zmarkdown: highlight latex and keep console as plainText

[artragis]

#257

[artragis]

I don't know why tests are failing here.

[vhf]

Something changed in an upstream package we transitively depend on: syntax-tree/mdast-util-to-hast@742a780

This fix makes remark-parse more compliant with the commonmark spec. We could either hold this dependency back or upgrade, I think it's better to upgrade. I'll ping you here as soon as it's fixed, you'll then have to rebase on master.

[vhf]

I'll prepare this on a separate branch, I might wait on syntax-tree/mdast-util-to-hast#22 before merging it.

[vhf]

I decided to pin the dependency instead. You can now rebase on master.

[artragis]

@vhf it's rebased.

[vhf]

From what I read here, only the \ is highlighted, end{document} is not highlighted.

This seems wrong. TeX highlighting should at least detect {} and its content, and highlight this.

[vhf]

If you have issues reading snapshots, you don't have to use snapshots.

Or before using a snapshot, you could check what the snapshot would be. Try .toBe(null) instead of .toMatchSnapshot(), it will usually tell you that you expected the output to be null but got … instead.

Read carefully what … is, and if you are 100% sure it is correct, use a snapshot.

Or use whatever trick you think of, the goal is to commit good code, not to follow a certain procedure.

[artragis]

output was good, and in latex, but input was not good due to the fact I use template and did not escape { and }

[vhf]

There is no need to escape {} in a template string.

Please run the following code:

const dedent = require('dedent')

const input = dedent `\`\`\`latex
\\usepackage{inputenc}[utf8]
\\begin\{document\}
\\texttt\{code\}
\\end\{document\}
\`\`\``

and check that this is the input your test is supposed to test.

[artragis]

th'ats what I did in my last update and as you can see it's ok.

[vhf]

It doesn't look right.

Why did you have to escape here: \{document\} but not here: {inputenc}?

[vhf]

Let's go back to my previous comment, I think you should check that the input of your test is correct:

const dedent = require('dedent')

const input = dedent `\`\`\`latex
\\usepackage{inputenc}[utf8]
\\begin\{document\}
\\texttt\{code\}
\\end\{document\}
\`\`\``

Is this the markdown string you want to test?

Here is the last line of your input: (a tag, a name, a tag, a name, a tag, a name)

<span class="hljs-tag">\<span class="hljs-name">\</span></span>end<span class="hljs-tag">\<span class="hljs-name">{</span></span>document<span class="hljs-tag">\<span class="hljs-name">}</span></span>

I think it should be: (a tag, a name, a string)

<span class="hljs-tag">\<span class="hljs-name">end</span><span class="hljs-string">{document}</span></span>

[artragis]

Ok, so,

const input = dedent `\`\`\`tex
\\usepackage{inputenc}[utf8]
\\begin{document}
\\texttt{code}
\\end{document}
\`\`\``

I got the same result as when I ask for latex.

So two possibilities :

• the output is really what we expect
• the input is not good, but I don't get where.

[artragis]

It was the second option.

I had a look on how we test it on lowlight and rehype-highlight side and I just adapted the test with it.

It seems that I do not master the string format management in javascript. Can you explain it please?

When I did:

const input = dedent `\`\`\`latex
\\usepackage{inputenc}[utf8]
\\begin{document}
\\texttt{code}
\\end{document}
\`\`\``

the result was not the same as

const input = ['````latex',
      '\\usepackage{inputenc}[utf8]',
      '\\begin{document}',
      '\\texttt{code}',
      '\\end{document}',
      '```'].join('\n')

[vhf]

Good job, that's much better :)

See how dedent behaves like String.raw:

> const dedent = require('dedent')
> dedent(`\\`)
'\\'
/* careful here: tagged template: */
> dedent`\\`
'\\\\'
> String.raw`\\`
'\\\\'

I'll rework the tests a bit to fix the formatting, add a test, remove an extra ` here: ```` and I'll merge.

[vhf]

Merged: bee1f54