About page-break

[Turtle91]

Originally Posted by jhowell

While I agree with your stance on property names that begin with "-webkit-" and other prefixes, if you look at the cited appendix it becomes clear that "Prefixed properties" in this case refers to property names that begin with the "-epub-" prefix.

I noted that as well. However, the spec is talking about ePub so it would make sense to only demo -epub- prefixes. In the real world the devices/apps DO support prefixes of all types to some degree or another. The argument was being made that using prefixes wasn't allowed...that is not true.

Best practice would be to use the non-prefix if it's supported, prefixed if necessary. Either way is legal and epubcheck doesn't give a fit over their use.

To Jon's latest goal post movement - ignoring an unsupported style is exactly what epub is supposed to do. That is not broken. That is what coder's rely on when they create epubs that are backwards compatible...it's all part of the fallback coding concept.

[JSWolf]

Originally Posted by Turtle91

I noted that as well. However, the spec is talking about ePub so it would make sense to only demo -epub- prefixes. In the real world the devices/apps DO support prefixes of all types to some degree or another. The argument was being made that using prefixes wasn't allowed...that is not true.

Best practice would be to use the non-prefix if it's supported, prefixed if necessary. Either way is legal and epubcheck doesn't give a fit over their use.

To Jon's latest goal post movement - ignoring an unsupported style is exactly what epub is supposed to do. That is not broken. That is what coder's rely on when they create epubs that are backwards compatible...it's all part of the fallback coding concept.

But due to older versions of RMSDK ignoring the entire CSS when it thinks there is an error, you have to take that into account. We have a number of people using devices with older versions of RMSDK and some are being sold now. We have Kobo & B&N being sold with older RMSDK. With the nook, you have no choice. We Kobo, you may or may not have a choice. The thing is, most users would not know this. All they would think is that there is something wrong with the eBook.

[Turtle91]

Originally Posted by JSWolf

But due to older versions of RMSDK ignoring the entire CSS when it thinks there is an error, you have to take that into account. We have a number of people using devices with older versions of RMSDK and some are being sold now. We have Kobo & B&N being sold with older RMSDK. With the nook, you have no choice. We Kobo, you may or may not have a choice. The thing is, most users would not know this. All they would think is that there is something wrong with the eBook.

You keep forgetting the word “sheet” when you say this. Those broken devices might ignore the entire css “sheet” if it sees something it doesn’t like… that’s why people actually have 2 (or more) css sheets where they put the styles that those old, broken, devices might barf on…. This also part of that backwards compatible - fallback coding that good coders plan for.

[DNSB]

Originally Posted by JSWolf

But due to older versions of RMSDK ignoring the entire CSS when it thinks there is an error, you have to take that into account. We have a number of people using devices with older versions of RMSDK and some are being sold now. We have Kobo & B&N being sold with older RMSDK. With the nook, you have no choice. We Kobo, you may or may not have a choice. The thing is, most users would not know this. All they would think is that there is something wrong with the eBook.

Testing on my Kobo Forma's version of RMSDK, it will disregard the entire stylesheet if a structural/syntax error is found (missing ;, missing { or }, etc.). For an unknown property, it properly treats that one item/line as a null operation while the rest of the stylesheet is used.

[JSWolf]

Originally Posted by DNSB

Testing on my Kobo Forma's version of RMSDK, it will disregard the entire stylesheet if a structural/syntax error is found (missing ;, missing { or }, etc.). For an unknown property, it properly treats that one item/line as a null operation while the rest of the stylesheet is used.

I just took the eBook Voyage of the Damned from StandardEBooks. As it comes, it's broken in an older RMSDK. But if you remove two simple things that aren't actually needed it works well enough.

What I'm pointing out is just because something is valid in epub3 does not mean it will work in an older RMSDK.

/* Invert images in dark mode. RMSDK requires a target media as well as a state. */
@media all and (prefers-color-scheme: dark){
	img.epub-type-contains-word-se-image-color-depth-black-on-transparent{
		filter: invert(100%);
	}

	img.epub-type-contains-word-se-image-color-depth-black-on-transparent.epub-type-contains-word-se-image-style-realistic{
		background: currentColor;
		filter: none;
	}
}
/* ...or if the device supports prefers-color-scheme. We’ll invert the image in core.css. RMSDK requires a target media as well as a state. */
@media all and (prefers-color-scheme){
	img.epub-type-contains-word-se-image-color-depth-black-on-transparent:not(.epub-type-contains-word-se-image-style-realistic){
		background: transparent !important;
	}
}

[Turtle91]

Originally Posted by JSWolf

I just took the eBook Voyage of the Damned from StandardEBooks. As it comes, it's broken in an older RMSDK. But if you remove two simple things that aren't actually needed it works well enough.

What I'm pointing out is just because something is valid in epub3 does not mean it will work in an older RMSDK.

/* Invert images in dark mode. RMSDK requires a target media as well as a state. */
@media all and (prefers-color-scheme: dark){
	img.epub-type-contains-word-se-image-color-depth-black-on-transparent{
		filter: invert(100%);
	}

	img.epub-type-contains-word-se-image-color-depth-black-on-transparent.epub-type-contains-word-se-image-style-realistic{
		background: currentColor;
		filter: none;
	}
}
/* ...or if the device supports prefers-color-scheme. We’ll invert the image in core.css. RMSDK requires a target media as well as a state. */
@media all and (prefers-color-scheme){
	img.epub-type-contains-word-se-image-color-depth-black-on-transparent:not(.epub-type-contains-word-se-image-style-realistic){
		background: transparent !important;
	}
}

That’s cool - but the standard has been there to accept prefixes since early on in ePub coding - and it is still valid today.

You should always test on your targeted market’s device(s) and code for proper fall-back options to accommodate older devices to a point.

[JSWolf]

Originally Posted by Turtle91

That’s cool - but the standard has been there to accept prefixes since early on in ePub coding - and it is still valid today.

You should always test on your targeted market’s device(s) and code for proper fall-back options to accommodate older devices to a point.

If that code was in a commercial ePub, it's entirely possible that is could be purchased and used by people who have a nook or a Kobo. Besides, that code is not needed at all. Once you get into the chapter 1, there are no images.

[Turtle91]

Whatever Jon - you do you. Just don’t tell people they can’t use prefixes.

[JSWolf]

Originally Posted by Turtle91

Whatever Jon - you do you. Just don’t tell people they can’t use prefixes.

Use them now, regret it later.

[Turtle91]

Originally Posted by JSWolf

Use them now, regret it later.

Yeah, not that worried about it. As has been mentioned multiple times, it will ignore a css it doesn’t support… so it’s not going to break anything to include it… you certainly won’t have to go back and fix them later. And if you include both versions, the prefixed and non prefixed, the device/app will simply use the one it supports…

They have a term for that…. Let me think about it… oh yeah:

Fallback coding

Might want to look into that concept!

[JSWolf]

If it's made invalid at some point, epubcheck will call it an error.

[Turtle91]

By your logic: If nook or kobo are made invalid at some point….

That’s a nonsensical argument.

It is not now, nor has it been flagged by epubcheck as an error.

Just admit when your argument is wrong instead of changing your argument…

Prefixes are allowed.

If you have the option to use a non prefix then use them both to cover all your bases.

[JSWolf]

Originally Posted by Turtle91

By your logic: If nook or kobo are made invalid at some point….

That’s a nonsensical argument.

It is not now, nor has it been flagged by epubcheck as an error.

Just admit when your argument is wrong instead of changing your argument…

Prefixes are allowed.

If you have the option to use a non prefix then use them both to cover all your bases.

If youi have to use a prefix instead of the proper ePub code, then your program is buggy.

[icearch]

Standards are created and maintained by mortal not god. What works are the true gold standard.

[RbnJrg]

Originally Posted by JSWolf

If youi have to use a prefix instead of the proper ePub code, then your program is buggy.

Totally wrong! A prefixed property is a proper one and the epub is not buggy for containing it. It's not wrong to include it (according to the current epub standard, these are valid properties). Where is the truth? In what the standard states or in your words? You're the one who's wrong, and you don't have the humility to admit your mistake.