Bug 452

Summary:	Cygwin won't build NSC
Product:	nsc	Reporter:	Gavin Weng <gavinweng>
Component:	core	Assignee:	Sam Jansen <sam.jansen>
Status:	RESOLVED WONTFIX
Severity:	blocker	CC:	sam.jansen
Priority:	P2
Version:	unspecified
Hardware:	PC
OS:	Windows

Description Gavin Weng 2008-12-20 01:01:33 UTC

[583/583] build-nsc
python: can't open file 'scons.py': [Errno 2] No such file or directory
Building NSC stack failed

Comment 1 Sam Jansen 2008-12-20 02:08:57 UTC

Gavin, you reported that NSC doesn't work in cygwin in bug 360. Is still doesn't work, and as I reported there, using a VM is the best solution to running it.

Unfortunately there are some very difficult technical problems to solve to make NSC work under Cygwin and I do not know of anyone with the time to work on them (not me for sure.)

You can still build ns-3 without NSC; just do not enable NSC at compile time.

The particular symptom you report looks a bit different. It is probably a good idea for the ns-3 build system to check for NSC+Cygwin and respond with a reasonable error message. That is probably a feature request rather than a bug.

Comment 2 Gavin Weng 2008-12-20 02:59:18 UTC

Actually, that's the reason why I posted it again since it behaves a lot different. It comes out to me that maybe this could be fixed.

I suggest enable-nsc be notified NOT supporting cygwin, just as the module "python bindings" does, which was also reported by me with the resolution WONTFIX.

> Gavin, you reported that NSC doesn't work in cygwin in bug 360. Is still
> doesn't work, and as I reported there, using a VM is the best solution to
> running it.
> Unfortunately there are some very difficult technical problems to solve to make
> NSC work under Cygwin and I do not know of anyone with the time to work on them
> (not me for sure.)
> You can still build ns-3 without NSC; just do not enable NSC at compile time.
> The particular symptom you report looks a bit different. It is probably a good
> idea for the ns-3 build system to check for NSC+Cygwin and respond with a
> reasonable error message. That is probably a feature request rather than a bug.