rblampain, you have discovered that different engines (browser, calibre, doesn't matter) render HTML differently.
This is a problem almost as old as HTML itself.
It seems to me that your statement (thread title) is really a hidden question.
It seems you hope that by repeating this question the engines will somehow stop behaving that way.
They won't.
Congratulations on your discovery; but there's no ultimate solution to this.
You can research and hopefully minimize the problem, that's all.

2 members found this post helpful.

Thanks for the clarification, it solves the problem.

My apologies.
Although the sample code gives the same result in Calibre, it does not work in the browser either (Firefox) while the HTML page it is extracted from does work (in the browser) and so far, I cannot see why.

I will follow ondoho's advice.

Somehow, I have raised the curiosity of the creator of Calibre himself and if I receive any advice from him, I will pass it here.

Consider that Calibre's source code includes this file...

https://github.com/kovidgoyal/calibr...rver/reset.css

1 members found this post helpful.

From the very clear answer from the creator of calibre:

"..most readers use the browser for the actual rendering.. "
"..calibre work in "paged" mode by default.."
which displays the page containing the targeted link - anywhere in the page.


The solution was to insert a page break before and after so only the targeted link is in that page and displayed after clicking the link.

Added css code in the external CSS file for the targeted links:
It may be that, except first converting files in pages, there may not be that much differences between browsers and Ebook readers although some tags may behave differently in one than in the other.