Version 8 roadmap (tracking issue)

[KillyMXI]

Goals

• Selectors support, to address the biggest user requests;
• When implemented it, I also figured how to improve the weird base elements selection feature (some breaking changes but hopefully it makes more sense now);
• Introduce compile and convert functions (an attempt to gain back some performance in batch processing jobs).

Related issues and PRs

• Feature Request: ignoreSelector Array #159
• [Feature] Ability to mark elements to be hidden from the text version #179
• Can support css rules? #143 (partly)
• selector for baseElement #96
• CSS selectors only match the first found element #215
• Ignore selectors #186

Checklist

• develop a suitable selectors parser (parseley);
• develop a selectors decision tree (selderee);
• code changes in html-to-text;
• deprecated options handling, tests, documentation, ...;
• push the changes to the repo (selectors branch);
• collect initial feedback;
• publish if everything is fine (probably in a week)
  • update version in package.json and changelog;
  • bump dependencies and regenerate package-lock.json;
  • publish the package;
• collect post-launch feedback.

Looking for feedback

(I'm no longer highlighting this issue. But if you wish I'm still happy to hear any feedback - in comments here or as a separate issue.)

Performance

In case you can benchmark the batch processing time on a large enough practical dataset, I am curious how it differs between 7.1.1 and 8.0.0.

Anything unclear

Let me know if you find any changes in configuration and documentation confusing (weird configuration format, weird defaults, unclear instructions, whatever).

Web bundlers experience

Works out of the box / needs a workaround, configuration tweaking / etc; with what framework/bundler version...
It would be helpful to map out this area before I dig into it myself and can provide the support outside of Node environment.

[Niek]

Wow, that's a lot of changes (and work involved), great job! :) I will do some tests on my codebase shortly and report my findings here.

[rootwork]

This is working great for me. I'm using it on MJML to remove the mobile navigation icons (hidden with CSS, but present in the HTML) with the following:

selectors: [{ selector: 'div.mj-menu-trigger', format: 'skip' }],

Much cleaner than the regex I was using before.

Thanks for all your hard work making this happen!

[KillyMXI]

@Niek Got anything to report?
I'll proceed with the release tomorrow.
If you haven't got around to do anything - no big deal.
It would be a shame if you actually put some effort and it's going to be wasted due to miscommunication.

[Niek]

Sorry for the delay @KillyMXI! I tested and can confirm it works great! I'm testing with a bunch of (fairly advanced) selector like blockquote[class^="twitter-tweet"] and I can skip them properly.
This is much easier than my previous workaround of re-loading the output in JSDom and removing the selectors with a manual loop.

[KillyMXI]

Awesome. Thanks everyone.
Version 8 is now live.

[dargmuesli]

@webstech would you mind upgrading the type definitions over at https://github.com/DefinitelyTyped/DefinitelyTyped/tree/master/types/html-to-text to v8? <3

[KillyMXI]

By the way, there is another issue related to type definitions: #223

[webstech]

@webstech would you mind upgrading the type definitions over at https://github.com/DefinitelyTyped/DefinitelyTyped/tree/master/types/html-to-text to v8? <3

Working on it. Wishing I had documented the process better - it was a mix of generated and hand-crafted code. The typescript code generator is having issues. Probably something on my end.

[webstech]

Working on it. Wishing I had documented the process better - it was a mix of generated and hand-crafted code. The typescript code generator is having issues. Probably something on my end.

PR 53812 has been opened. There is a commit for the v7 changes and another for v8. Deprecated code is still supported so there should be no breaking changes.

Thanks for bringing this to my attention.

fyi, I have updated my notifications for this repo to all activity so I won't miss issues.

[KillyMXI]

All in all, this was quite a quiet release.
Still some great things have happened.
Thanks everyone who helped to bring @types/html-to-text up to date.
The package also got a mention in JavaScriptDaily.

Time to close this issue. (That doesn't mean I stop listening for feedback though.)

[dargmuesli]

Thanks for your dedicated work!

[webstech]

The package also got a mention in JavaScriptDaily.

Do you have a link for this? My searches did not go well. Thanks.

[Niek]

The package also got a mention in JavaScriptDaily.

Do you have a link for this? My searches did not go well. Thanks.

https://twitter.com/JavaScriptDaily/status/1402980878230376449?s=20