From: Linus Torvalds <torvalds@linux-foundation.org>
Date: Sun, 9 Oct 2011 02:19:16 +0000 (+1200)
Subject: Don't drop precision from floating point GP coordinates
X-Git-Url: http://git.tdb.fi/?a=commitdiff_plain;h=0f9d1757a4590f554bcfc87f79c7a72baa5ad95d;p=ext%2Fsubsurface.git

Don't drop precision from floating point GP coordinates

Using '%f' limits the precision to 6 decimals, which may well be
perfectly ok.  But at least in theory you *could* have higher precision,
and gps units will report it, so don't mindlessly limit us to what %f
shows.

This arbitrarily uses '%.12g' instead.  %g will drop excess zeroes at
the end, so it actually results in the same (or shorter) ascii
representation unless you have the extra precision.

Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
---

diff --git a/save-xml.c b/save-xml.c
index d6774b5..4183b1c 100644
--- a/save-xml.c
+++ b/save-xml.c
@@ -162,7 +162,7 @@ static void show_location(FILE *f, struct dive *dive)
 	 */
 	if (latitude || longitude) {
 		int len = snprintf(buffer, sizeof(buffer)-4,
-			"  <location gps='%f %f'>",
+			"  <location gps='%.12g %.12g'>",
 			latitude, longitude);
 		if (!dive->location) {
 			memcpy(&buffer[len-1], "/>\n", 4);