➜
This issue tracker has been migrated to GitHub,
and is currently read-only.
For more information,
see the GitHub FAQs in the Python's Developer Guide.
Created on 2010-01-03 20:13 by stefan.schweizer, last changed 2022-04-11 14:56 by admin. This issue is now closed.
| Messages (5) | |||
|---|---|---|---|
| msg97177 - (view) | Author: Stefan Schweizer (stefan.schweizer) | Date: 2010-01-03 20:13 | |
HTMLParser should only handle entity references that are terminated with a semicolon. I know that the semicolon can be omitted in some cases (http://www.w3.org/TR/html4/charset.html#h-5.3) and that some browsers are more tolerant, but the following example causes some odd output: >>> import HTMLParser >>> class EntityrefParser(HTMLParser.HTMLParser): ... def handle_data(self, data): ... print "handle_data '%s'" % data ... def handle_entityref(self, name): ... print "handle_entityref '%s'" % name ... >>> p = EntityrefParser() >>> p.feed("<p>spam&eggs are delicious</p>") Expected Result: handle_data 'spam&eggs are delicious' Actual Result: handle_data 'spam' handle_entityref 'eggs' handle_data ' are delicious' |
|||
| msg97263 - (view) | Author: Florent Xicluna (flox) *  |
Date: 2010-01-05 15:42 | |
It is a documented behavior. http://bip.cnrs-mrs.fr/bip10/scowl.htm#semi Quoted from issue500073: "If you want to process such a document in a specific way, I recommend to subclass HTMLParser, overriding unknown_entityref." |
|||
| msg97266 - (view) | Author: Stefan Schweizer (stefan.schweizer) | Date: 2010-01-05 17:16 | |
I do not think that the semicolon can be omitted here, because it is not at a line break or immediately before a tag, it is in the middle of a paragraph. Anyway, I guess I have to live with the decision in issue500073. Also I could not find an 'unknown_entityref' method in the HTMLParser module, only the now deprecated parser in htmllib had one. |
|||
| msg97270 - (view) | Author: Florent Xicluna (flox) * |
Date: 2010-01-05 18:02 | |
For the record, this is valid HTML 4.01 Strict: <!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01//EN"> <html> <head> <title>Sample</title> <meta http-equiv="content-type" content="text/html;charset=utf-8"> </head> <body> <p>La clé</p> <p>La clé des champs</p> <p>La clé des champs</p> </body> </html> Tested with http://validator.w3.org/check and Mozilla Firefox 3.5.6 Reference: http://www.is-thought.co.uk/book/sgml-6.htm#General But HTML5 should prohibit such ambiguous syntax: http://dev.w3.org/html5/spec/Overview.html#syntax-ambiguous-ampersand |
|||
| msg97278 - (view) | Author: R. David Murray (r.david.murray) * |
Date: 2010-01-05 20:13 | |
w3m (a text mode browser) does not treat the é without the ; as an entity ref (it puts é literally into the display), while firefox does turn it into an eacute with or without the ;. I'm sure somebody somewhere has a table listing which browsers have what behavior. Firefox does render, eg, &test without a trailing semi as &test. If you want to mirror that result in code using HTMLParser, you can implement the behavior in your entityref handler. However, this brings up an interesting issue. Firefox also renders "&test;" literally. You can't implement that full behavior using HTMLParser, as far as I can see, since you loose the information as to whether the entity ref was terminated by a semicolon or not. So there may be a legitimate feature request with respect to that issue. |
|||
| History | |||
|---|---|---|---|
| Date | User | Action | Args |
| 2022-04-11 14:56:56 | admin | set | github: 51875 |
| 2010-01-05 20:13:30 | r.david.murray | set | nosy:
+ r.david.murray messages: + msg97278 |
| 2010-01-05 18:02:42 | flox | set | messages: + msg97270 |
| 2010-01-05 17:16:15 | stefan.schweizer | set | messages: + msg97266 |
| 2010-01-05 15:42:07 | flox | set | status: open -> closed nosy: + flox messages: + msg97263 resolution: duplicate stage: resolved |
| 2010-01-03 21:51:59 | ezio.melotti | set | nosy:
+ ezio.melotti |
| 2010-01-03 20:13:28 | stefan.schweizer | create | |