Xiaofan Chen <xiaofanc <at> gmail.com> writes:
> I am not sure if I buy this or not. I can tolerate Firefox to pop
> up script error under Linux but not crashing or exit disgracefully
> even if the cause is the incompatible website or some not-well-written
> JavaScript.

There is nothing to buy the browsers are all free. A JavaScript equivalent of 
an unintentional fork bomb is as bad as any resource exhaustion D.O.S. failure.
The difference can mainly be seen between a system reboot on other platforms and
a simple application shutdown and restart on *nix derivatives (with or without
logout). This has got to do with the amount of 'integration' evangelized by the
respective platform developers, as does the amount of unsaved work destroyed by
such an event. Pick the one that you claim to hurt you least ... in my case
typical browsing means 2-3 browser instances with 15+ tabs open, probably a
third of them datasheets or PDF files with 100+ pages where the exact page is
pretty hard to find again in case of crash. If one goes down it takes all of
them with it. Fortunately Firefox has a 'restore previous session' setting and 
I can be back where I was before disaster struck in 60 seconds or so.

There is no single browser that works everywhere, f.ex. I have good luck 
trying alternatively Firefox, Opera and IE on certain sites. IE is not the one
that 'usually works', by the way. Also with the latest virus craze I am
reluctant to use IE. Keeping a Windows machine clean of c**p is a full time job
I cannot afford.

For really picky instances it is possible to select resource limitation of any
client applications (such as browsers) on *nix. This is done using quotas. The
result is that a browser crash will not log the session and the window manager
out and will not starve the system of free heap or cpu cycles, such that other
more critical applications will keep running.

Peter


-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist