Re: changes in testing sub-system: umlplutotest
Richard Guy Briggs <[email protected]> Wed, 26 Feb 2003 11:36:50 -0500
| Newsgroups | gmane.network.freeswan.devel |
|---|---|
| Message-ID | <[email protected]> |
On Thu, Feb 20, 2003 at 02:05:52PM -0500, Michael Richardson wrote:
> In order to write a new test case for co-terminal tunnels, I needed
> to have three UMLs under "management" (meaning, I can control the
> setup and configuration of them via commands sent to the console)
>
> The current test "umlplutotest" can manage at most two such systems.
>
> I started by generalizing the "2host-test.tcl" to 3, and decided to
> do 'X' systems instead. I then rewrote the umlplutotest to actually use
> the X-capable test, "Xhost-test.tcl" (Maybe I should called it "Nhost-test")
I would have used Nhost-test...
> I realize that this can be confusing. The Xhost test now has a list. A
> test may define:
> XHOST_LIST="EAST WEST FOO BAR"
>
> to create slots by those names. It could equally well be:
> XHOST_LIST="HOST1 HOST2 HOST3 HOST4"
>
>
> For each host defined, there are five variables that must be set for
> Xhost-test.tcl.
>
> ${host}_START - program to start UML. Usually set
> for you by functions.sh
>
> ${host}_{INIT,RUN,FINAL}_SCRIPT - script to run during init,
> during runtime, and to shutdown the UML.
>
> REF_${host}_CONSOLE_OUTPUT - copy of console output to be
> compared against.
Ok, so all of this should be backwards compatible with existing tests,
right?
> This code is in HEAD, and snapshots, but is not expected to release
> with 2.00.
Excellent! Thanks for this generalisation.
> ] Michael Richardson, Sandelman Software Works, Ottawa, ON |net architect[
slainte mhath, RGB
--
Richard Guy Briggs -- ~\ Auto-Free Ottawa! Canada
<www.TriColour.net> -- \@ @ <www.flora.org/afo/>
No Internet Wiretapping! -- _\\/\%___\\/\% Vote! -- <Green.ca>
<www.FreeSWAN.org>_______GTVS6#790__(*)_______(*)(*)_______<www.Marillion.com>