I do I turn off the escaping/quoting about C# for a name field?

[diberry]

The original TOC looks like:

- name: LUIS Documentation
  href: index.yml
- name: Overview
  expanded: true
  items:
  - name: C#
    href: what-is-luis.md

I want to manipulate the yaml so I change in into JSON with load. Then I check and manipulate (business logic) then I want to dump. The dump adds quotes around the C# which I don't want. Is there an option to turn that off?

• name: LUIS Documentation
  href: index.yml
• name: Overview
  expanded: true
  items:
  • name: "C#"
    href: what-is-luis.md

[puzrin]

Dumper has some euristic and tries to minimize quoting/escaping automatically. But it's not ideal. This can be "improved" via PR only.

[diberry]

@puzrin Can I create PR to control it through options?

[puzrin]

If you can improve existing algorythm - PR will be accepted with high probability.

If you need option - it's better discuss it first to save your time (i don't understand why option needed), but note, i'm against "kludges".

At first glance, string with # is quoted, because # means comment in yaml. But probably it worth to skip if # is not prepended with space and so on. I don't remember exact condition (you will need read yaml spec to not break something else)

[diberry]

The Microsoft Docs system doesn't use quotes around characters and phrases that this tool is adding quotes to for escaping. The build system is separated from authoring so the author's don't have to know about escaping or quotes in the same way.

I need the markdown to not be changed from the original in order to make use of this tool's dump function.

[puzrin]

This package is about YAML, not something else. Dumper MUST provide guarantee that you can load data back and it will not become broken. So, it tries to not use quoting when possible. But sometime false positives possible because algorythm is not ideal.

Just removing quotes everywhere is not acceptable because that will break spec. But you can try to improve existing euristic to reduce false positives.

[diberry]

@puzrin I meant yaml, sorry, too may formats in my head right now. I'll investigate the spec to see where the limitations are.

[puzrin]

Keep in mind, this parser is focused on speed, it does NOT have AST and does not keep src formatting details. For example, you can write in source "C\#" or "C#", but "\" will be lost after dump if not required.

In general, this can not be fixed right without full rewrite. But specific things like quoting can be improved.