Fossil SCM 1.23 released

The problem is that if links are turned on, bots will kill server by going through each checkin, downloading each and every file and diff. Imagine reconstructing each checkin from SQLite repository, which has 10348 checkins with 1233 files each. (http://www.sqlite.org/src/stat).

Before this, the only way to prevent this from happening is by only showing history and files to logged in users (thus, "anonymous login" feature) -- this was the default.

A few versions back, the links were turned on for some users where Fossil could detect that it wasn't a spider (basically, by looking at user-agent header). It seems like this wasn't enough, thus, more agressive detection using JavaScript.

You can configure everything, of course: the setting is "auto-hyperlink", and you can enable links for everyone by giving the "h" permission to user "nobody".