ns snapshot installation, simple-wireless.tcl and tcp-friendly unicast congestion control

[Robin POSS <rgp5@cornell.edu>]

Hello, 
I work on a red hat Linux PC.
I installed ns-allinone-2.1b5 , and then the daily snapshot of the 12/5/99
in the ns-allinone-2.1b5/ns-2.1b6-current directory, and the daily snapshot
of the same date of tclcl in the ns-allinone-2.1b5/tclcl-1.0b8 directory.

** Is that the right way to do it? 

I ask because I had to change the virtual link in
/root/ns1/ns-allinone-2.1b5/bin by myself (I changed ns ->
/root/ns1/ns-allinone-2.1b5/ns-2.1b5/ns to ns ->
/root/ns1/ns-allinone-2.1b5/ns-2.1b6-current/ns (and by the way in my case
it solved the problem with simple-wireless.tcl Max Larsson and Raja had the
8th of november)), so I wonder if there are other links I should also change.

** ./validate, on my brand new compiled snapshot of ns, went smoothly exept
at the end:

***********************************************
../../ns test-suite-friendly.tcl impulse QUIET
Test output differs from reference output
Diagnose with: diff test-output-friendly/impulse.test
test-output-friendly/impulse
Or see URL "http://www-mash.cs.berkeley.edu/ns/ns-problems.html".
../../ns test-suite-friendly.tcl impulseMultReport QUIET
Test output differs from reference output
Diagnose with: diff test-output-friendly/impulseMultReport.test
test-output-friendly/impulseMultReport
Or see URL "http://www-mash.cs.berkeley.edu/ns/ns-problems.html".
[...]
Some test failed.
validate overall report: some tests failed:
     ./test-all-friendly
to re-run a specific test, cd tcl/test; ../../ns test-all-TEST-NAME
************************************************

It is strange, there is no reference to "test-suite-friendly.tcl" in the
Linux run (http://www.isi.edu/~yuri/ns-validate/ns_validate_19991205) of
validate for that date. The diagnostic hint doesn't help much. 
Here is what is said in the change log about this "friendly" thing:

************************
[floyd] Tue Jul 6 21:19:27 PDT 1999
     Updated files connected with tfrm.cc and tfrm-sink.cc for
tcp-friendly unicast congestion control. The original code is from Mark
     Handley, with structural modifications by Jitendra Padhye. The
validation test is "test-all-friendly". 
************************

I am not too worried since I don't want to use "tcp-friendly unicast
congestion control", I'm just pointing out to you that in my case the
validate program found some problems.






  
--
Robin POSS