escape tag for url special character

[kevchentw]

Issue

The generate process failed when tag cantains special character (e.g. "/")

Reproduce

1. set a tag with "/" such as "CI/CD" on notion
2. run notablog generate .
3. Error as below

E/notablog-cli(20200505T234305.045): Error: ENOENT: no such file or directory, open 'public/tag/CI/CD.html'
    at Object.openSync (fs.js:457:3)
    at Object.writeFileSync (fs.js:1282:35)
    at /home/kevink_chen/.nvm/versions/node/v12.16.1/lib/node_modules/notablog/src/render-index.js:34:8
    at Map.forEach (<anonymous>)
    at renderIndex (/home/kevink_chen/.nvm/versions/node/v12.16.1/lib/node_modules/notablog/src/render-index.js:27:19)
    at generate (/home/kevink_chen/.nvm/versions/node/v12.16.1/lib/node_modules/notablog/src/generate.js:90:3)
    at processTicksAndRejections (internal/process/task_queues.js:97:5)
    at async cmdGenerate (/home/kevink_chen/.nvm/versions/node/v12.16.1/lib/node_modules/notablog/bin/cli.js:28:5) {
  errno: -2,
  syscall: 'open',
  code: 'ENOENT',
  path: 'public/tag/CI/CD.html'
}

Solution

replace special character with "-" by using squirrelly filter

reference:

• https://developers.google.com/maps/documentation/urls/url-encoding
• https://www.geeksforgeeks.org/replace-special-characters-in-a-string-with-underscore-_-in-javascript/

[dragonman225]

Hi, this regex [&\/\\#, +()$~%.'":*?<>{}] seems to effect a lot of characters, is it really necessary?
And does it means that CI-CD instead of CI/CD will be displayed in the tag, not just the URL and the filename being changed?
Personally, I suggest using encodeURIComponent() for the URL and the filename, and HTML entities for the displayed text.

[kevchentw]

The tag will display as "CI/CD" since I only escape href only.
<a href="tag/{{@this.value|escapeTag}}.html">{{@this.value}}</a>

The reason for replacing with "-" is for the url readability
e.g.
/tag/CI-CD.html -> replace
/tag/CI%252FCD.html -> encodeURIComponent (encode two times)

For the regex, we can preserve necessary characters such as "%", "/", "?", "&", "#". But I need to check out which character will effect the url links.

[dragonman225]

The tag will display as "CI/CD" since I only escape href only.
<a href="tag/{{@this.value|escapeTag}}.html">{{@this.value}}</a>

This sounds great! At first, I was worried that changing the displayed text might confuse the user.

The reason for replacing with "-" is for the url readability
e.g.
/tag/CI-CD.html -> replace
/tag/CI%252FCD.html -> encodeURIComponent (encode two times)

For the regex, we can preserve necessary characters such as "%", "/", "?", "&", "#". But I need to check out which character will effect the url links.

Agreed. Readability is a good reason to adopt character replacing instead of encodeURIComponent.