[Gdal-dev] GTOPO30 problem

[Frank Warmerdam]

Clay, Bruce wrote:
> Frank:
>   Thanks for the update.  After nearly a week of debugging and pulling
> my hair out (what little is left that is), I think I found the problem.
> It appears to be in the least likely of all places, Winzip.  All of the
> GTOPO30 files are tarred and gzipped.  I have been using Winzip to
> unpack the files and that appears to be the mistake.  MicroDem has the
> ability to uncompress and untar files.  When I used MicroDem to unpack
> the files I get the correct min and max values as defined in the GTOPO30
> document (at least for the file in question).  As a sanity check I went
> back to Winzip and checked the files out of the same compressed data
> file the run the latest gdalinfo I get
>    "Computed Min/Max=-32768.000,32767.000"
> 
> I can't explain why this is the case just that it certainly appears to
> be.  I would be real happy to have independent confirmation from anyone
> that has the time to run the same test "gdalinfo -mm e020n90.dem" on a
> dem file extracted with Winzip version 8.1 
> 
> Another interesting tidbit I just tried is a freeware program
> (PowerArchiver 2001, v6.1) unpacks the data correctly such that the
> results of the above test are "Computed Min/Max=-137.000,5483.000"

Bruce,

I hope I never meet a WinZip developer.  They provide a "convenient default
option" to apply DOS/UNIX CR/LF conversion when unpacking files.  Unfortunately
this corrupts binary files.  I have had no end of hassles due to this brain
dead decision on their part.

We used to run into a similar problem with people downloading files by ftp
and the ftp client applying default CRLF conversion.

Damn, and double damn winzip.

Best regards,
-- 
---------------------------------------+--------------------------------------
I set the clouds in motion - turn up   | Frank Warmerdam, warmerdam at pobox.com
light and sound - activate the windows | http://pobox.com/~warmerdam
and watch the world go round - Rush    | Geospatial Programmer for Rent