Code review

[mojavelinux]

• enable web fonts in stylesheet
• refactor conversion and rendering code
  • allow XHR to work for local files (status = 0)
  • prevent flash of unrendered content
  • use clearer ipc message name
  • improve method names
  • better error message when XHR fails
• bump max Firefox version
• add dollar sign before command in README
• set safe to server by default to enable includes
• set implicit attributes (no need to disable toc)

[mojavelinux]

I absolutely could not figure out how to use nsParserUtils from the content script. I guess this just isn't possible. The benefit would be that we could get a document fragment back when sanitizing instead of a string. I suppose it's plenty fast to do it the way we are doing, and technically more correct.

I noticed, however, that the Markdown Viewer is able to use the nsParserUtils because they don't use contentScript. (See https://github.com/Thiht/markdown-viewer/blob/master/chrome/content/markdown-viewer.js) I'm not sure if that's better. Is it more correct to use contentScript for an extension?

[ggrossetie]

Thanks that's way better now!

I absolutely could not figure out how to use nsParserUtils from the content script. I guess this just isn't possible.

Same here.

The benefit would be that we could get a document fragment back when sanitizing instead of a string.

Yes but I asked Kris Maglione (the reviewer) and he's OK with the sanitize method.

I noticed, however, that the Markdown Viewer is able to use the nsParserUtils because they don't use contentScript. (See https://github.com/Thiht/markdown-viewer/blob/master/chrome/content/markdown-viewer.js) I'm not sure if that's better. Is it more correct to use contentScript for an extension?

I don't know, I used the same "architecture" as the Chrome extension with a "background" script and a "content script". I didn't know that we can retrieve the document content from the main script (i.e. have access to the DOM of the page).