| Re: Windows/Linux reading binary data - sign problem [message #57872] |
Wed, 09 January 2008 02:06  |
Nigel Wade
Messages: 286
Registered: March 1998
|
Senior Member |
|
|
RussellGrew wrote:
> Hello,
>
> Interesting scenario here. I have some code (not written by myself)
> that reads a bunch of data series from a binary file.
>
> Most of the data series contain positive floating point numbers. One
> of them contains negative numbers. On windows, this works fine. On
> linux, whenever the negative numbers should appear, the values have
> reverted to -2147.48. The series with only positive numbers reads fine
> on both machines.
That looks very suspicious. It's 32bit -MAXINT, with a decimal scaling factor.
>
> There is some manipulation to produce the data series.
Is the number above what is read by IDL, or the result of the "manipulation"?
>
> Both machines are little endian, checked with
http://www.dfanning.com/tips/endian_machines.html,
> using IDL6.3 in both cases. Linux is 64bit gentoo, windows is a 64bit
> processor running 32 bit windows.
>
> Ideas? Perhaps there is some obvious difference between platforms that
> I am unaware of?
There shouldn't be. What method are you using to read the floating point binary
data? Are you running a 32bit or 64bit version of IDL on Linux, and are the
floating point numbers 32bit or 64bit (float or double)? READU should work the
same on all platforms provided the data is in the correct machine format and
you ask it to read floats/doubles.
For example, this writes a 32bit and 64bit floating point values to a file, then
reads them back. The platform is 64bit Linux and IDL is 64bit.
IDL Version 6.4 (linux x86_64 m64). (c) 2007, ITT Visual Information Solutions
IDL> a=float(-32.0)
IDL> b=double(-64)
IDL> openw,1,'tmp.tmp'
IDL> writeu,1,a,b
IDL> close,1
IDL> openr,2,'tmp.tmp'
IDL> c=float(1)
IDL> d=double(1)
IDL> readu,2,c,d
IDL> print,c,d
-32.0000 -64.000000
IDL> exit
I can also read it back using 32bit IDL:
IDL Version 6.4 (linux x86 m32). (c) 2007, ITT Visual Information Solutions
IDL> c=float(1)
IDL> d=double(1)
IDL> openr,2,'tmp.tmp'
IDL> readu,2,c,d
IDL> print,c,d
-32.0000 -64.000000
If I transfer the binary file tmp.tmp to a 32bit Windows machine I can still
read it using the same code.
--
Nigel Wade, System Administrator, Space Plasma Physics Group,
University of Leicester, Leicester, LE1 7RH, UK
E-mail : nmw@ion.le.ac.uk
Phone : +44 (0)116 2523548, Fax : +44 (0)116 2523555
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
