|
From: Davide C. <din...@sm...> - 2001-05-03 06:57:49
|
Dear Chris, the first idea that comes to my mind is you could have forgotten the assignment CFLAGS = -DLITTLE in the makefile for compiling c sources, since Intel-like machines have "little endian" byte ordering, as opposed to SGI; you could also verify this with the command od -a <myfile.v5d>|more if the first bytes on the file are something like " D5V" instead of "V5D ", that's the problem. I hope this helps, Davide On Wed, 2 May 2001, christopher moore wrote: > > I'm having a problem. I modified foo_to_v5d.c to read GMT grid files > (really netCDF files) to create a v5d file. > > It runs great on my SGI machine, but when I give it to a friend running > Linux (RH7) it doesn't work in the following way: > > We got it to compile and run (after hunting down a rouge fclose(f)). And > it creates a file, but vis5d and v5dinfo don't recognize it. > > Error: not a v5d file > Error: couldn't open pache3.v5d > > Frustrating, because it compiles and runs without errors. And it > compiles, runs, and gives good data files on my SGI. I'm not up on > Linux... anyone have trouble Creating v5d in Linux? We're using g++ (and > the netCDF c++ libraries, but they seem to be reading the files just fine > - it's writing v5d that's the trouble). Vis5d-5.2. > > Sorry so long. > > Chris > ______________________________________________________________ > Christopher Moore tel 206.526.6779 > Research Scientist, Oceanography fax 206.526.6744 > University of Washington/JISAO/NOAA-PMEL > -------------------------------------------------------------- > > -- __________________________________________________________ Davide Cesari Servizio Meteorologico Regionale __ tel. (39) 051/284541 ||\ fax (39) 051/284664 |||\ e-mail din...@sm... |||/ Address: ARPA-SMR, Viale Silvani 6, 40122 Bologna, Italy __________________________________________________________ |