python-markdown-discuss

[Python-markdown-discuss] Footnote ext and random links

[Waylan L. <wa...@gm...>]

Yuri,

I got the footnote extension working again last night and was looking
at the output. There are a few little oddities I'll fix, but the real
problem (IMO) is the random strings in the generated links from marker
to note and back. As they are randomly generated, they are never the
same. This means that if a user was to link to a specific footnote
(i.e.: ``http://example.com/page.html#fn1-xxxxxxxxx``), the link
wouldn't work for future visits. Well, the correct page would load,
but the browser wouldn't jump to the proper footnote - which could be
a useful feature. I also noted that other implementations do not do
this. They simply use the key (i.e. ``fn1``) which should be unique
anyway. On top of all that, as the randomly generated links are
different each time, it is imposable to test with our current testing
framework (tests exist, but they're turned off) - which is why no one
noticed we broke footnotes when we implemented Treap.

So, my question is: what purpose do the random strings serve? Why
aren't we just using the key for each footnote?

-- 
----
Waylan Limberg
wa...@gm...

Re: [Python-markdown-discuss] Footnote ext and random links

[Yuri T. <qar...@gm...>]

> at the output. There are a few little oddities I'll fix, but the real
> problem (IMO) is the random strings in the generated links from marker
> to note and back. As they are randomly generated, they are never the
> same. This means that if a user was to link to a specific footnote
> (i.e.: ``http://example.com/page.html#fn1-xxxxxxxxx``), the link

I did this to make it possible to display multiple markdown documents
on the same HTML page.  E.g., several entries on the front page of the
blog.  But I agree, this is an ugly solution.  You can get rid of the
random suffix, and meanwhile I will think about a better way to deal
with multiple documents on a page and similar issues.

- yuri

-- 
http://sputnik.freewisdom.org/

Re: [Python-markdown-discuss] Footnote ext and random links

[Waylan L. <wa...@gm...>]

I just pushed my changes to footnotes.py. It's a significant refactor
under the hood. I threw out the old hash to keep the order of
footnotes and used OrderedDict. It worked quite well. I was about to
have the setFootnote method make sure the ID was always unique (by
appending a num), but then it wouldn't match the marker anymore. Not
sure how that would work.

Anyway, the output exactly matches PHP Markdown Extra's output now -
except we include a title attribute in the backlink. This addresses
various little oddities with the output we were getting before. And,
we have working tests now.

On Wed, Oct 29, 2008 at 2:49 PM, Yuri Takhteyev <qar...@gm...> wrote:
>> at the output. There are a few little oddities I'll fix, but the real
>> problem (IMO) is the random strings in the generated links from marker
>> to note and back. As they are randomly generated, they are never the
>> same. This means that if a user was to link to a specific footnote
>> (i.e.: ``http://example.com/page.html#fn1-xxxxxxxxx``), the link
>
> I did this to make it possible to display multiple markdown documents
> on the same HTML page.  E.g., several entries on the front page of the
> blog.  But I agree, this is an ugly solution.  You can get rid of the
> random suffix, and meanwhile I will think about a better way to deal
> with multiple documents on a page and similar issues.
>
> - yuri
>
> --
> http://sputnik.freewisdom.org/
>



-- 
----
Waylan Limberg
wa...@gm...