Implement \neq and \notin via macros like LaTeX's

[edemaine]

As discussed in KaTeX/katex-fonts#47 (comment), I simplified LaTeX's definitions of \neq and \notin into KaTeX macros.

They seem visually very close to LaTeX's; here's a texcmp for the tweaked Not test (only third line is new):

I assume this will remain true after #1267 but it would be good to test.

This PR removes the need for the \neq and \notin symbols from the fonts, so it would fix KaTeX/katex-fonts#47

The new \neq should be essentially the same as the font-based one. Here is a comparison of the new \notin symbol (top) vs. the old \notin symbol (bottom):

[codecov]

Codecov Report

Merging #1400 into master will increase coverage by <.01%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master    #1400      +/-   ##
==========================================
+ Coverage   82.33%   82.33%   +<.01%     
==========================================
  Files          77       77              
  Lines        4239     4241       +2     
  Branches      731      731              
==========================================
+ Hits         3490     3492       +2     
  Misses        647      647              
  Partials      102      102

Impacted Files	Coverage Δ
src/symbols.js	100% <ø> (ø)	⬆️
src/macros.js	88.28% <100%> (+0.26%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 97d6229...9b8abc7. Read the comment docs.

[kevinbarabash]

I think this is a move in the right direction. I'm confused though why the slant of the slash is different in \notin in the third line especially if \notin is now defined as \not\in now.

[kevinbarabash]

Shouldn't the / be a \not? It seems weird to use a different glyph for \notin.

[ronkok]

@kevinbarabash If I remember correctly, \notin was created shortly before \not was available.

[edemaine]

This is what LaTeX does; see the definition in comments. I believe it's a conscious decision because the \in glyph is relatively narrow. So I think we should follow suit.

[kevinbarabash]

They seem visually very close to LaTeX's; here's a texcmp for the tweaked Not test (only third line is new):

SGTM.

[edemaine]

...unless there's another LaTeX package that redefines \notin differently? I know AMS doesn't redefine it. I'm on a phone so can't easily check symbols-a4.

[edemaine]

There is a rather large variety of \notin symbols listed in symbols-a4, from packages fdsymbol, stix, txfonts, mathabx, MnSymbol.

Not sure what to make of all this. I guess I lean toward matching LaTeX's default behavior, but I'm also happy to design a \not-based macro if that's the preference.

[kevinbarabash]

I'd lean towards LaTeX's default behaviour as well.

[ronkok]

For what it's worth, here is the MathML from (1) font-based \notin and (2) macro-based \notin.

<mrow>
  <mo>∉</mo>
</mrow>

<mrow>
  <mstyle>
    <mrow>
      <mo>∈</mo>
    </mrow>
    <mpadded lspace="-1width" width="0px">
      <mrow>
        <mi mathvariant="normal">/</mi>
        <mspace width="0.05555555555555555em"></mspace>
      </mrow>
    </mpadded>
  </mstyle>
</mrow>

[edemaine]

@ronkok Yes, this is a downside to the macro approach in general. I was thinking we could fix all of these with a \mathml{LaTeX code used for HTML}{MathML equivalent} command.

[ronkok]

Now that would be cool.

[kevinbarabash]

LGTM. Thanks for posting a texcmp screenshot.