MobileRead Forums - View Single Post

KevinH · 08-12-2016, 10:24 AM

What you are seeing here is that the BV (QtWebkit widget) is allowing the ul tag to be inside a p tag which is not allowed by the current standard. The difference is that some browser engines support it and some do not but officially it is not allowed.

In BV you selected a part of a paragraph and then changed it to be a bulleted list. The QtWebkit widget allowed this change and shows it the way you would expect but it is not valid html (nor valid xhtml) and so when you click to actually see the tags in CodeView, it forces you to properly fix it (either manually or automatically) before it will load it back into the BV widget.

The real issue is that you have selected part of the contents of a tag not knowing that this was the case because BV does not show you the tags, and then changed it to something else. There is no way for Sigil to detect you have not selected an entire paragraph and so it tries to do what you ask. But the internal QtWebkit is being more "forgiving" than what it should.

If you really care about the actual code generated, then only use BookView to correct single text words as you proof read. Everything else should be done in CodeView.

If you do not know anything about how to generate an ordered list using xhtml, then you need to get a good resource and learn how to do that.

If you do NOT care about the generated xhtml, then simply enable automatic correction. That warning about losing data is a bit absurd now that Sigil has moved away from Tidy to Gumbo (as it is autocorrecting parser just like most browsers) as there have been no documented cases of gumbo actually losing anything (yet!).

You appear to turn off autocorrection and then select partial tag structures in BV and using the tools and then complaining when it generates bad code that Sigil could automatically fix for you if you would simply let it.

KevinH

08-12-2016, 10:24 AM	#7
KevinH Sigil Developer Posts: 8,816 Karma: 6000000 Join Date: Nov 2009 Device: many	What you are seeing here is that the BV (QtWebkit widget) is allowing the ul tag to be inside a p tag which is not allowed by the current standard. The difference is that some browser engines support it and some do not but officially it is not allowed. In BV you selected a part of a paragraph and then changed it to be a bulleted list. The QtWebkit widget allowed this change and shows it the way you would expect but it is not valid html (nor valid xhtml) and so when you click to actually see the tags in CodeView, it forces you to properly fix it (either manually or automatically) before it will load it back into the BV widget. The real issue is that you have selected part of the contents of a tag not knowing that this was the case because BV does not show you the tags, and then changed it to something else. There is no way for Sigil to detect you have not selected an entire paragraph and so it tries to do what you ask. But the internal QtWebkit is being more "forgiving" than what it should. If you really care about the actual code generated, then only use BookView to correct single text words as you proof read. Everything else should be done in CodeView. If you do not know anything about how to generate an ordered list using xhtml, then you need to get a good resource and learn how to do that. If you do NOT care about the generated xhtml, then simply enable automatic correction. That warning about losing data is a bit absurd now that Sigil has moved away from Tidy to Gumbo (as it is autocorrecting parser just like most browsers) as there have been no documented cases of gumbo actually losing anything (yet!). You appear to turn off autocorrection and then select partial tag structures in BV and using the tools and then complaining when it generates bad code that Sigil could automatically fix for you if you would simply let it. KevinH Last edited by KevinH; 08-12-2016 at 10:26 AM.