encoding issue when moving to a new host, FYI

Hi All,
First let me say that Brad’s plugin is one of the best ever, I’ve used it on a bunch of sites without a hitch. This time, I finally ran into an issue, and it may not even be the plugin’s fault, but here it is as a cautionary tale.

I was asked to move a huge existing site to Dreamhost. I got my backup from the plugin. Over on the other side, the backup was too big to load through the browser. I came back and did the backup again, this time compressing it, and then the import went fine. [By the way, Dreamhost’s admin panels are impressively annoying and confusing, and it’s a big pain to view a site there before propagation!]

The “new” site ran fine, but I had the dreaded “question mark/diamond” symbols on Posts. I recognized this as an encoding issue. I was going to just leave it, as it wasn’t ubiquitous, but then I found that many pages/posts would not display in the WP editor, a bigger deal! I looked at the database itself, and noted that it had a different encoding then the first one, oops!! (that could have been the whole problem, and it’s something worth checking, although non-geekie humans probably wouldn’t be able to do that. this was also the first time I compressed the data, but I still doubt that that hurt anything) I did some reading and investigating, but couldn’t find any clear way to fix this en masse.

I ended up deleting all Posts/Pages and then doing a normal WordPress XML export/import (Tools… export, and so on), and then the whole thing was fine, and no funny characters. 🙂

This is an awesome plugin, and I’m thinking there may be no way for it to “know” that you’re importing one type of character-encoded data into a completely different type of database encoding.

Cheers, Dave

http://wordpress.org/plugins/wp-migrate-db/