Unable to use Tidy on html

[wichtounet]

Hi,

I enabled the tidy filter on html file:

from nikola import filters
FILTERS = {
".html": [filters.tidy],
}

and now NIkola refuses to build my site:

TaskError - taskid:render_tags:../categories/index.html
PythonAction Error
Traceback (most recent call last):
File "/usr/lib64/python3.3/site-packages/doit/action.py", line 360, in execute
returned_value = self.py_callable(_self.args, *_kwargs)
File "/usr/lib64/python3.3/site-packages/nikola/utils.py", line 529, in unlessLink
action(target)
File "/usr/lib64/python3.3/site-packages/nikola/filters.py", line 144, in tidy
for line in output.split("\n"):
TypeError: Type str doesn't support the buffer AP

Is there a problem with tidy ? Can I use another thing as a replacement ?

[ralsina]

You are using python 3, right? Probably it's just that noone tried this and it's not working right.

[wichtounet]

Yes, I'm using Python 3.3, but the 2.7 version is also installed. If I have installed nikola using pip, is there a way to change the Python version used ?

[ralsina]

Easiest is just to install it again: pip-2.7 install nikola

Or wait 15' and I'm sure I can find you a simple patch ;-)

I was lying, probably won't have time to look at it today.

[Kwpolska]

I’ll attempt a fix soon.

[ralsina]

With all that brokenness in the tool, it's sort of pointless.

[Kwpolska]

So, what is the official resolution?

(a) nuke tidy-in-Nikola altogether; or
(b) pull in #1167 and mark tidy as “not production-ready as it destroys many HTML5 things (don’t forget it was last updated in 2009)”

[da2x]

(c) Tidy-HTML-Light were we strip empty lines and indenting spaces altogether. Force line breaks before and after closing sectioning elements. Better than nothing.

[Kwpolska]

Google doesn’t seem to know what Tidy-HTML-Light is. Same goes for DuckDuckGo. I’m pretty sure it doesn’t exist.

[ralsina]

Well, if the goal is to produce pretty HTML, we could just use lxml and pretty-print it.

[da2x]

Two kind of goals here: smaller output (stripped whitespace) or pretty HTML.

Tidy-HTML-Light is better known as regex.

[wichtounet]

Thanks for the help. I tried using 2.7, but after that, my site refused to build with an UTF-8 error. I will continue 3.3. It was more a test than a real need, so don't worry too much about it for me.

Do I let the issue open ?

[Kwpolska]

@wichtounet you should create a new issue describing the UTF-8 problem. We will close this issue if we believe we fixed it well enough (or that we won’t fix it).

[ralsina]

@Aeyoun

For stripping whitespace, yui-compress or whatever does a much better job. For prettier HTML, lxml does a much better job. So, let's fix this by removing tidy.