* Improve rule coding.
* Rule widths (overline, underline, fraction, sqrt) actually scale
with the current font size in AMS-TeX. Implement that. (Sqrt is
a special case: the rule width depends on the font size *of the
surd*.)
* Change the CSS coding for rules. The old, complex coding prevented
variable-width lines and may have contributed to issues like #696.
Its purpose, according to 0a3a2271f4,
was IE8 support; but KaTeX no longer supports IE8.
* The 0.5px offset makes larger sizes better, smaller sizes worse.
Smaller sizes seem more important.
* Cleanup (intended to be squashed).
Some delicate surgery is required on the `vlist-s` table-cell CSS.
Problem with #768 reported by @ronkok; issue #779. Fixes:
* `font-size: 1px` (rather than `font-size: 0.05em`) solves render
issues with user overrides of the `.katex` `font-size`.
* Also fix a width issue.
TeX and CSS treat line heights in fundamentally different ways. In
TeX, every character is treated as a box of its precise height and
depth; the line height (\baselineskip) applies after characters have
been assembled into lines. In CSS, in contrast, every character
creates a "line box" corresponding to the accompanying font. When
characters of different fonts and sizes are placed into the same
span, the resulting line box contains the line boxes of all children.
This is unfortunate because, for example, we want `\frac{1}{2}` to
behave in vertical spacing contexts like it is exactly as tall and
deep as the visible fraction (which is the TeX behavior). Given CSS
constraints, though, in most contexts the fraction has extra vertical
space: the line boxes for the numerator and denominator create
padding. For small boxes, this isn't so bad. To really see the
problem put a tall rule in the denominator of a fraction, or check
out the VerticalSpacing screenshotter test, which has way more space
than it should.
Solving this problem in CSS is difficult. There is no easy way to get
rid of the extra line boxes.
But there is *a* way, namely tables. A table-cell with vertical-align
top, bottom, or middle is ignored for the purposes of line height
calculation.
So in this commit, makeVList puts its contents into a
vertical-align:bottom table-cell (to clear unwanted line boxes), with
an extra row used to represent depth.
Many Chrome screenshotter tests change. This is because Chrome rounds
table dimensions to integral numbers of pixels, while it uses
sub-pixel positioning for non-table displayed tabs. That makes many
vlists a fraction of a pixel wider than they used to be.
* Fix interaction between styles and sizes by implementing styles as sizes.
Rather than having both `textstyle` CSS classes and `size5` CSS classes
affect the font size (and step on each other), implement sizes more the
way TeX does: a command like `\displaystyle` changes the current size.
This is actually a simplification, since now only `size` affects the size.
Simplifies CSS and computation. Many screenshotter tests change; they
change to be more like TeX. For instance, `\sqrt` fixes some
discrepancies in size treatment.
Also:
Remove the `Options.withX()` methods in favor of `.havingX()`, which
might return the same `options`.
Remove `Style.cls()` and `Style.reset()`.
Remove `Options.reset()`. You should never modify an `Options`; they
should change only by the `havingX()` methods.
* Implement TeX sizing for scriptsize/scriptscriptsize.
At every size level. Also make the sizes match TeX to the last decimal.
* Review comments.
* Internal: Pass full `options` objects to makeSpan/makeSymbol.
Not just the current color. This will facilitate applying options
to built nodes in a standardized way, rather than changing all
callsites.
* Add style switching test: text and scriptstyle in the same group.
* Apply style-specific spacing using different CSS coding.
Specifically, infer style from a class on the *current* element,
rather than the parent element. Use "mtight" class to denote elements
with tight spacing (scriptstyle or scriptscriptstyle). Apply that
class automatically based on options.
* Fix#533, #534, #541.
- #534: Implement getTypeOfGroup for font groups.
- #533, #541: Improve the ways spaces are applied to lists. Since
CSS adjacency implements mathematical spacing, it's incorrect to
introduce "convenience spans" for spaces and display changes into
the generated HTML -- those spans break adjacency. Apply display
changes directly, and shift space spans into adjacent atoms.
Requires updates to two screenshotter tests, LimitControls and
SupSubLeftAlignReset. The new results for these tests are closer
to TeX output than the old results.
Also requires updates to Jasmine tests, since those assumed output
structures that have changed.
* Fix#136: Size commands generate fragments, not spans.
This is so the size commands don't hide the types of their enclosed
atoms. Addresses #136.
This slightly changes the vertical position of the Sizing test. Not
sure the vertical position matters, so change the test.
Summary:
This diff provides support for Latin-1, Cyrillic, and CJK characters
inside \text{} groups. For Latin-1 and Cyrillic characters we use
glyph metrics from a glyph from Basic Latin that has roughly the same
bounding box. We use the metrics for a capital 'M' to approximate the
full-width CJK characters. Half-width characters are not supported yet.
Test Plan:
- make test
- make screenshots
Reviewers: emily
It looks like the `text-align: center` is affecting the text in sub and
superscripts. Fixes#447
Test Plan:
- Visit [this
example](http://localhost:7936/?text=x%5El_%7Bi%5E%7Bl%2B1%7D%2Bi%2C%20j%5E%7Bl%2B1%7D%2Bj%2C%20d%7D)
- Edit the HTML to add `<span class="katex-display">...</span>` around the
`<span class="katex">` node.
- See that the sub and superscripts are left-aligned, not centered
(It looks like we don't have a way to test this in the screenshotter for now)
@kevinb
Summary: For some reason, adding `border-style: solid` also adds a 3px
border around elements, which means that all of the rules that we
created are 3px too large. This sets the default size to 0 for all the
edges, which makes them correct.
Test plan:
- See that `a\rule{0em}{0em}b` produces no visible rule.
- See the new screenshots look reasonable.
This adds the ability to add `|` to a column description and have
vertical separators be added. I added types to the column descriptions
and added some logic to handle the separators when building the vertical
lists of the array.
Test plan:
- See the Arrays screenshot looks good.
- `make test`
This commit introduces environments, and implements the parser
infrastructure to handle them, even including arguments after the
“\begin{name}” construct. It also offers a way to turn array-like data
structures, i.e. delimited by “&” and “\\”, into nested arrays of groups.
Environments are essentially functions which call back to the parser to
parse their body. It is their responsibility to stop at the next “\end”,
while the parser takes care of verifing that the names match between
“\begin” and “\end”. The environment has to return a ParseResult, to
provide the position that goes with the resulting node.
One application of this is the “array” environment. So far, it supports
column alignment, but no column separators, and no multi-column shorthands
using “*{…}”. Building on the same infrastructure, there are “matrix”,
“pmatrix”, “bmatrix”, “vmatrix” and “Vmatrix” environments. Internally
these are just “\left..\right” wrapped around an array with no margins at
its ends. Spacing for arrays and matrices was derived from the LaTeX
sources, and comments indicate the appropriate references.
Now we have hard-wired breaks in parseExpression, to always break on “}”,
“\end”, “\right”, “&”, “\\” and “\cr”. This means that these symbols are
never PART of an expression, at least not without some nesting. They may
follow AFTER an expression, and the caller of parseExpression should be
expecting them. The implicit groups for sizing or styling don't care what
ended the expression, which is all right for them. We still have support
for breakOnToken, but now it is only used for “]” since that MAY be used to
terminate an optional argument, but otherwise it's an ordinary symbol.
Summary:
Use the TeX definitions of `\root` to get the optional `\sqrt`
argument in the right place. Also add the MathML version.
Fixes#48
Test Plan:
- `make test`
- See that the images look good
Reviewers: kevinb, alpert
Reviewed By: alpert
Differential Revision: https://phabricator.khanacademy.org/D17236
Summary: Make the CSS rules that hide the MathML rendering not
`!important`, so that they can be overridden. Also, remove the useless
`math.katex` rule.
Fixes#197
Test plan:
- `make test`
- See that huxley screenshots haven't changed
Auditors: alpert
Summary:
This adds support for rendering KaTeX to both HTML and MathML
with the intent of improving accessibility. To accomplish this, both
MathML and HTML are rendered, but with the MathML visually hidden and
the HTML spans aria-hidden. Hopefully, this should produce much better
accessibility for KaTeX.
Should fix/improve #38Closes#189
Test Plan:
- Ensure all the tests, and the new tests, still pass.
- Ensure that for each of the group types in `buildHTML.js`, there is a
corresponding one in `buildMathML.js`.
- Ensure that the huxley screenshots didn't change (except for
BinomTest, which changed because I fixed a bug in `buildHTML` where
`genfrac` didn't have a `groupToType` mapping).
- Run ChromeVox on the test page, render some math. (for example,
`\sqrt{x^2}`)
- Ensure that a mathy-sounding expression is read. (I hear "group
square root of x squared math").
- Ensure that nothing else is read (like no "x" or "2").
- Ensure that MathML markup is generated correctly and is interpreted
by the browser correctly by running
`document.getElementById("math").innerHTML =
katex.renderToString("\\sqrt{x^2}");` and seeing that the same speech
is read.
Reviewers: john, alpert
Reviewed By: john, alpert
Subscribers: alpert, john
Differential Revision: https://phabricator.khanacademy.org/D16373
Summary:
Add the ability to pass in options to the render calls which contain information about the parse. This information is passed around to the parser and builder, which parse and render differently depending on the options. Currently, this includes an option to render the math in display mode (i.e. centered, block level, and in displaystyle).
Also added some changes to make it easier to add new data to functions (now that new data doesn't need to be copied into the ParseFuncOrArg data structure, it is looked up when it is needed) and has more sane support for the `'original'` argType (as suggested by pull request #93).
Test Plan:
- Make sure tests and lint pass
- Make sure huxley screenshots didn't change, and new screenshot looks correct
Reviewers: alpert
Reviewed By: alpert
Differential Revision: https://phabricator.khanacademy.org/D13810
Summary:
Add correct parsing of optional arguments. Now, things like `\rule` can shift
based on its argument, and parsing of `\sqrt[3]{x}` fails (correctly) because we
don't support that yet.
Also, cleaned up the lexing code a bit. There was a vestige of the old types in
the lexer (they have now been completely moved to symbols.js). As a byproduct,
this made it hard to call `expect("]")`, because it would look at the type of
the Token and the type for "]" was "close". Now, all functions just look at the
text of the parsed token, and in special occasions (like in the dimension lexer)
it can return some data along with it.
Test Plan:
- Make sure tests still work, and new tests work
- Make sure no huxley screenshots changed
- Make EXTRA SURE `\sqrt[3]{x}` fails.
Reviewers: alpert
Reviewed By: alpert
Differential Revision: http://phabricator.khanacademy.org/D13505
Summary:
Create a fonts.less file which generates identical css to fonts.css, but using
less rules to be more understandable and customizable. For example, add the
ability to change where the fonts directory is located (instead of mandating it
be located next to the less file), and add the ability to disable specific font
formats (like disable EOTs when IE8 support isn't needed).
Test Plan:
- Ensure that the test page and huxley page still work
- Ensure that the output of `./node_modules/.bin/lessc static/fonts.less` is
the same as the original css by running both through
`./node_modules/.bin/cleancss` and diffing them.
- Ensure that the huxley screenshots haven't changed
- Ensure that the build step still works
Reviewers: alpert
Reviewed By: alpert
Differential Revision: http://phabricator.khanacademy.org/D13326
Summary:
Improve the build process by combining the CSS files, automatically
selecting font files to include, and building .tar.gz and .zip files.
Test Plan:
- Make sure tests work
- Make sure huxley screenshots didn't change
- Make sure zip and tar.gz files contain only files we want to distribute
Reviewers: alpert
Reviewed By: alpert
Differential Revision: http://phabricator.khanacademy.org/D13159
Summary:
Add support for math-mode accents. This involves a couple changes.
First, in order to correctly position the accents, we must know the kern between
every character and the "skewchar" in that font. To do this, we improve our tfm
parser to run the mini-kern-language and calculate kerns. We then export these
into fontMetrics.js.
Then, we add normal support for accents. In particular, we do some special
handling for supsubs around accents. This involves building the supsub
separately without the accent, and then replacing its base with the built
accent.
Finally, the character in the fonts for the \vec command is a combining unicode
character, so it is shifted to the left, but none of the other characters do
this. We add some special handling for \vec to account for this.
Fixes#7
Test Plan:
- Make sure tests pass
- Make sure no huxley screenshots changed, and the new one looks good
Reviewers: alpert
Reviewed By: alpert
Differential Revision: http://phabricator.khanacademy.org/D13157
Summary:
Add support for all of the other operators, including the ones with symbols and
limits. This also fixes the bug where subscripts were shifted the same amount as
subscripts.
To accomplish this, the domTree.textNode has been repurposed into symbolNode
which is no longer an actual text node, but instead represents an element with a
single symbol in it. This lets us access properties like the italic correction
of a symbol in a reasonable manner without having to recursively look through
children of spans.
Depends on D13082
Fixes#8
Test Plan:
- Make sure tests work
- Make sure huxley screenshots didn't change much, and new screenshot looks good
Reviewers: alpert
Reviewed By: alpert
Differential Revision: http://phabricator.khanacademy.org/D13122
Summary:
Add a way to automatically build vlists correctly. Previously, we built
vlists manually in ~4 different places, which made it difficult to manage
changes, and led to a large amount of duplication in the less. This also fixes
the vlist construction in safari, where the `display: inline-table` wasn't being
applied because of CSS specificity. This leads to the only significant change in
the huxley tests, with the vertical spacing.
Test Plan:
- Make sure the tests still work
- Make sure most of the huxley screenshots didn't change, and that the new
changes are insignificant.
- Make sure vlists now work in Safari
- Make sure the change to the VerticalSpacing screenshot is caused by the
fix-baseline span now correctly applying `display: inline-table` by creating
the construct in master and adding `display: inline-table !important` to the
`.fix-ie` css rule
Reviewers: alpert
Reviewed By: alpert
Differential Revision: http://phabricator.khanacademy.org/D13082
Fix the \frac and \overline lines so they actually render at the correct sizes
(i.e. don't get smaller when in script and scriptscript sizes). This also seems
to fix#33.
Test Plan:
- Make sure new huxley screenshot look good.
- Render
\frac{1}{\Bigl(\sqrt{\phi \sqrt{5}}-\phi\Bigr) e^{\frac25 \pi}} =
1+\frac{e^{-2\pi}} {1+\frac{e^{-4\pi}} {1+\frac{e^{-6\pi}}
{1+\frac{e^{-8\pi}} {1+...} } } }
at small font sizes like 30pt and smaller, see that it looks much better
Auditors: alpert
Summary:
Follow the TeXbook instructions on how to construct square roots. Using
makeCustomSizedDelim, this becomes nearly trivial.
Test Plan:
- Make sure normal tests work
- Make sure the new huxley test looks good, and other huxley tests haven't changed.
Reviewers: alpert
Reviewed By: alpert
Differential Revision: http://phabricator.khanacademy.org/D12918
Summary:
Added stacked delimiter support for more delimiters. Split out delimiter
functions into its own file, and split out some tree building functions into a
common file. Supports the empty `.` delimiter with \left and \right, and doesn't
try to produce huge /, \backslash, <, or > delimiters. Depends on D7844
Test input:
\left( \left) \left[ \left\lbrack \left] \left\rbrack \left\{ \left\lbrace
\left\} \left\rbrace \left\lfloor \left\rfloor \left\lceil \left\rceil
\left\langle \left\rangle \left/ \left\backslash \left| \left\vert \left\|
\left\Vert \left\uparrow \left\Uparrow \left\downarrow \left\Downarrow
\left\updownarrow \left\Updownarrow {x^{x^{x^{x^{x^{x^{x^{x^{x^{x^x}}}}}}}}}}
\right.\right.\right.\right.\right.\right.\right.\right.\right.\right.
\right.\right.\right.\right.\right.\right.\right.\right.\right.\right.
\right.\right.\right.\right.\right.\right.\right.\right.
Test Plan:
- Run the test input, see that it works
- Run the tests, see that they work
- Look at huxley screenshots (not here yet :( ) and make sure they look good
Reviewers: alpert
Reviewed By: alpert
Differential Revision: http://phabricator.khanacademy.org/D11602
Summary:
I don't know why we originally used 2/3em for scriptstyle, but both TeX
and MathJax use 0.7em for the scriptstyle size.
Test Plan:
- Look at huxley tests, make sure everything that changed is due to the change
in font size
Reviewers: alpert
Reviewed By: alpert
Differential Revision: http://phabricator.khanacademy.org/D12869
Summary:
Instead of using the metrics from our generated TTF files, use the fonts that
TeX ships with. Pull the mapping out of the MathJax-dev repo from makeFF to get
the correct mapping of metrics to font characters, and use our own tfm reader to
extract metrics out of the tfm files into a useable format. Add a README and
Makefile rule to make this process easier in the future.
Also remove the silly 0.05em we put on supsubs because our italic correction
works now.
Test Plan:
- Run huxley tests, see that changes are because of font metric changes.
- See that the extension piece of `\bigl |` now extends above the top, as it is
supposed to.
Reviewers: alpert
Reviewed By: alpert
Differential Revision: http://phabricator.khanacademy.org/D12867
Summary:
Supports the 'ex' and 'em' units for sizes. Doesn't support the optional depth
argument.
Test Plan:
- See that the huxley test looks good, and nothing else changed
- See that the tests pass
Reviewers: alpert
Reviewed By: alpert
Differential Revision: http://phabricator.khanacademy.org/D12777
Summary:
In LaTeX, large delimiters are the same font size as they are at a normal size,
regardless of the actual size. This means that we need to scale up the font size
in the inner nodes, which is annoying because we run into the same problem we
had with \Huge, etc in those nodes. Thus, this fixes both problems at once.
The problem was that when we used our baseline-align-hack and then increased the
font size inside of one of the middle (display: block and height: 0) nodes, the
node with the increased font size would shift downards (misaligning its
baseline). To fix this, we add a method for calculating the maximum font size
used in each of the nodes, and adding a small node with this font size to each
of the other nodes (including the fix-ie node). This shifts all of the nodes
down the same amount, and gets their baselines aligned.
Test Plan:
- Do dumb things by putting \Huge and \big in places they shouldn't be, and
make sure they behave responsibly
- Do the same thing in IE 8, 9, 10, 11, Safari, Firefox, and make sure they all
behave the same (to some approximation)
- Make sure the new huxley image looks good, and the images that changed don't
have significant changes
Reviewers: alpert
Reviewed By: alpert
Differential Revision: http://phabricator.khanacademy.org/D12684
Summary:
Also, rename all of our uses of fonts to use the uppercased versions. We want to
use the uppercase versions because it makes updating and modifying the fonts
much easier (since the font names inside the actual font files are uppercased).
Test Plan:
- Make sure the huxley screenshots look good (You can compare a diff of them
on github at
f90d093361
By my eye, it seems like some things have moved up ~1/2 pixel, and some of
the fonts have maybe slightly changed shape, like the large `b` in
SizingBaseline)
Reviewers: alpert
Reviewed By: alpert
Differential Revision: http://phabricator.khanacademy.org/D11979
Summary:
Follow the instructions in the TeX book for drawing \overlines. This uses the
same code as fractions to produce the bars. Also added the ability to cramp
styles (i.e. T -> T' and T' -> T').
Test Plan:
- Look at `\overline{x}`, `\overline{\dfrac{x}{y}+z}`, and
`\blue{\overline{x}}` to make sure they look good.
- Make sure the tests work
- Make sure the huxley tests look good (Not here yet :( )
Reviewers: alpert
Reviewed By: alpert
Differential Revision: http://phabricator.khanacademy.org/D11604
Match the naming styles used for the css styles in the variable names.
Test Plan:
- Make sure `x\Huge x \tiny x \Huge x` renders correctly.
- Make sure the huxley tests stay the same
Auditors: alpert
Summary:
Make delimiter sizing work. This involved
- Adding the symbols for the remaining delimiters (like `\lfloor` and `\{`)
- Adding metrics for the size1, size2, size3, and size4 fonts
- Parsing delimiter sizing functions
- Using the big fonts when possible, otherwise building large copies of the
delimiters from scratch
Test Plan:
- See that
`\bigl\uparrow\Bigl\downarrow\biggl\updownarrow\Biggl\Uparrow
\Biggr\Downarrow\biggr\Updownarrow\bigm/\Bigm\backslash\biggm|
\Biggm|\big\lceil\Big\rceil\bigg\langle\Bigg\rangle\bigl(\Bigl)
\biggl[\Biggl]\Biggr\{\biggr\}\Bigr\lfloor\bigr\rfloor`
parses correctly (this contains all of the delimiters, and all of the sizing
modes)
- See that the huxley tests didn't change, and the new one looks good
- See the normal tests work
Reviewers: alpert
Reviewed By: alpert
Differential Revision: http://phabricator.khanacademy.org/D7844
Summary:
Implicit groups are objects that act like groups but don't have brackets around
them. This is used for things like sizing functions or font-change functions
that can occur in the middle of the group, but act like they apply to a group
after them which stops when the current group stops. E.g. `Hello {world \Large
hello} world` produces normal, normal, Large, normal text. (Note, I just came up
with the name implicit group, I don't think this is actually how it is parsed in
LaTeX but it fits nicely with the current parsing scheme and seems to work
well).
For now, apply this to the sizing functions (we don't have any other functions
that act like this). Also note that this doesn't really do much practically
because we limit sizing functions to only be on the top level of the expression,
but it does let people do `\Large x` and get a large `x`, without having to add
braces.
Test Plan:
- Run the tests, see they work
- Make sure `abc \Large abc` looks correct
Reviewers: alpert
Reviewed By: alpert
Differential Revision: http://phabricator.khanacademy.org/D10876
Summary:
For some reason, `vertical-align: top` doesn't work in IE9, or at least doesn't
work the same as in any other browser (I'm too lazy to figure out what the
correct behavior really should be). Giving a literal value for `vertical-align`
works the same in all browsers, so use that instead.
Test Plan:
- Replace line 13 of static/index.html with
<br>m<span id="math"></span>m
- See that the test page now looks good in Chrome, Firefox, and IE9
- See that huxley tests haven't changed
Reviewers: alpert
Reviewed By: alpert
Differential Revision: http://phabricator.khanacademy.org/D8384
For some reason, when you have a nested elements that look like
`display: inline-block; > position: relative; > position: absolute;`
then the `position: absolute;` element is shifted down a bunch. If there is
anything* else inside either of the other two elements, then this behavior
disappears. (This can be seen at [this fiddle](http://jsfiddle.net/qZXRr/). We
have this structure when we create `\llap` and `\rlap`s, and this weird behavior
So, to fix this I added an empty `display: inline-block;` span inside the llap
to fix this.
Test plan:
- See that the new huxley image looks good
- Test this in a bunch of browsers and see they also look good
Auditors: alpert
Summary:
Keep track of the color inside the style now, and use that when we are
rendering things. Added a custom lexing mode to handle lexing colors correctly.
Prefixed the old katex colors (internally) with "katex-" to distinguish them
from CSS colors.
Test Plan:
- Run the normal tests, see they work
- Run the huxley tests, see that they didn't change except for the color one
which looks right
Reviewers: alpert
Reviewed By: alpert
Differential Revision: http://phabricator.khanacademy.org/D7763
Summary:
Remove a single `vertical-align: top`, and somewhow it now works. May
the gods of CSS have mercy on us. Also added some tests.
Test Plan:
- See that the huxley tests don't have any changes
- See that the new huxley screenshots look reasonable
- Run the normal tests and see that they work
Reviewers: alpert
Reviewed By: alpert
Differential Revision: http://phabricator.khanacademy.org/D7494
Summary:
TeX spacing is complicated - there's \enskip and \enspace and the two are
subtly different in some weird way. But \enspace seems to be the preferred
half em space in LaTeX, and the only one which MathJax supports.
Test Plan: Parsed and rendered correctly.
Reviewers: alpert
Reviewed By: alpert
Differential Revision: http://phabricator.khanacademy.org/D6437
Test Plan: Parses and renders locally. \!\, and \,\! are now no-ops.
Reviewers: alpert
Reviewed By: alpert
Differential Revision: http://phabricator.khanacademy.org/D6395
