docbook / xslTNG

DocBook xslTNG Stylesheets
https://xsltng.docbook.org
MIT License
42 stars 20 forks source link

Losing text after footnote #360

Closed tgraham-antenna closed 1 year ago

tgraham-antenna commented 1 year ago

With the following paragraph, the text beginning "Graphics are..." does not appear in the HTML.

    <para>Papers
            typically start on an odd-numbered, right-hand page: this could just follow from
            right-hand pages being the ‘front’ of a leaf, or it could be a by-product of
            single-article reprints of papers needing to start on a right-hand page.<footnote>
                <para>Magazines and newspapers start articles as a two-page spread when it suits them, but in my limited experience, papers in journals and in conference proceedings do not start with a two-page spread.</para>
            </footnote> Graphics are typically floated to the top
            (and sometimes also to the bottom) of the page. In a two-column paper, graphics can be
            either one column wide or span both columns. Large graphics might instead be grouped at
            the back of the paper, after the references. The combination of a two-column layout plus
            floating figures generally takes fewer pages than if all text, headings, and figures
            span the width of the page.</para>
tgraham-antenna commented 1 year ago

Generates:

    <p>Papers
            typically start on an odd-numbered, right-hand page: this could just follow from
            right-hand pages being the ‘front’ of a leaf, or it could be a by-product of
            single-article reprints of papers needing to start on a right-hand page.<span class="footnote" id="R_article1_para2_footnote1-fnote">Magazines and newspapers start articles as a two-page spread when it suits them, but in my limited experience, papers in journals and in conference proceedings do not start with a two-page spread.</span></p>
ndw commented 1 year ago

Can you provide a sample document that demonstrates this? On its own, that paragraph seems to format just fine as HTML:

Screenshot 2023-05-28 at 08 20 06

and PDF:

Screenshot 2023-05-28 at 08 20 35
tgraham-antenna commented 1 year ago

The document is https://github.com/tgraham-antenna/MUK-proc/blob/css-proc/proceedings-css.xml, which XIncludes https://github.com/tgraham-antenna/MUK-proc/blob/css-proc/papers-2021/graham/paper.xml

If you want to regress and run Ant rather than Gradle, the command that I use with the css-proc branch is:

ant -emacs -f build-xslTNG.xml gui.single

This expects the xslTNG clone, with gradlew zipStage having been run, in a sibling directory to MUK-proc.

ndw commented 1 year ago

Interesting. A quick test from those sources also doesn't reproduce the problem. I can't imagine how it's related to XInclude, and yet...still investigating.

Nope, it's not XInclude.

Sigh. How do I make ant work, I wonder...

ndw commented 1 year ago

I've got the css-proc branch, but there seem to be files missing:

ant -emacs -f build-xslTNG.xml gui.single
Buildfile: /Volumes/Projects/docbook/MUK-proc/build-xslTNG.xml

html.prep:

single.basename:

html.single.uptodate:

html.single:

css.prep:

BUILD FAILED
/Volumes/Projects/docbook/MUK-proc/build-xslTNG.xml:151: /Volumes/Projects/docbook/MUK-proc/xsl/img does not exist.

Total time: 0 seconds

If I just create xsl/img to make that error go away, there's an error attempting to copy the dejavu fonts.

ndw commented 1 year ago

Right. This is (another) bug in unwrap-paragraphs. I have to go out for a bit, but I'll try to fix this today.

tgraham-antenna commented 1 year ago

There would have been a --recurse-submodules to clone xsl that I had forgotten about because I always started from an existing MUK-proc clone and just added/pulled the css-proc branch. The instructions in README.md for working with the muk2019 branch should apply to the css-proc branch.

tgraham-antenna commented 1 year ago

Right. This is (another) bug in unwrap-paragraphs. I have to go out for a bit, but I'll try to fix this today.

I have enough to go on with, and you have done more than enough.

I need to shift the abstract margins by 1pt, and then I really need to stop fiddling with the styles and actually write my paper and slides. I don't need to see the missing text nearly as much as I need to see the missing paper, so it's fine by me if you don't get to it today.