feat(cli): --trust flag for trusting cli input

[ngzhian]

The trust option has much finer grained control, which is hard to
specify on the command line, so expose just a single boolean flag, which
changes trust from false (default) to true.

[edemaine]

Is this a work-in-progress? The new option doesn't seem to be getting used/checked anywhere.

I also wonder whether --trust-all (or even --trust) would be a better spelling for the option. There's only one input in an execution of the CLI, so it doesn't seem to make sense to call out inputs.

[codecov-commenter]

Codecov Report

Merging #3339 (5269823) into master (b1e24f2) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master    #3339   +/-   ##
=======================================
  Coverage   93.54%   93.54%           
=======================================
  Files          88       88           
  Lines        6552     6552           
  Branches     1509     1509           
=======================================
  Hits         6129     6129           
  Misses        393      393           
  Partials       30       30           

---

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 b1e24f2...5269823. Read the comment docs.

[ngzhian]

Is this a work-in-progress? The new option doesn't seem to be getting used/checked anywhere.

Is there a particular place I should add the test to? I'm not sure where such cli tests should go.

I also wonder whether --trust-all (or even --trust) would be a better spelling for the option. There's only one input in an execution of the CLI, so it doesn't seem to make sense to call out inputs.

--trust-all looks good, I can change it

[edemaine]

I wasn't talking about tests. I mean that you never actually look at options.trustAllInputs, so your code doesn't "do" anything, other than accept a new argument. The trust option remains unaffected.

[ngzhian]

I wasn't talking about tests. I mean that you never actually look at options.trustAllInputs, so your code doesn't "do" anything, other than accept a new argument. The trust option remains unaffected.

Right, thanks! I will rename it to trust, then program.parse(process.argv) should take care of it, since SettingOptions.trust takes a boolean as well.

[edemaine]

Cool, this does seem consistent with how --strict works. By that analogy, perhaps we should add -T too?

[ngzhian]

Done!

[edemaine]

I just tested that this works for me (with a \url input). Awesome!

I forgot but confirmed that the documentation is generated automatically from cli.js itself. See https://deploy-preview-3339--katex.netlify.app/docs/cli.html

I just revised the documentation; assuming you approve, I'll merge.

Update: I found PR #2839 which is identical, so I'll add @summersz as a coauthor when merging.

[ngzhian]

Thanks for improving the documentation and finding the older issue, looks great, please merge whenever convenient, thanks!

[KaTeX-bot]

🎉 This PR is included in version 0.13.19 🎉

The release is available on:

• npm package (@latest dist-tag)
• GitHub release

Your semantic-release bot 📦🚀

[edemaine]

@ngzhian Thanks for the PR!