Re: [Python-markdown-discuss] Passing Extension objects to markdown() and markdownFromFile() does not work

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

Thomas, thanks for the feedback. It is much appreciated.

On Sun, Aug 29, 2010 at 7:14 AM, Thomas ten Cate <tte...@gm...> wrote:
> Since I can't seem to find an issue tracker, I'm posting this to the
> mailing list. Please note that I'm not subscribed.

Out issue tracker is here:

http://www.freewisdom.org/projects/python-markdown/Tickets

It should probably be linked from the front page. Yuri, seeing that
page is locked to avoid spam, could you add the link there.  People
seem to keep having the problem of not finding it.

> Using Markdown 2.0.3. At the bottom of
> http://www.freewisdom.org/projects/python-markdown/Writing_Extensions
> it says:
>
>> However, Markdown will also accept an already existing instance of an extension. For example:
>>
>>     import markdown
>>     import myextension
>>     configs = {...}
>>     myext = myextension.MyExtension(configs=configs)
>>     md = markdown.Markdown(extensions=[myext])
>
> This is partly true, and the code works. However, one would expect the
> same thing to work when using the shortcut functions markdown() and
> markdownFromFile(). But the former is defined as follows:
>
>    def markdown(text,
>                 extensions = [],
>                 safe_mode = False,
>                 output_format = DEFAULT_OUTPUT_FORMAT):
>        md = Markdown(extensions=load_extensions(extensions),
>                      safe_mode=safe_mode,
>                      output_format=output_format)
>        md.convertFile(input, output, encoding)
>
> The load_extensions function expects a list of `basestring`s, and
> chokes on Extension objects. The extra call seems redundant.

Funny you should mention this. Just the other day while working on
something else, I noticed that code int he wrapper function and
wondered about that very thing. I just didn't have the time to
investigate it. Thanks for confirming this for me.
>
> Is there a reason for this implementation, or is it better to remove
> the call to load_extensions and pass the list of extensions directly,
> like this?
>
>        md = Markdown(extensions=extensions,
>                      safe_mode=safe_mode,
>                      output_format=output_format)
>
> As an aside, the markdown() and markdownFromFile() functions don't
> accept extension_configs either, which also seems unnecessarily
> limiting.
>

For the record, this is for historical reasons. Extension configs can
be passed in as part of the extension name, In fact, the command line
script still does it this way. i.e.: ``extensions =
['extname(var=foo)']``. I suspect that is why the wrapper classes do
there own thing with the extensions. At one time, the class didn't
know about that syntax, only the wrapper functions did.

Anyway, I agree that the wrappers should accept anything that the
class will.  I've been intending to do some work on this. Thanks for
the reminder.

-- 
----
\X/ /-\ `/ |_ /-\ |\|
Waylan Limberg