[Ns-developers] Ns-3.2 Daily Bug Status for Monday, September 8, 2008

[Mathieu Lacage]

On Mon, 2008-09-08 at 14:10 -0700, craigdo at ee.washington.edu wrote:

> Currently the big problem seems to be the failing regression tests.  We get
> regression comparison errors on Linux and OSX, a regression crash on cygwin,
> and MinGW refuses to run regression tests in its vanilla flavor (at least in
> my environment).
> 
> I think we need a HOWTO that describes how to set up a MinGW environment
> "correctly" -- so that one can follow the instructions and just have it
> work, rather than having to figure it all out by debugging cryptic error
> messages :-)

http://www.nsnam.org/wiki/index.php/User_FAQ#How_to_configure_cygwin
http://www.nsnam.org/wiki/index.php/User_FAQ#How_to_configure_mingw

I should point out that our supported windows platform is cygwin, not
mingw (I did not try to find where this was discussed but I know that it
was because at one point, I was very much in favor of mingw and gave up
on it). MingW seems to somewhat work but I don't think we should invest
any ressources in it or bother with writing documentation about it.
i.e., if someone wants to do this kind of support, I am all for it, but
I see little point in spreading our current resources on two win32
platforms.

CygWin is higher priority and, right now, it is what we should expect
(modulo the regression bug you mentioned and which is on my TODO): works
in debug mode, fails in optimized mode, no python bindings and no nsc.
Basically, most of the work we did on ns-3 since ns-3.1 is effectively
linux-only. Nsc should eventually work on cygwin because sam seems to
care about this platform but I doubt python will work on cygwin anytime
soon.

And, in case someone really wonders, yes, our windows port is
functional, although it is effectively a second-class citizen but,
unless someone really wants to invest time and effort in it, this is
quite unlikely to change a lot soon.

regards,
Mathieu