Use the last 'diveid' when merging dives for one dive computer

When we merge two dives into the same dive because a divecomputer had
incorrectly considered it two separate dives (due to surface time within
the dive) we should pick the dive ID from the later dive to be the
diveid of the resulting merged dive.  Otherwise we might re-download the
(now merged) partial dive.

This is a rather unusual special case, but it actually hit me with the
Uemis on my last dive in Palau: Chandelier Cave has multiple surface
points where you can spend time admiring the cave above water, and the
Uemis (but not my Suunto's) decided that the dive was actually four
short dives back-to-back.

Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>

1 files changed, 3 insertions, 0 deletions

diff --git a/dive.c b/dive.c
index cc673f122..d9943c23f 100644
--- a/dive.c
+++ b/dive.c
@@ -1926,6 +1926,9 @@ static void interleave_dive_computers(struct divecomputer *res,
 		if (match) {
 			merge_events(res, a, match, offset);
 			merge_samples(res, a, match, offset);
+			/* Use the diveid of the later dive! */
+			if (offset > 0)
+				res->diveid = match->diveid;
 		} else {
 			res->sample = a->sample;
 			res->samples = a->samples;