Showing posts with label python. Show all posts
Showing posts with label python. Show all posts

Wednesday, November 11, 2009

META: More on the Image Break

I have figured out a way to fix all the links in one fell swoop. However, it is fairly drastic mojo. It will likely remove any flags (funny, cool, interesting, etc) that have been set on posts. Trial runs have indicated that it probably will not erase comments, but it might result in some strange feed behavior. I am not doing it immediately, but will probably do it in the next few days.

For those of you interested in what this entails....

A quick Google search didn't turn up many promising avenues. An Ask Metafilter thread from 2005 said it might require hacking the Blogger API to implement changes of this magnitude.

It turns out that it doesn't require going quite that far. It is possible to export all of the posts of a Blogger blog into an XML file, and then import it later either to the same blog or a different one. While it is on the local machine passing it through a quickie Python script easily fixes the links, once I get the old URL format solidly recognized (for the record, Comics.com's new URL system is rather simpler than the old one). It turns out that even comments get exported to the archive file.

The problems arise from the fact that, while I can restore the blog posts to a new blog then change its address to match the old one, not all of the old blog's settings get restored. The kinds of issues this produces ranges from minor (having to reupload the banner) to somewhat harsh (anyone following the blog will have to refollow).

The alternative is to delete all of the posts on the current blog and reimport them from the hacked backup. This should be safe since I have the blog backed up. It will keep all of the settings and followers, but I don't know if it will do something nutty like resend all of the pages as new RSS entries.

Will probably take action on this in a day or two.