
 Using the Envoy Filesystem  Sept. 3, 1992
 -----------------------------------------

 Contained in this archive are very early revisions of the Envoy filesystem.
 It is very preliminary, and is likely to change in the coming months.

 The EFS is split into two pieces -- a "client" and a "server".  The client
 is actually an AmigaDOS filesystem, and resides in l:.  The server is an
 Envoy Service, and resides in the Services directory of your Envoy
 distribution.

 Any machine can export and import mounts for EFS.  Which mounts are to
 exported or imported, however, must be specified by the user.

 First, some setup is required.  Immediately go to the Config drawer of your
 Envoy installation and select "Services Configuration".  Once there, use the
 configuration editor to Add and Enable the filesystem.service.  Be careful
 to select the full path of the filesystem.service, and not a relative
 path from where the requester "begins" in the directory tree.  This is a
 known bug in the Services configuration editor.

 Next, use the "Filesystem Configuration" editor.  This is where the user is
 to determine what directories are to be exported -- and who is capable of
 accessing them.

 Select 'Add' under the 'Shared Directories' listview.  A requester
 will appear, allowing you to specify what directory or volume
 you wish to make available via the network.  Click on 'Volumes', and select
 either a volume or a directory to export (do not select a file -- only a
 volume or directory).  Once selected, click 'OK'.  The directory or volume
 you selected should now be listed under "Shared Directories".  Click on it
 to verify that it's activated.  Click in the 'Name' string gadget on the
 right side of the screen.  This gadget specifies the Volume name that
 all clients will know your mount as.  Enter an informative string here.
 Also, use the 'Add' button below the 'Users and Groups' listview to add
 the names of whichever users you wish to permit to access the mount that
 you've just created.  The three checkboxes in the upper-right corner of
 the editor specify whether you wish someone using the client to be able
 to Snapshot the disk icon for your mount, whether clients are allowed to
 utilize the 'Leave Out' Workbench feature on this mount, and whether you
 wish to utilize "full" permission protection on this mount ("full" implies
 having ownership in terms of Group and User ID's and protection bits; this
 isn't yet supported).

 Once you've added all of the volumes or directories you wish to make
 available, click 'Save'.  On the next reboot, the exporting machine is
 prepared to act as a fileserver for the directories you selected.

 The client, or importing machine(s) are a bit more of a problem.  A GUI
 program for selecting and creating mountfiles on the Client doesn't
 exist yet -- so these mountfiles must be edited by hand.  A sample
 mountfile has been included in this archive.  The important information
 appears on the UNIT line.  Following is the 'Unit' line from the
 mountfile that I use to access my test Amiga:

 Unit = "siberia:scratchywork:gregmthepw"

 This line has the format of:

 Unit = "(Realm:)machinenameRemoteMountNameClientsUserNamePasswordflags"

    'machinename' simply describes the name of the computer that has a mount
            available that you wish to access.  The Realm: portion is
            optional.

    'RemoteMountName' is the name of the mount that was made available on the
            server.  This must be the same string that was listed in
            the 'Filesystem Configuration' editor.  (Hence, Sys: and System2.0:
            are NOT interchangable - even if they reference the same thing.)

    'ClientsUserName' is the username of whoever is trying to mount something.
            For instance, if your username is "bob", and your're trying to
            reference something on Jack's machine, you should enter "bob"
            here.

    'Password' is currently ignored.  It will, however, be necessary in the
            future.

    'flags' is an optional field.  If you place a 'b' in this field, you
            will not have access to the .backdrop file on that mount -- even
            if it exists.  If you place a 'd' here, you will not have access
            to the Disk.info file -- if it exists.  These switches are a local
            method of preventing Left-Out icons, and Disk snapshotting.  The
            selection set by the server, however, has precedence.

    (Also, the seperating character is a vertical bar -- <Shift><Alt><I>.)

    Eventually, a GUI editor will create these mountfiles for the user,
    requiring little more than point-and-click actions.

    Once you've edited a mountfile, you can either double-click on it to cause
    a mount attempt, or place it in either DosDrivers (in AmigaDOS 2.1 or
    above), or in WBStartup.  If a mount attempt fails, no icon can appear
    on workbench. However, due to a DOS quirk, you cannot re-attempt a mount
    merely by double-clicking the icon again.  You may cause the filesystem
    to re-attempt the mount by making direct reference to the mounted device.
    (The name of the mountfile is the name of the mounted device.  If your
    mountfile is named 'scratchy', the device name of the mount will be
    'scratchy:'.)  A simple "dir scratchy:" will force the filesystem to
    attempt a reconnection.

    You're allowed to import or export as many mounts as your machine's
    memory can handle.










