode MathML spaces
We used to have a complex mechanic for escaping strings, marking them as
`needsEscape = false`, and then not escaping those strings (for
combining strings in `\operatorname`). But this doesn't really work
with `render`ing directly to a node, as `document.createTextNode` can't
be stopped from escaping.
I've thus removed this mechanic, which required the following changes:
* Switch MathML "smart space" encoding to use Unicode instead of
`&LongNames;` (which weren't working with `render` for the same reason).
* Hack our HTML/MathML serializer to not use `String.trim`, which wrecks
havoc with emitted Unicode spaces.
Now `toText()` doesn't escape, so strings concatenate in unescaped form,
and `toHTML()` only does the necessary escaping. Thus fix#1782.
* Fix src/utils.js
Co-Authored-By: edemaine <edemaine@mit.edu>
* Fix src/mathMLTree.js documentation
Co-Authored-By: edemaine <edemaine@mit.edu>
* Remove trim hack thanks to diffable-html@4.0.0
* Switch back to jest-serializer-html
* Update mathMLTree.js
* Fix \\ and \newline after operator
Fix#1790. `\\` and `\newline` render as a span with classes
`mspace` and `newline`. We need to check for `newline` when bringing
spaces into the same `base` group.
* Add tests and comment
* Update comment for defineFunction handler.
* Rename .value to .text in all token-style ParseNodes.
* Rename symbolNode's .value to .text.
* Rename color-token's .value to .color.
* Comments without terminating newlines in nonstrict mode
Fix#1506 by allowing single-line comments (`%` without terminating newline)
in nonstrict mode. `Lexer` and `MacroExpander` now store the `Settings`
object, so the `Lexer` can complain about missing newline according to the
`strict` setting. I filtered this out from the snapshot tests with a slightly
different `replacer`.
* Reimplement \href like \verb, add \url
Major restructuring to lex URL arguments differently, e.g. to support
`\href%{hello}` and `\href{http://foo.com/#test%}{hello}`. The new URL
parsing code is simpler, but involves a special case in `parseSymbol`
like `\verb`.
Also add support for `\url` while we're here.
* Cleanup
* Fix flow errors and improve error messages
* Add \url to documentation
* Improve doc formatting
* Refactor test helpers
* Combine common codes from `parse` and `build`, and dispatch using
`Mode`
* Remove `toNotBuild` and `toNotParse` and use `expect.not`
* Remove `Warning` and instead mock `console.warn`
* Add `expect.toHavePassed` and check whether parse/build succeeded
lazily
* Improve failed test `message`:
- Use color
- Print stack traces(excluding internals)
- Print diff
- Follow jest matcher output style
* Update helpers.js
* Remove toBeTruthy checks
getParsed throws an error if parsing fails.
* Used tagged literals
* Use .toHaveLength
* Use tagged literals
* Use to{Build,Parse}Like where possible
* Remove compareParseTree
* Use snapshot where possible
* Join into one line where <88 chars
* Revert console.warn() to throw an error
Merge `expectToWarn` to `expectKaTeX`, like
`expect.toFailWithParseError`
* Remove call to console.warn from stack traces
* Fix merge errors
* Remove `getTree`
* Move `_getBuilt` into `getBuilt`
* Move default settings and tagging literal support in to `getParsed`
and `getBuilt`
* Remove stack traces clean-up
* Remove `toHavePassed` matcher
* Extract `expected` string construction into `printExpectedResult`
* \html@mathml
Fix#1452
* Add missing file
* Implement \char (via internal \@char)
* Remove excess <mstyle> wrapper on \mathbin etc.
* Fix tests
* Add Unicode support for \copyright and \textregistered
Testing that this doesn't lead to an infinite loop thanks to \char` escaping.
* Add tests
* Use assertNodeType
* Switch from regex to lookup table, and no parseInt
* Automatic mrel/mbin spacing for \boldsymbol
Fix#1154
* Add test
* Add missing mode
* Separate out \boldsymbol support to correct flow types
* Fix ternaries thanks to @ronkok's comment
* Revert package-lock.json
* Add screenshot test
* Add and expose addFontMetrics function
* fix typo
* recreate package.json-lock
* Add setFontMetrics function to change the defualt metricMap - change getFontMetrics to getGlobalMetrics
* use new setFontMetrics on main KaTeX object
* fix package-lock.json error by rebuilding it
* Add appropriate tests
* update the snapshot
* Support \hdashline
Support `\hdashline` from package `arydshln`. This PR acts as a complement to PR #1395.
Similarly to #1395, the dashed line is rendered as `border-bottom-style: dashed;`. That does not exactly match the `dashsegment` and `dashgap` lengths in `arydshln`, but it does render black lines with sharp edges.
* Fixed top line position
* Add screenshots
* Fix lint error
* Fix another lint error
* Deleted HLine screenshots
* Pick up comments
* Fix ligatures not always triggering
Example: \text{\rm --} would not cause ligature but should
* Remove ligatures in \texttt mode
* Add screenshot tests
* Fix MathML ligatures
* Fix type
* Handle \text... fonts in MathML building
* Remove leftover console deubgging
* Support \arraystretch as a macro definition
Also add `expandMacro` and `expandMacroAsText` helpers to `MacroExpander`.
* Remove excess defaulting
* Add test
* Rewrite spacing commands as macros
Fix#1130 by defining `\!`, `\negthinspace`, `\,`, `\thinspace`, `\:`,
`\medspace`, `\;`, `\thickspace`, `\negmedspace`, `\negthickspace`,
`\enspace`, `\enskip`, `\qquad`, `\quad` as macros.
Fix#1036 by defining a new `SpaceNode` in mathMLTree that recognizes
all special space amounts from
https://www.w3.org/TR/2000/WD-MathML2-20000328/chapter6.html
* Fix <mspace> rendering and add a test to catch it
* Update screenshots
* Wrap special space characters in <mtext>
* Update screenshots
* Fix MathML escaping behavior
* Fix flow typos
* Fix Unicode-chrome
* Reimplement mod operators in terms of macros
* Rerun ModScript
* Remove excess table entries (now already macros)
* Fix bmod
* Avoid unnecessary <mrow> wrapping
buildMathML gains two helpers:
* `makeRow` helper wraps an array of nodes in `<mrow>`,
unless the array has length 1, in which case no wrapping is necessary.
* `buildExpressionRow` for common case of `makeRow(buildExpression(...))`
* Combine adjacent <mtext>s in all cases
No more need for `makeTextRow` helper or anything fancy in text MathML handler.
* Concatenate <mn>s and decimal point into single <mn>
Fix#203
* Fix snapshots
* Tag sketch
* Drop objectAssign; already using Object.assign elsewhere
* Basic \gdef support
* Bug fix
* Finish \tag
* MathML numbers equations with <mlabeledtr>
* Fix flow bugs
* \gdef tests
* Add basic \tag tests
* Screenshot test for \tag
* \tag* test
* Add missing file
* Bug fix screenshot
* Major refactor
* Represent tag at top level of parse tree, requiring less hackery
* No more \@tag function; it was essentially just doing \text
* Wrap tag in group so e.g. ( and ) are formatted the same
* Add `feed` method to MacroExpander for multiple inputs (for tag)
* Bug fixes in buildHTML, makeTextRow, _getBuilt (for display mode)
* Remove excess <mrow> wrapper when unnecessary
* Update screenshot from tag being wrapped in group
* Add maxExpand limit
* Move test helpers into common modules
* helpers.js gets all the helper functions
* setup.js gets the common Jest setup (serializer, expect extensions)
* Exclude test from coverage testing
* @ylemkimon's comments: parsing -> building, settings || defaultSettings
* Default argument for settings
* Fix lint errors
* @ylemklemon's comment: use buildAndSetResult
* Use template literals
* Stop throwing ParseError when throwOnError is false
`render`, `renderToString`, etc. now catch ParseError and render it to the
raw LaTeX (with proper escaping) and a hover title with the error.
Along the way:
* Use new `katex.__renderToDomTree` in katex-spec's `_getBuilt`.
(This was necessary to get the new error handling in `_getBuilt`.)
* Fix jest results which must always be functions, not strings.
* fix lint
* Fix flow error
Leave error type unspecified, as we check it with instanceof.
* Update katex-spec.js
* update fonts and metrics so that accents are positioned correctly
* fix degree symbol, move more accents into the 0x2c0-0x2df range
* fix accents in Unicode screenshots
* update jest snapshots, update AccentsText to include requested characters
* update fonts submodule, add unicode chars to AccentsText test
* update submodules after merging its PR
* Unicode accents
* Lexer now looks for combining dicritical marks and adds them to the same character
* Parser's `parseSymbol` now recognizes both combined and uncombined forms of Unicode accents, and builds accent objects just like the accent functions
* Added CJK support to math mode (not just text mode)
* Add invalid combining character test
* Add MathML test
* Add weak support for other Latin-1 characters
This maintains backwards compatibility, but it uses the wrong font.
There's a TODO to fix this later.
Also refactor symbol code to use for..of
* Update Unicode screenshot
* Remove dot from accented i and j (in math mode)
Also add dotless Unicode characters to support some accented i's and j's
* Fix \imath, \jmath, \pounds, and more tests
* Switch from for..of to .split().forEach()
Save around 800 bytes in minified code
* Fix split
* normalize() detection
* Convert back to vanilla for loops
* Fix merge
* Move normalize dependency to unicodeMake.js
* Make unicodeSymbols into a lookup table instead of macros
This is important for multi-accented characters.
* Add comments about when to run
* Move symbols definition into unicodeMake/Symbols.js
* Remove CJK support in text mode
* Add missing semicolon
* Refactor unicodeAccents to its own file
* Dotless i/j support in text mode
* Remove excess character mappings
* Fix Åå in math mode (still via Times)
* Update to support #1030
* Add accented Greek letter support (for supported Greek symbols)
* Update screenshot
* remove Æ, æ, Ø, ø, and ß from math mode test
* added support for \boldsymbol
* added \boldsymbol tests
* added \boldsymbol screenshots
* added \bm, snapshot, updated screenshots
* updated snapshot for \boldsymbol
* properly display \imath and 2 in boldsymbol
* implemented \boldsymbol for +
* fix typo, remove boldsymbolLetters hardcode
* make comment about using Main-Bold more general
* fixed trailing space CI error
This PR implements `\mathchoice` function.
I once created PR on the wrong branch. Sorry for the mess.
This is particularly useful when one defines custom macro for "big operators".
For example:
```latex
\newcommand{\infdisj}{%
\mathop{%
\mathchoice{% display
\bigvee\hspace{-2ex}\bigvee%
}{ % inline
\bigvee\hspace{-1.75ex}\bigvee%
}{ % script
\bigvee\hspace{-1.4ex}\bigvee%
}{ % scriptscript
\bigvee\hspace{-1ex}\bigvee%
}}}
```
* Support \colorbox and \\fcolorbox
These are functions from the `color` package. They accept text, not math. They also have padding similar to `\fbox`.
Because of the padding, the code in `buildHTML` is intermixed with the `\fbox` code. I have not (yet) created a new file in the functions folder. This way, the reviewer gets a cleaner diff.
* Fix lint error
* colorbox screenshots
* Pick up review comments
Summary: (@kevinbarabash)
This diff extracts MathML builder tests out into a separate file and
updates them to use jest snapshot testing. It updates the output of
prime ' from using identifier nodes <mi> to operator nodes <mo>. It
also updates large operators with limits to use munderover instead
of msupsub. I added an option to remove unnecessary mrows to buildGroup.
Right now it's only used for by groupTypes.supsub. I'll see if it can be used
elsewhere (everywhere?) in a follow up PR.
Test Plan:
- make test w/o errors
- verify mathml snapshots contain the desired markup
