Anchors in Heading don't work.

[ruoyu0088]

I recently updated to IPython 2.2, and the anchors in heading don't work any more. I debugged this, and found that the the HTML of the rendered heading is as

<h3 id="This-is-a-heading
">This is a heading<a href="#This-is-a-heading
" class="anchor-link">¶</a></h3>

there is a "\n" char at the end of the id and href attribute. I can reproduct this by some simplified code in the HeadingCell.prototype.render() function in textcell.js

$($.parseHTML(IPython.security.sanitize_html(marked.parser(marked.lexer("this is a heading"))))).text()

There is a \n char at the end of the result. To fix this I added an line of code that removes '\n' in hash.

hash = hash.replace(/\n/g, '');

[minrk]

Might have been caused by an update to marked.js. A judiciously placed .trim() might be what we need.

[dsblank]

Thanks for finding this! I was just trying to demo our Table of Contents extension, and it had stopped working.

[dsblank]

As the talk of backporting and releasing a new 2.x release is being made, it would be great to get this fixed and out. I confirm that this is still an issue in master 3.0.0-dev. Looks like ruoyu0088 has a fix, so it shouldn't be too hard.

[Carreau]

Backport are usually only critical bug fix. Not doing so might expose us at not beeing back ported on Debian. So I wouldn't be too optimistic on having this back ported once fixed.

Envoyé de mon iPhone

Le 14 sept. 2014 à 07:02, Doug Blank notifications@github.com a écrit :

As the talk of backporting and releasing a new 2.x release is being made, it would be great to get this fixed and out. I confirm that this is still an issue in master 3.0.0-dev. Looks like ruoyu0088 has a fix, so it shouldn't be too hard.

—
Reply to this email directly or view it on GitHub.

[dsblank]

I contribute to lots of software that is distributed by debian, and I can assure you that a bugfix that:

• fixes a regression (it did work in a previous version, and later versions broke it)
• fixes internal links in a system that is used by a large number of people in a browser
• is a one-line change
• can arguably be declared "critical" by those of us that use things like "table of contents" with large notebooks
• can have a test written for it
• there is no workaround

would not risk your status as a stable project. I appreciate being conservative in changing things, but such a usability bug can make the system look very broken, because it is broken.

[Carreau]

fixes a regression (it did work in a previous version, and later versions broke it)

My

[Carreau]

Rhahh, sorry for previous post, try again:

Any way it's marked as 2.3 milestone, so I supposed it is accepted by the rest of the core team, so it will be done. I guess the limitting factor is someone fixing the bug and sending a PR :-)