| Endian-ness [message #29236] |
Fri, 08 February 2002 12:47  |
Jonathan Joseph
Messages: 69
Registered: September 1998
|
Member |
|
|
Is there a system variable that gives the endian-ness of the current
hardware? I am reading a file which tells me the endian-ness of the
data, and I'd like to swap_endian if it is different from the current
hardware. In lieu of finding a system variable to compare to, I have
done this:
test_int = 1
byteorder,test_int,/ntohs
big_endian = test_int eq 1
"network" byte order is big-endian, so I convert a 1 to the host
byte-order and see if it's still a 1.
This way works fine, but it seems as thought I'm missing something.
Is there a better way? Using the byteorder routine to convert the data
is not an option (unless it's been improved for 5.5).
Thanks.
-Jonathan
|
|
|
|
| Re: Endian-ness [message #29302 is a reply to message #29236] |
Mon, 11 February 2002 10:01  |
Robert Stockwell
Messages: 74
Registered: October 2001
|
Member |
|
|
Jonathan Joseph wrote:
> Thank you David and Liam,
>
> Liam's result seems more aesthetically pleasing (no offense David).
> Unfortunately, I don't think I can use the swap_if_big_endian and
> swap_if_little_endian keywords to OPEN, because whether I swap or
> not depends on BOTH the hardware and the file. I have no a-priori
> knowledge of the endian-ness of the file until I've already
> opened it. Although, I guess I could open the file, figure out
> what it is, close it and then re-open it.
This is exactly what I do. I have a rule that all data I write
shall be little endian, but there still can be files that are
big endian. I have a small range of possible values, so I read the
data in little endian, check the range, and if it exceeds the
range, reread the data as bigendian.
That seems to be the only way to tell if a file was written
as big/little endian. Of course, one could put a header
on the file indicating how it was written, but if someone else
writes the data, you are out of luck.
Cheers,
bob
|
|
|
|
| Re: Endian-ness [message #29305 is a reply to message #29236] |
Mon, 11 February 2002 08:12 |
Liam E. Gumley
Messages: 378
Registered: January 2000
|
Senior Member |
|
|
Jonathan Joseph wrote:
[stuff deleted]
> Am I correct in assuming that Liam's functions don't need to be
> specifically cast to Long? In other words,
>
> byte(1,0) would yield the same result as byte(1L, 0L)
That's just my FORTRAN background showing: I prefer to use explicit
casts.
Cheers,
Liam.
Practical IDL Programming
http://www.gumley.com/
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
