Recompiling under CUDA 3.0 new rule used and compiler gives error

I’m trying to recompile app from CUDA 2.3 to CUDA 3.0
New rule file was used for *.cu files, but NVCC compiler gives same error as with old rule:

1>Compiling with CUDA Build Rule…
1>“C:\CUDA\bin\nvcc.exe” -ccbin “P:\bin\VS9\VC\bin” -I"C:\CUDA\include" -I"./" -I"…/…/common/inc" -I"…/…/…/shared/inc" -I"…/" -I"…/…/db" -I"“p:\bin\NVIDIA GPU Computing SDK/c/common/inc”" -Xcompiler "/EHsc /W3 /nologo /O2 /Zi /MT " -maxrregcount=32 -gencode=arch=compute_10,code=“sm_10,compute_10” -gencode=arch=compute_20,code=“sm_20,compute_20” --compile -o “…\bin/MB_CUDA/Win32/VLAR_kill_enabled/Intermediate\” “d:\R\SETI6\CUDA_MB\client\cuda\”
1>nvcc fatal : A single input file is required for a non-link phase when an outputfile is specified

Where it finds another input file to complain and how to avoid this error?

EDIT: found answer already.
double “” “” after environment variable substitution (bolded part)