Bug converting html css text-indent and left-margin

[bhoyt]

Back in version 0.9.25 I converted by NIV Bible. It worked well.

Now I have revised the indentation of poetry and genealogical lists. But using calibre 1.21 the   are removed and negative text-indent used with left-margin is incorrect.

In the attached files (test.html, test.css) the indentation is correct in my browsers but not when I convert to azw3 for my kindle.

Note: line 11: <p class="poet04">&nbsp;&nbsp;<sup class="vn" id="v01010002">2</sup>The sons ...

It converts to: <p class="poet"><sup class="vn" id="v01010002">2</sup>The sons ...

The &nbsp; are removed which I am using for proper indentation. And the class="poet04" which has margin-left:4em; and text-indent:-4em; is changed to class="poet" with margin-left:4em; but text-indent:3% WRONG!!

Note: line 9: <p>&nbsp;&nbsp;<sup class="vn" id="v01010001">1</sup>This is ...

It converts to: <p class="calibre1"><sup class="vn" id="v01010001">1</sup>This is ...

Again the &nbsp; are removed and the <p> tag is replaced by <p class="calibre1"> with a text-indent:3% which is close to correct. OK.

Can the &nbsp; be left and the margin and text-indent settings be left so that my indentations will work correctly?

If not what kind of fix should I use?

Thanks, Kovid.
Your program is great!! Thanks muchly for all the work you have put into it.

[kovidgoyal]

The nbsp are not removed they are converted to the unicode nbsp character, which renders exactly the same.

[bhoyt]

Thanks, Kovid. But I gather that the unicode nbsp characters are then replaced by a text-indent which in the test.zip I have attached is wrong since it is not negative. See line 11 that I mentioned above for this.

I presume this is a bug.

[JimmyR]

When I convert your test.zip to azw3, this is what I end up with:

<p class="poet">**<sup class="vn" id="v01010002">2</sup>The sons<sup class="fnr"><a class="calibre3" href="part0000.html#fn01010002a">1</a></sup> of Japheth:</p>

CSS:
.poet {
    display: block;
    text-align: left;
    text-indent: -4em;
    margin: 0 0 0 4em
    }

The asterisks are the unicode nbsps.

I'm not sure why you would end up with a text-indent of 3%.

[bhoyt]

That's very strange! I just re-added the test.zip to make sure and then converted to azw3 and I get this (note there are no nbsp characters):

<p class="poet"><sup class="vn" id="v01010002">2</sup>The sons<sup class="fnr"><a href="part0000.html#fn01010002a">1</a></sup> of Japheth:</p>

CSS:
.poet {
display: block;
text-align: left;
text-indent: 3%;
margin: 0 0 0 4em
}

Why the difference? I am running Win7 and using Calibre 1.21 64bit

Bruce

[JimmyR]

I'm running Win7 - Calibre 1.21 32bit. What input/output profiles are you using on Page Setup in the conversion dialogue? I don't think that would matter, however. I tried various combinations and ended up with the same results.

It is strange and I'm at a loss to explain it. You might try restoring defaults in the conversion options- though I don't see any settings that would cause your problem.

Edit: I used the default input and Kindle Paperwhite for output.

[bhoyt]

Also strange is that a line down a bit with only one nbsp is rendered correctly:

<p class="poet3">*<sup class="vn" id="v01010013">13</sup>Mizraim was the father of</p>

CSS:
.poet3 {
display: block;
text-align: left;
text-indent: -4em;
margin: 0 0 0 4em
}

[JimmyR]

I got your results when I enabled "Smarten punctuation" in the Look & Feel options. Do you have that enabled? If so, disable it.

[bhoyt]

Ah, yes!!! Thank you very much. That solved the problem.
But I wonder what smarten punctuation is doing with text-indent?

[JimmyR]

Originally Posted by bhoyt

Ah, yes!!! Thank you very much. That solved the problem.

Great! Glad that sorted it.

But I wonder what smarten punctuation is doing with text-indent?

I really don't know enough about how it works to proffer an explanation.

Again, glad it's working for you now.

[susan_cassidy]

text-indent doesn't have anything to do with &nbsp; You shouldn't be using non-breaking spaces for indentation anyway, just the text-indent specification.

[bhoyt]

Yes, I know that I should use text-indent for indent purposes but for this purpose it seems that nbsp is the better option. Why? because some verses have one digit numbers (1-9) and others have two digit numbers (10-99) and a few have three digit numbers (100+). That is a content matter not a CSS formatting matter.

I could do the indentation by requiring all numbers to be three digits (001-100+) but leading zeros doesn't look good.

The text-indent option requires me to make a different class for three different sizes of verse numbers. That seems like forcing content into the CSS formatting.

Any advice?

[JimmyR]

Apparently, multiple nbsps are removed and a 3% indent is added:

Originally Posted by ldolse on 09-13-2010

Some logic just got hooked into the preprocess option to take care of multiple nbsp's being used as indents, converting them to css styles. The code was just checked in, so it will be in the next release.

...

Right now it just takes any number(2 or more) of nbsps and normalizes them to 3% indent. I've been debating changing the indent based on the actual number of nbsps, but a lot of books I've seen use entirely too many anyway.

https://www.mobileread.com/forums/sho...26#post1108526

[jackie_w]

@bhoyt,
For this particular book, you should check the settings you had on the Convert - HeuristicProcessing page. If you have checked both of

- the first box (Enable heuristics) and
- the last box (Replace entity indents with CSS indents),

you might find this is where the problem lies

[bhoyt]

jackie_w, I have heuristics disabled since I make format my own html files directly.


Jimmy-Jim, thanks for that link. It explains why 2 nbsp were converted to 3% indent, whereas 1 was not.