Joey Hess commented on the image linking issue.

I don't care much myself (yet, maybe). But others do care. Especially when it makes a bigger amount of traffic (you might have to pay for) just for some Warez-kiddies linking the movie-poster in their warez-forum. You maybe do not want them to link you at all.

Ich werde jetzt auch gelegentlich in Deutsch bloggen. Dazu habe ich mein Blog ziemlich umorganisiert. Wenn Ihr nur die Englischen Einträge wollt, dann könnt ihr diese über den RSS-Feed http://blog.drinsama.de/erich/en/index.rss bekommen, entsprechend die Deutschen über http://blog.drinsama.de/erich/de/index.rss oder alle über http://blog.drinsama.de/erich/index.rss

I'm going to blog a bit in german, too. For this i restructured my blog a lot. If you want to get only the englisch entries, please use the feed location http://blog.drinsama.de/erich/en/index.rss

You can also get only the german entries using http://blog.drinsama.de/erich/de/index.rss, or all there is using http://blog.drinsama.de/erich/index.rss

Today i migrated from blosxom (perl version) to pyblosxom (python version)

Apart from some layout issues of the wiki formatter (doesn't to itemizations or preformatted, but i can probably just re-edit these entries and use HTML) this went fairly well.

The multiuser modifications i had made to blosxom were easy to make in python, too (well, reading an env-var set by mod-rewrite is easy, and getting the homedir for a user isn't hard either)

The biggest part was to make the entries keep their GUID. I hope i manage to avoid trashing planet.debian.org this way... don't flame me otherwise: i did try not to do so. If you are interested in the code, contact me. The principle is basic: if for an entry coming from "filename" a file named "filename.guid" is found, the first line is read and used as guid for the entry. Then the rss2 renderer checks this guid variable and uses it if present. Otherwise it will fall back to the traditional generation fo guids.

Non-german speakers will want to change my feed location from http://blog.drinsama.de/erich/index.rss to http://blog.drinsama.de/erich/en/index.rss to just get the english entries.

Lars Wirzenius recently [http://liw.iki.fi/liw/log/2004-10.html#20041007b blogged about preventing image linking]. I helped a friend setup this for his [http://www.moviemaze.de/ Movie pages].

There are several tarpits you can run into. Most notably some "smart" web privacy tools such as outpost will not just empty your referrer field, but place some bogus data in there "Field blocked by Output" or so.

Therefore i suggest only blocking links that come from offsite-locations that do start with http://

It is also a nice thing to actually send a warning image then, instead of just dening access. This makes the linking site uglier, and maybe someone will tell the referrer that he has "stolen pictures" on his site.

Because usually the one linking you will NOT get the error message due to caches. He probably found the image on your site, loaded it into his browser cache with a valid referrer, then included it on his page and still can see it just fine.

Also I suggest only applying this to some directory. You'll want a way to include a image in your blog posts and have it being displayed in sites such as planet.debian.org correctly.