From mboxrd@z Thu Jan  1 00:00:00 1970
Date: Fri, 16 Jun 2006 14:25:57 +0200
From: =?ISO-8859-1?Q?=22Nils_O=2E_Sel=E5sdal=22?= <NOS@Utel.no>
Subject: Re: [9fans] Shared Library Service?
In-reply-to: <000101c6913e$8d321f70$14aaa8c0@utelsystems.local>
To: Fans of the OS Plan 9 from Bell Labs <9fans@cse.psu.edu>
Message-id: <4492A355.2070407@Utel.no>
MIME-version: 1.0
Content-type: text/plain; charset=ISO-8859-1; format=flowed
Content-transfer-encoding: 7BIT
References: <000101c6913e$8d321f70$14aaa8c0@utelsystems.local>
User-Agent: Thunderbird 1.5.0.4 (Windows/20060516)
Topicbox-Message-UUID: 6b9d2d18-ead1-11e9-9d60-3106f5b1d025

LiteStar numnums wrote:
> Hello all,
> Had an odd idea: shared library service...
> Executables placed in,say, /lib/shsvc, would be executed with pipes 
> redirected to a mounted service (for now /n/shsvc).
> The file server at /n/shvc simply has files for each library loaded. An 
> application that wishes to use the service
> can simply open a file handle to /n/shvc/library & simply read/write to 
> the library file using a FastCGI/SCGI like
> protocol. The file service could do some simple checks on the data 
> (making sure it's actually protocol data, &c),
> and handle if a 'library' puked while working. This may not make as much 
> sense as I thought, since I am writing
> this after only two hours sleep...
> Cheers,
> --Stefan

This sounds more like a weird way of writing fileservers than "shared
libraries" - which I'd imagine should actually merge code/data into the
address space of the loading process.