Boostnote: Some URL-encoded string can't be decoded correctly

[cignoir]

context

URL including Japanese which is URL-encoded by EUC-JP or Shift-JIS can not be decoded correctly in view mode.

screenshot of Boostnote with developer tool open
https://i.gyazo.com/b3dc7697f51cb2ba8fcebbdeb4022a94.png

environment

• windows10
• boostnote v0.8.15

expected

I don't want to be automatically decoded.

more example

This behavior corrupts some of existing URLs.

Paste on boostnote and a web browser for comparison

http://metasearch.osdn.jp/wiki/index.php?Samba%2BLDAP%2Bsmbldap-tools%A4%C7%A5%EF%A1%BC%A5%AF%A5%B0%A5%EB%A1%BC%A5%D7%A4%CE%C5%FD%B9%E7%C7%A7%BE%DA

[aquibm]

Hey @kazup01, I'd love to give this one a go :)

[kazup01]

👍

[aquibm]

Looks like the artefacts are generated by mdurl's decode function, which is used by linkify-it, which is in turn used by markdown-it to parse links that aren't wrapped in markdown's link syntax.

One way we could get around this weird behaviour would be to retroactively fix the links decoded by mdurl after markdown-it renders the editor content. I've raised #978 to address this issue.

[asmsuechan]

URL including Japanese which is URL-encoded by EUC-JP or Shift-JIS can not be decoded correctly in view mode.

The issue was fixed by https://github.com/BoostIO/Boostnote/pull/816

[kazup01]

Thanks for your contribution @aquibm !