[PATCH] Be much more careful about merging dives

[Dirk Hohndel]

On Tue, Sep 22, 2015 at 12:43:30PM -0700, Linus Torvalds wrote:
> +
> +/*
> + * The end of a dive is actually not trivial, because "duration"
> + * is not the duration until the end, but the time we spend under
> + * water, which can be very different if there are surface events
> + * during the dive.
> + *
> + * So walk the dive computers, looking for the longest actual
> + * time in the samples (and just default to the dive duration if
> + * there are no samples).
> + */
> +static inline int dive_totaltime(const struct dive *dive)
> +{
> +	int time =  dive->duration.seconds;
> +	struct divecomputer *dc;
> +
> +	for_each_dc(dive, dc) {

This gives me a warning when I compile it:

In file included from /home/hohndel/subsurface/dive.c:8:0:
/home/hohndel/subsurface/dive.c: In function ‘dive_totaltime’:
/home/hohndel/subsurface/dive.h:582:11: warning: assignment discards
‘const’ qualifier from pointer target type [-Wdiscarded-qualifiers]
  for (_dc = &_dive->dc; _dc; _dc = _dc->next)
           ^
/home/hohndel/subsurface/dive.c:2853:2: note: in expansion of macro ‘for_each_dc’
  for_each_dc(dive, dc) {
  ^


Declaring dc as

	const struct divecomputer *dc;

fixes that. Any issues with that?

/D