When you visit
http://www.fossil-scm.org/fossil/timeline
You should automatically see all of the hyperlinks without having to log in
as anonymous. If this is not the case, then please send me a snapshot of
http://www.fossil-scm.org/fossil/test_env via private email. Thanks.
It is important for the Fossil server to distinguish between robots and
humans and to only show the plethora of hyperlinks to humans. If robots
see all of the hyperlinks and start walking them, they will end up
downloading every historical version of every check-in and file and diff
and ZIP archive and tarball and annotated source listing, etc. That
results in an unreasonable load on the server and unreasonable bandwidth
utilization.
The latest Fossil code checks the User-Agent string to try to distinguish
between humans and robots. We want to err on the side of caution and
assume robot if we do not recognize the browser. But we'd also like to
avoid excluding real humans where necessary. If you think your browser has
been excluded unnecessarily, please let me know, as described above.
Yes, I know that the User-Agent string is easily forged, and a hostile bot
could trivially impersonate a human. Finding the right balance between
security and ease-of-use is hard. We'll continue with the current strategy
for the time being and see how things evolve. Note that the automatic
humanness detector can be turned off for individual servers on the
Admin/Access page.
The automatic humanness detector has actually been in Fossil for several
weeks and seems to have been working pretty well. But yesterday, the
Systrix web crawler was misclassified as human and caused a surge in the
server load, so I had to adjusted the discriminator algorithm.
--
D. Richard Hipp
[email protected]
_______________________________________________
fossil-users mailing list
[email protected]
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
