https://bugzilla.novell.com/show_bug.cgi?id=289007#c6
--- Comment #6 from Egbert Eich 2007-07-02 22:01:41 MST ---
I've been looking at this some more: the old transport selection mechanism was
much more involved (and sometimes seemed to lack a general logic):
with hostname == foo and DISPLAY=foo:n.m the original code tried unix sockets
first. With DISPLAY=localhost:n.m it tried tcp socket on localhost.
Furthermore there was a distinction between 'local' and 'unix' transport,
local/:n.m as also a valid DISPLAY string.
'local' ment whatever local transport was the preferred one on the target
platform while 'unix' meant "use UNIX sockets". The choice was left to
libxtrans which had the platform dependent code.
My patch doesn't support 'local' yet. Since xcb doesn't support any other local
transport than UNIX sockets it doesn't really make sense to add code for that.
On the other hand we might want to support 'local' as protocol string to avoid
breaking old stuff.
I'm not sure yet how far the old heuristic should be followed. Certainly
existing scripts can make all sorts of assumptions based on the old heuristic
which are no longer true today.
--
Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.