Ne pas transformer tous les caractères non-ascii en leur version html

[ghost]

Je viens de mettre en place sur un blog à http://traductions.sploing.fr/index.html et quand on regarde le code source de la page, tous les caractères non-ascii sont transformés en leur équivalent html, ce qui le rend plus difficilement lisible et réutilisable par Facebook et co, comme par exemple pour le titre de cet article  : http://traductions.sploing.fr/novlangue/2013/04/06/peche-fiscal/ où un nbsp; remplace un bête espace. Est-ce qu'il ne serait pas possible de laisser le texte utf-8 tel quel ? Après tout, c'est ce qui est déclaré en en-tête de la page ici.

[justinmayer]

Have a look here:

https://github.com/getpelican/pelican/blob/master/pelican/themes/notmyidea/templates/article.html#L2

Does that help address the   in the <title> tag behavior you described?

(cc: @bbinet, since he may be better-positioned linguistically to assist)

[ghost]

Oh thanks a lot ! I do speak english but somehow I just used my mothertongue. Probably because the doc also exists in French. I'll give it a try tonight and again thanks for your kind reply !

Justin Mayer notifications@github.com schrieb:

Have a look here:

https://github.com/getpelican/pelican/blob/master/pelican/themes/notmyidea/templates/article.html#L2

Does that help address the   in the <title> tag behavior you
described?

(cc: @bbinet, since he may be better-positioned linguistically to
assist)

---

Reply to this email directly or view it on GitHub:
#831 (comment)

Diese Nachricht wurde von meinem Android-Mobiltelefon mit K-9 Mail gesendet.

[ghost]

Your suggestion did work, you can check it there with the facebook debugger : https://developers.facebook.com/tools/debug Wonderful !

In order for me to validate html5, do you also know how to ask to pelican to escape correctly urls, I mean with %20 instead of spaces. There : https://github.com/sploinga/traductions/blob/master/bootstrap2/templates/article_infos.html for example, if I could not only give to facebook and twitter the article.title, but also the article.title|correct format for the url, it would be very nice!

Btw, thanks a lot for this great piece of software. Simple and powerful at the same time.

[justinmayer]

You could try looking at the Jinja docs; perhaps something in the following section might work: http://jinja.pocoo.org/docs/templates/#builtin-filters

[ghost]

OK great this is the doc I missed. Will first look there next time :) Thanks again !

You could try looking at the Jinja docs; perhaps something in the
following section might work:
http://jinja.pocoo.org/docs/templates/#builtin-filters

---

Reply to this email directly or view it on GitHub:
#831 (comment)

Diese Nachricht wurde von meinem Android-Mobiltelefon mit K-9 Mail gesendet.

[justinmayer]

No problem. Is there anything else we can do to assist regarding this issue?

[ghost]

nope i think i found out. In jinja 1 there was an urlencode function, but they dropped it in jinja 2 apparently to be encoding agnostic. See pallets/jinja@37303a8 and pallets/jinja#17 . What I did was simply to grab the code proposed there and add it locally to my filters.py in my jinja installation. This is not very safe because next time i update it it will be erased, but at least it works.

code is simply : 

def do_urlencode(value)
    if type(value) == 'Markup':
    value = value.unescape()
    value = value.encode('utf8')
    value = quote_plus(value)
    return Markup(value)

quote_plus comes from urllib.

Then, in my template, i striptag before i urlencode, and pof it works correctly according to facebook debugger, the w3c validator, and the html source code.

It might be a good idea to integrate some custom jinja filters such as this one in pelican, or to recommand striptagging title and description for sharing websites ? For this filter you only need to force utf-8 on everyone, don't know if you want that.

Again, thanks a lot for your very helpful and much appreciated remarks ! 

[avaris]

@sploinga, you can add additional Jinja filters with Pelican. Define that function in your settings file and also add the following in settings.

JINJA_FILTERS = {'do_urlencode': do_urlencode}

Now you can use that filter and you won't lose it when you update.

[ghost]

Great, I'll do it soon that way, thanks ! :)

Now that you two did answer everything I could hope for, I think it's safe to assume you can close the issue ;)

[justinmayer]

Glad to hear it. Issue closed.