bug-hurd
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Defualt socket server overriding


From: Neal H. Walfield
Subject: Re: Defualt socket server overriding
Date: Tue, 19 Jun 2007 21:27:05 +0200
User-agent: Wanderlust/2.14.0 (Africa) SEMI/1.14.6 (Maruoka) FLIM/1.14.6 (Marutamachi) APEL/10.6 Emacs/21.4 (i386-pc-linux-gnu) MULE/5.0 (SAKAKI)

At Tue, 19 Jun 2007 23:20:27 +0800,
Wei Shen wrote:
> So I decide to narrow the range of the task to
> a small set of specific servers - socket servers - first as some of you
> suggest.

This is good.

> I list a straightforward approach below.
> 
> (1) Add a set of new environment variables, e.g. PFINETSERVER for the
> pf_inet server and PFLOCALSERVER for the pf_local server.

We should have consistent naming between node names and environment
variables.  The default node names are /servers/socket/{2,pfinet},
etc.  Perhaps have SERVERS_SOCKET_PFINET be the name of the
environment variable to use.

> (2) Hack *_hurd_socket_server* in Glibc to check for these environment
> variables.

Right.

> If one of them is set (to a list of overriding server names), try
> the server names specified by its value prior to the default socket server
> name, in the order of their appearance in the overriding list until
> successfully getting a server port.

A nice feature would be the option to use a file descriptor rather
than a symbolic name.  This could take the form fd:X where X is some
number.  This allows for selected access delegation (consider:
SERVERS_SOCKET_PFINET=fd:3 prog 3</path/to/my/pfinet) and removes a
naming issue.  See section 3.4 of the critique [1] for an example of
the problems that can arise through the use of symbolic names.

Also, I think that if the environment variable is set, then it should
not fall back to the default.  This can hide errors.  If the user also
wants the program to consider the default, then it should list it
explicitly.

Neal


[1] 
http://plato.walfield.org/papers/200707-walfield-critique-of-the-GNU-Hurd.pdf




reply via email to

[Prev in Thread] Current Thread [Next in Thread]