html2pdf line breaks

[xw-mk]

Hey,

Is there a way to control line breaks in html2pdf output? I can see that top level sections will always page break, but this logic doesn't apply to a top level requirement. The particular document i'm working has a lot of text and tables per requirement and it's causing some messy looking page breaks (not anything I would class as a bug).

I think I could achieve what I need with any of the following:

• Change the section level that causes a page break
• Force a page break for each requirement
• Manually insert a page break in text somehow (perhaps similar to how linking works [LINK: ])

Appreciate any thoughts or feedback.

[mettta]

@xw-mk Hey,

In html2pdf itself, page breaks can be controlled, but in StrictDoc the layout is fixed and those options aren’t currently exposed to the user (in the way you suggest above).

That said, we’re happy to see if there’s something reasonable we can do within the current constraints. Could you share an example of what you mean by “messy looking page breaks” (for example, a screenshot)? That would help us understand the issue better.

[xw-mk]

Sure, this is obviously subjective and when I say messy page breaks I'm generally happy with how they are handled by strictdoc/html2pdf . My problem is that if I was working on a word document, rather than having a small section of a requirement (example shown below) at the bottom of a page, I would insert a page break for cleaner reading. The document I am working on is particularly bad because the requirement length and page length are just different enough that most pages will have something like this.

Ideally we would like to use some of our requirements documents as customer facing documentation (in this instance CAN message specifications which requires a lot of tables.) rather than create another document that has to be managed separately. Because of this it's harder to ignore small issues like this. Perhaps this is having our cake and wanting to eat it too.

When you say "those options aren’t currently exposed to the user (in the way you suggest above)." does that there options exposed to the user that can force page breaks?

[xw-mk]

I've been experimenting with some of the jinja templates and I think I've found a way to force line breaks for each requirement by adding 'sdoc-node[show-node-type-name="REQUIREMENT"]' to 'data-page-break-after-selectors'

.venv\Lib\site-packages\strictdoc\export\html\templates\screens\document\pdf\index.jinja

  {%- if view_object.project_config.is_activated_html2pdf() -%}
  <script
  defer
    {# data-debug-mode='false' #}
    data-preloader='true'
    data-preloader-target='[html2pdf-preloader]'
    data-preloader-background='#F2F5F9'
    data-forced-page-break-selectors=''
    data-page-break-after-selectors='.pdf-toc .free_text sdoc-node[show-node-type-name="REQUIREMENT"]'
    data-page-break-before-selectors='H2'
    data-no-break-selectors='.html2pdf4doc-no-break sdoc-section-title sdoc-meta sdoc-node-title .pdf-toc-row'
    data-no-hanging-selectors='.html2pdf4doc-no-hanging sdoc-section-title sdoc-anchor sdoc-meta .admonition-title sdoc-node-title sdoc-node-field-label strong:only-child'
  src="{{ view_object.render_static_url('html2pdf4doc.min.js') }}"></script>
  {%- endif -%}

I will run a fork for know with this change as I appreciate this is not something that everyone will want.

During my testing I've encountered a peculiar bug that causes additional blank pages, but as far as I can tell, only when bold text is present. I've put a minimal sdoc below and attached the pdf export.

document (5).pdf

Details

```

[DOCUMENT]
TITLE: Test document
OPTIONS:
VIEW_STYLE: Plain

[[SECTION]]
TITLE: Section 1

[TEXT]
STATEMENT: >>>
Vivamus consectetur mollis varius. Quisque posuere venenatis nulla, sit amet pulvinar metus vestibulum sed. Sed at libero nec justo leo.
<<<

[[/SECTION]]

[[SECTION]]
TITLE: Section 2

[TEXT]
STATEMENT: >>>
Lorem ipsum dolor sit amet, c
<<<

[[/SECTION]]

</p>
</details> 

[mettta]

Yes, you’ve found exactly the parameters I was referring to.
At the moment they are not user-configurable, and for cases like “one requirement per page” the only viable solution right now is indeed what you did - maintaining a fork.

Making this behavior configurable is on our roadmap.

[mettta]

Regarding the bug related to bold text:
in your example I see a collision of two different mechanisms - thank you for highlighting this, it’s very useful.

One of these mechanisms is related to the strong:only-child selector logic.

There is a data-no-hanging-selectors parameter which prevents an element from being left as the last item on a page (similar to how we don’t want a heading to be separated from its following paragraph).

In RST content, subheadings are sometimes created using bold text instead of <h*> tags.
Because of that, a CSS rule strong:only-child was introduced to prevent a page break right after such “pseudo-headings”.

However, this rule is too aggressive:
if a paragraph contains only a single bold text fragment and the rest of the text is not wrapped, the strong:only-child selector still matches.
This is a known limitation and is currently not solved at the html2pdf level.

As a temporary workaround on the StrictDoc side, ensuring that such paragraphs contain an additional inline node is sufficient to break this behavior.

In your specific example, the paragraph with bold “sed” under "Section 1" is incorrectly classified as no-hanging, which causes "Section 2" to be pulled onto the same page.
The appearance of an empty page at the same time is a combined side effect of this interaction.
Once an extra inline tag is present inside the paragraph, the strong:only-child selector no longer matches and the incorrect no-hanging classification disappears.

[xw-mk]

Ah okay, I think I understand. So if a paragraph had 2 pieces of bold text that would not have the additional page break?