Comments on: Draconian Error Handling in XML http://lucumr.pocoo.org/cogitations/2008/03/11/draconian-error-handling-in-xml/ Armin Ronacher thinking Fri, 25 Jul 2008 13:00:58 +0000 http://wordpress.org/?v=2.2 By: Armin Ronacher http://lucumr.pocoo.org/cogitations/2008/03/11/draconian-error-handling-in-xml/#comment-2951 Armin Ronacher Fri, 21 Mar 2008 16:05:35 +0000 http://lucumr.pocoo.org/cogitations/2008/03/11/draconian-error-handling-in-xml/#comment-2951 Well. I don't want to be a jerk so I don't blame you for having a page that once had invalid XML on it. Without a proper XML toolchain it's quite hard to get it right. Especially when dealing with data from external websites encodings, and broken markup makes it tricky to still produce parsable XML. Well. I don’t want to be a jerk so I don’t blame you for having a page that once had invalid XML on it. Without a proper XML toolchain it’s quite hard to get it right. Especially when dealing with data from external websites encodings, and broken markup makes it tricky to still produce parsable XML.

]]> By: karl dubost, w3c http://lucumr.pocoo.org/cogitations/2008/03/11/draconian-error-handling-in-xml/#comment-2950 karl dubost, w3c Fri, 21 Mar 2008 15:16:04 +0000 http://lucumr.pocoo.org/cogitations/2008/03/11/draconian-error-handling-in-xml/#comment-2950 ahah good catch. We are using MT which doesn't catch all mistakes. MT publishes the entry on a machine. This machine has a cron running locally and check if there are new pages, if yes, it cvs add and commit them to the W3C server. We have a program on W3C server which every day takes the cvs commit in the cvs log and checks if the files are valid. If not, it sends an email to the committer saying "dude, your code is not valid, fix it." maybe less casual :p The code we use for checking the URI is the LogValidator (you can find it on W3C site.). so not perfect, but we try to catch the errors when they show up. Maybe we could improve a bit by checking before committing, but that would require to hack MT a bit more, which makes it a bit more fragile with regards to future updates. ahah good catch.

We are using MT which doesn’t catch all mistakes.
MT publishes the entry on a machine. This machine has a cron running locally and check if there are new pages, if yes, it cvs add and commit them to the W3C server.
We have a program on W3C server which every day takes the cvs commit in the cvs log and checks if the files are valid. If not, it sends an email to the committer saying “dude, your code is not valid, fix it.” maybe less casual :p

The code we use for checking the URI is the LogValidator (you can find it on W3C site.).

so not perfect, but we try to catch the errors when they show up. Maybe we could improve a bit by checking before committing, but that would require to hack MT a bit more, which makes it a bit more fragile with regards to future updates.

]]>