BlogList macro: Allow the base heading level to be specified

[James Teh]

When using the BlogList macro with the full or float formats, blog posts are rendered with their heading inside an h1 tag. Structurally, this infers a top level heading and associated section. This breaks the structure of a document when the BlogList macro is used to render posts beneath the top level. For example, on the front page of ​http://www.nvda-project.org/, I use heading level 1 for the page title and heading level 2 for the main section headings. "News" is a main section heading (level 2), so the blog post titles should be at heading level 3.

[James Teh]

This patch uses a stream filter to "rebase" the heading levels. I don't think there is a nice way to do this from within the templates. py:replace or py:match might be used to achieve this, but it would be a bit ugly at best. Nevertheless, a better implementation is welcome.

Note that the web site given as an example above (​http://www.nvda-project.org/) has this patch applied. The blog posts appear at heading level 3 instead of heading level 1. Heading level 1 is of course still the default.

[Mitar]

Why was this patch not integrated? I would also like this feature. Even more, I would like to be able to specify base heading level of posts also when listed or displayed standalone (through some configuration option). I would like to do something like for lists:

<h1>Listing description</h1> ... <h2>Post title</h2> ... <h3>First heading level, for example = ... = in post content<h3>

and when displayed standalone:

<h1>Post title</h1> ... <h2>First heading level, for example = ... = in post content<h2>

This would make blog pages semantically valid XML/HTML.

I have tried to do this with site.html template (using #6023):

<py:match path="div[@class='blog-main-list']//h6"><h6>${select('*|text()')}</h6></py:match>
<py:match path="div[@class='blog-main-list']//h5"><h6>${select('*|text()')}</h6></py:match>
<py:match path="div[@class='blog-main-list']//h4"><h5>${select('*|text()')}</h5></py:match>
<py:match path="div[@class='blog-main-list']//h3"><h4>${select('*|text()')}</h4></py:match>
<py:match path="div[@class='blog-main-list']//h2"><h3>${select('*|text()')}</h3></py:match>
<py:match path="div[@class='blog-main-list']//h1"><h2>${select('*|text()')}</h2></py:match>
<py:match path="div[@class='blog-main-list']//div[@class='blog-body']//h6"><h6>${select('*|text()')}</h6></py:match>
<py:match path="div[@class='blog-main-list']//div[@class='blog-body']//h5"><h6>${select('*|text()')}</h6></py:match>
<py:match path="div[@class='blog-main-list']//div[@class='blog-body']//h4"><h6>${select('*|text()')}</h6></py:match>
<py:match path="div[@class='blog-main-list']//div[@class='blog-body']//h3"><h5>${select('*|text()')}</h5></py:match>
<py:match path="div[@class='blog-main-list']//div[@class='blog-body']//h2"><h4>${select('*|text()')}</h4></py:match>
<py:match path="div[@class='blog-main-list']//div[@class='blog-body']//h1"><h3>${select('*|text()')}</h3></py:match>

but it does not work because of the ​#8770. So the other way I see this to make work is to do it in plugins code. With for example two configuration options which would output from wiki_to_html in post render​ pipe through header rebaser.

[osimons]

I've somehow overlooked this ticket and patch. It is useful, and a nice implementation. Thanks jteh!

I'll look to get it integrated as a macro-only thing, and don't think I want to redo the regular blog listings.

[Mitar]

I doubt it has to be much reworked. Would not it be enough to change line​ to:

${wiki_to_html(context(post.resource), do_shorten and post.body[:blog_max_size] + ' ... ' or post.body) | HeadingRebaser(list_heading_level_from_config if list_mode else alone_heading_level_from_config)}

with HeadingRebaser from above patch?

[James Teh]

Replying to osimons:

I'll look to get it integrated as a macro-only thing

Is there something that needs to be fixed/changed before this can be committed? I'm happy to try to do the required work if any.

[James Teh]

Patch updated for latest code

[anonymous]

Has any changes with this problem?