alphabetizing stylesheet, check book, and remove unused styles

[rjwse@aol.com]

I alphabetize the stylesheet using 'dummy' commands. Here are some examples. I would appreciate comments. Best regards, Pop

[JSWolf]

In English please.

[hobnail]

Seems like a bad idea if I understand correctly. See also this thread:

https://www.mobileread.com/forums/sh...d.php?t=326852

[rjwse@aol.com]

Originally Posted by JSWolf

In English please.

Sorry for not being clear. Here are two screenshots with further explanation. Best regards, Pop

[JSWolf]

Originally Posted by hobnail

Seems like a bad idea if I understand correctly. See also this thread:

https://www.mobileread.com/forums/sh...d.php?t=326852

Once you remove the unused CSS and are left with a much smaller CSS, sorting is usually OK. I've sorted CSS where I then needed to fix a few things. That was easy to do. Most publisher CSS is not doing anything fancy enough where sorting would be an issue.

[rjwse@aol.com]

Originally Posted by JSWolf

Once you remove the unused CSS and are left with a much smaller CSS, sorting is usually OK. I've sorted CSS where I then needed to fix a few things. That was easy to do. Most publisher CSS is not doing anything fancy enough where sorting would be an issue.

You are right. Most CSS in books is simple and minimum. I am pushing the extremes of fanciness, as much as I can, but it comes at a price of garishness.
When you ask calibre to remove unused code, I think it leaves alone any line with multiple items, so long as at least one of them is being used:

div, ol, p, ul, .class1, .class2 {whatever}

So far as I can tell, it will not remove any of the above so long as one of them is in the book. It would be great if it could remove the dead-wood items from a list.

Best regards, Pop

[Brett Merkey]

When you ask calibre to remove unused code, I think it leaves alone any line with multiple items, so long as at least one of them is being used:

div, ol, p, ul, .class1, .class2 {whatever}
So far as I can tell, it will not remove any of the above so long as one of them is in the book. It would be great if it could remove the dead-wood items from a list.

I think I may have already posted this, but you can unstack those selectors in the CSS with a regular expression:

--unstack selectors so Calibre "Remove unused rules" tool can be used:
Search: ([.]?[^ ]+),\s([^{]+)(\{[^}]+\})
Replace: \1 \3\n\n\2\3

'Replace All' until 0 returns.

[JSWolf]

Originally Posted by rjwse@aol.com

You are right. Most CSS in books is simple and minimum. I am pushing the extremes of fanciness, as much as I can, but it comes at a price of garishness.
When you ask calibre to remove unused code, I think it leaves alone any line with multiple items, so long as at least one of them is being used:

div, ol, p, ul, .class1, .class2 {whatever}

So far as I can tell, it will not remove any of the above so long as one of them is in the book. It would be great if it could remove the dead-wood items from a list.

Best regards, Pop

That is one reason why I don't like listed classes. never have and I don't create listed classes. I have taken some listed styles and separated them so the ones not used could be dumped.

Some publisher CSS is not simple. It can be a bit of a mess. Bit it's not setup to need a specific order most of the time.

[rjwse@aol.com]

Originally Posted by Brett Merkey

I think I may have already posted this, but you can unstack those selectors in the CSS with a regular expression:

--unstack selectors so Calibre "Remove unused rules" tool can be used:
Search: ([.]?[^ ]+),\s([^{]+)(\{[^}]+\})
Replace: \1 \3\n\n\2\3

'Replace All' until 0 returns.

That is slick! I will copy and save to calibre's snipets. After the dead-wood is deleted, can you again regroup the ones which are identical? Best regards, Pop

[Brett Merkey]

That is slick!

It is, kind of. When I set out to make that expression a couple of years ago, I figured it was impossible (for me) but I wanted to give it a try. I was surprised at how quickly it came together.

can you again regroup the ones which are identical?

I never bothered to work the other end because usually most of the stacked rules get swept to the great Bitbucket in the Sky. The only likely path to developing such a regex that I can think of is to have a regex that works on marked text only, within the CSS file. That is, a person wishing to "re-stack" identical rules would have to manually select the identical rules first. This really would not save much labor over doing the restacking manually, so I didn't bother to go into it.