Skip to content
Commits on Feb 29, 2016
Commits on Feb 26, 2016
  1. Merge pull request #533 from mathquill/feature.drop_embedded

    committed Feb 26, 2016
    Added experimental feature dropEmbedded
  2. Merge pull request #532 from mathquill/build.make_server_message

    committed Feb 26, 2016
    Add console output to show url
  3. @stufflebear
  4. Fix semantic merge conflict between #528 and #495 #ImAnIdiot

    committed Feb 25, 2016
    488ebca on #495 changed the Controller
    constructor and tests had to change. #528 was based on an old branch,
    copied some test boilerplate that should also have changed but didn't.
Commits on Feb 25, 2016
  1. @stufflebear
  2. @stufflebear

    Fixed merge conflicts

    stufflebear committed Feb 25, 2016
  3. @stufflebear

    Merge pull request #528 from mathquill/fix.textblock-cursor-movement

    stufflebear committed Feb 25, 2016
    Fix TextBlock cursor movement bug
  4. @stufflebear
  5. Merge pull request #529 from mathquill/fix.text_fractions

    committed Feb 24, 2016
    Wrap math in a fraction in parenthesis for text mode
  6. @stufflebear
Commits on Feb 24, 2016
  1. Fix #429 can't move cursor left out of TextBlock

    committed Feb 24, 2016
    Enforce contract that by the time .blur() on a node is being called, all
    has been said and done and the cursor has already been moved out of the
    Note that the main other case that calls .blur() is intentional blur,
    where already the cursor is hidden before .blur() is called on its
    Supersedes and closes #516 and #430.
  2. Merge pull request #526 from mathquill/fix.text_caret

    committed Feb 24, 2016
    Changed ** in text mode to ^
Commits on Feb 23, 2016
  1. @stufflebear

    Changed ** in text mode to ^

    stufflebear committed Feb 23, 2016
  2. @jwmerrill

    Merge pull request #525 from mathquill/

    jwmerrill committed Feb 23, 2016
    Fix Tab while there's a selection
  3. Fix selecting then Tabbing away and back then selecting

    committed Feb 23, 2016
    Steps to repro:
    1. select stuff
    2. while there's a selection, Tab away and back
    3. at this point the math field should be focused and have a blinking
       cursor and no selection. Do Shift-Left or Shift-Right
    The resulting selection will be quite wrong. This is because the
    anticursor used is from the selection that was cleared when Tabbing
Commits on Feb 20, 2016
  1. Use GitHub Releases instead of downloads page

    committed Feb 20, 2016
    Why maintain our own downloads page now that GitHub releases exists?
    Fix #521
  2. Fix Tab while there's a selection not Tabbing away

    committed Feb 19, 2016
    By "Tabbing away" I mean that if something is focused, and you hit Tab,
    you expect focus to jump to the next thing in tab order.
    Normally that happens in MathQuill editable fields whenever the cursor
    is in the root block/at the top level. But, a bug was intro'd by #294
    and reported in #413 that if there's a selection, then even if you're in
    the root block, when you hit Tab, focus wouldn't actually jump to the
    next thing. Instead, the math field would look blurred but the hidden
    textarea would "steal back" focus, and you could type in the seemingly-
    blurred MathQuill editable field, see GIF:
    (The "Show Textareas" button in the top-right corner of test/visual.html
     was used to show the normally hidden textareas.)
    Note that this bug also affected API clients who in a handler like
    `upOutOf` that's called synchronously on `keydown`, blur the textarea
    (for example, by focusing something above the math field), which is
    exactly what #355 was supposed to fix. Turns out the test case in #355
    isn't representative, the fix only seemed to work because for
    jQuery ≤1.8.3 (our tests use 1.7.2), when you call jQuery's `.blur()`
    method, the `blur` event handlers are called while the hidden textarea
    is still focused, so the `textarea[0].select()` doesn't "steal back"
    focus. However, when Tabbing away, or when the `upOutOf` or whatever
    handler focuses something else, or when you call jQuery ≥1.9.0's
    `.blur()` method, or (luckily for our tests) when you call the DOM's
    "native" `.blur()` method, the hidden textarea is already blurred, so
    the `textarea[0].select()` "steals back" focus, hence bug and fix.
    Isolated test case demonstrating this for jQuery 1.8.3 vs jQuery 1.9.0
    vs the DOM's "native" `.blur()` method:,output
Commits on Feb 18, 2016
  1. Merge branch 'master' into dev

    committed Feb 18, 2016
    	#484 fixed some broken links that were gone/already fixed
  2. Merge pull request #457 from mathquill/docs.changelog

    committed Feb 18, 2016
    Changelog/Migration guide for v0.9.4->v0.10.0
  3. Kill nonworking IRC link in README

    committed Feb 17, 2016
    GitHub-Flavored Markdown apparently doesn't allow `irc://` links
Commits on Feb 17, 2016
  1. Merge pull request #509 from mathquill/fix.latex-breaking-changes

    committed Feb 17, 2016
    Fix LaTeX breaking changes
  2. Merge pull request #506 from mathquill/rm.bad-latex

    committed Feb 17, 2016
    Kill \caret and \underscore
  3. Add back some pre-#247 nonstandard operator names

    committed Feb 17, 2016
    This is actually all of the nonstandard ones that aren't \mod (which is
    a real LaTeX command that's pretty different) and aren't like \asin,
    \acos etc.
    Along with the previous commit, this further reduces breaking of
    backcompat with LaTeX exported before #247.
  4. Accept nonstandard operator names as LaTeX commands eg \arcsec

    committed Feb 17, 2016
    They're still exported as correct LaTeX, e.g. \operatorname{arcsec}.
    We used to accept but also export these, and in #301 we fixed our export
    to only be correct LaTeX (i.e. use \operatorname{} for stuff like
    'arcsec' that, unlike \sin, isn't built into LaTeX) but also stopped
    accepting them when importing. That needlessly broke backcompat with
    the LaTeX that we output before #301, there's no conflict with any real
    LaTeX commands, so now restore backcompat.
  5. Add back \editable as alias for \MathQuillMathField

    committed Feb 16, 2016
    Why unnecessarily break backcompat, it doesn't conflict with any real
    LaTeX commands.
  6. Add back \lbrack and \rbrack, but as plain symbols

    committed Feb 16, 2016
    ...that are just aliases for the '[' and ']' vanilla symbols.
    Before #262, \lbrack and \rbrack were parsed in the same totally broken
    way that \lparen, \lbrace etc were.
    \lparen is not a real LaTeX command and since it was so broken I highly
    doubt anyone was relying on how it was parsed, so it's okay that it was
    removed in #262.
    However, unlike \lparen but like \lbrace, \lbrack and \rbrack are real
    LaTeX commands [1]. So, like we did for \lbrace and \rbrace in 38c2b09
    on #323, add vanilla symbols for them.
Commits on Feb 16, 2016
  1. Kill \caret and \underscore

    committed Feb 16, 2016
    As noted in 0afc6ab on #323, neither of these are real LaTeX commands.
    The real LaTeX command for an underscore character is simply `\_`:
    That'd be great except that there's no easy way for MathQuill to parse
    `\_` differently from plain `_` because the way the parser works,
    there's just one registry entry `LatexCmds._` that parses both.
    There is no one real LaTeX command for a caret character. This post
    suggests 8 ways to show various caret-like characters in real LaTeX:
    Only 4 of those 8 ways work in MathJax, and only one of those 4 is a
    normal caret character:
    Only 2 work in KaTeX v0.5.1, and neither are normal caret characters.
    The thing that works in MathJax is `\verb`, whose name is short for
    "verbatim" and is essentially the LaTeX equivalent of HTML's <pre>,
    which is really not worth our time fully implementing (which is why
    KaTeX doesn't have it).
    Simpler to just have no way of displaying a caret character, like KaTeX.
Commits on Feb 13, 2016
Something went wrong with that request. Please try again.