Hi, Chris Metzler schrieb:
hgtchop didn't freak out for me; it was more of a subtle clue that there were problems that manifested themselves later at the end of the chain. When you run hgtchop on the data and produce the *.arr.gz files, hgtchop prints status info on each tile to the screen. Does that all look OK? Mine looked just fine, *except* when it printed the lat/lon of the tile where it was working, there were non-Latin-1 characters (non alphanumeric symbols) in there, which occasionally caused beeping (ctrl-g, the keyboard bell) when it tried to print them on the screen. Otherwise it looked perfectly fine. But that was the first sign that something was up.
I have read your mails on terragear-devel but until now nothing of that seemed familiar. Until now. I think the non-latin-characters stem from the fact that HGT.write_area - which is called from hgtchop - writes the bucket to cout. SGBucket stores x and y as char. The operator<< declared in simgear/bucket/newbucket.hxx simply prints that as char. Try the attached patch on SimGear.
If anything else doesn't work, I suspect that it doesn't have anything to do with this.
Cheers, Ralf
Index: simgear/bucket/newbucket.hxx =================================================================== RCS file: /var/cvs/SimGear-0.3/source/simgear/bucket/newbucket.hxx,v retrieving revision 1.9 diff -u -r1.9 newbucket.hxx --- simgear/bucket/newbucket.hxx 8 Mar 2006 18:16:08 -0000 1.9 +++ simgear/bucket/newbucket.hxx 13 Apr 2006 13:25:11 -0000 @@ -316,7 +316,7 @@ inline ostream& operator<< ( ostream& out, const SGBucket& b ) { - return out << b.lon << ":" << b.x << ", " << b.lat << ":" << b.y; + return out << b.lon << ":" << (int)b.x << ", " << b.lat << ":" << (int)b.y; }