[WIP] Allow using Firefox / WebDriver BiDi protocol during conversion #597
Conversation
yhatt
changed the title
| }) | ||
|
|
||
| try { | ||
| await page.emulateMediaType('print') |
There was a problem hiding this comment.
WebDriver BiDi does not yet support media type emulation, but it's OK because using bare template, that is almost compatible with the print media, whenever converting to image(s).
| return await sharp()(buf) | ||
| .jpeg({ mozjpeg: true, quality: opts.quality }) | ||
| .toBuffer() |
There was a problem hiding this comment.
Use sharp to convert PNG to JPG because WebDriver Bidi does only support PNG screenshot.
- ToDo
- Check compatibility with the standalone binary
| this._firefoxPDFConversionWarning = true | ||
|
|
||
| warn( | ||
| 'Using Firefox to convert Markdown to PDF: The output may include some incompatible renderings compared to the PDF generated by Chrome.' |
There was a problem hiding this comment.
Currently, PDF conversion by Firefox has some incompatible rendeings compared to PDF generated by Chrome. e.g. There are some vanished elements in PDF, such as the background gradient like gaia theme, the element with CSS filter, etc.
➡️ Should treat PDF conversion backed by Firefox as fallback
| return await page.screenshot({ | ||
| clip, | ||
| quality: opts.quality, | ||
| type: 'jpeg', | ||
| omitBackground: true, |
There was a problem hiding this comment.
It is only supported by CDP. WebDriver BiDi will not work #555.
Related: #565
Replaces the browser management for converting with new one, implemented in #595. That supports using Firefox, and WebDriver BiDi protocol for Chrome.
This PR will replace only the browser engine. I'm going to work for new options about new browser manager and finders in another pull request.