Author martin.panter
Recipients Arfrever, berker.peksag, eli.bendersky, ezio.melotti, martin.panter, serhiy.storchaka
Date 2015-06-12.02:54:03
SpamBayes Score -1.0
Marked as misclassified Yes
Message-id <1434077644.53.0.833847240771.issue19176@psf.upfronthosting.co.za>
In-reply-to
Content
Ideally I guess the Python native behaviour is better: only call target.doctype() if available. It might allow you to easily implement doctype() in both the old and new versions of the API, without worrying about the API being called twice, and without experiencing any DeprecationWarning. But I do not have a real-world use case to demonstrate this, and I don’t think this decision should hold up committing inherit-doctype.v2.patch. They are separate bugs.

BTW: Another difference between the implementations is that the C version accepts my <!DOCTYPE blaua> test case, but the Python version ignores it. It only works with a more elaborate case like <!DOCTYPE blaua SYSTEM "dtd">. I’m no expert, but I think my original XML should be allowed.
History
Date User Action Args
2015-06-12 02:54:04martin.pantersetrecipients: + martin.panter, ezio.melotti, Arfrever, eli.bendersky, berker.peksag, serhiy.storchaka
2015-06-12 02:54:04martin.pantersetmessageid: <1434077644.53.0.833847240771.issue19176@psf.upfronthosting.co.za>
2015-06-12 02:54:04martin.panterlinkissue19176 messages
2015-06-12 02:54:03martin.pantercreate