Silly bug - while we normally fill the cylinders in a tight group with no
"null" cylinders in the middle, since we copy from an existing dive and
since that may end up with an odd sequence of cylinders, we need to
continue looking at ALL cylinders and not stop with the first one that's
empty.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
In analyze_gaslist() we were only keeping gases that had their .depth
above the current depth, i.e. where we could switch to in the future. Now
we take note as well of the strongest gas that we could have already
switched to and switch to it before we attempt to ascend.
[Dirk Hohndel: minor whitespace and code cleanup]
Signed-off-by: Robert C. Helling <helling@atdotde.de>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
This commit is a little bigger than I usually prefer, but it's all
somewhat interconnected.
- we pass around the cylinders throughout the planning process and as we
create the plan we calculate the gas consumption in every segment and
track this both in the end pressure of the cylinder and over time in
the samples
- because of that we no longer try to calculate the gas consumption after
being done planning; we just use what we calculated along the way
- we also no longer add gases during the planning process - all gases
have to come from the list of cylinders passed in (which makes sense
as we should only use those gases that the user added in the UI or
inherited from a the selected dive (when starting to plan with a dive
already selected)
With this patch I think we are close do being able to move all of the
planning logic back into the planner.c code where it belongs. The one
issue that still bothers me is that we are juggling so many dive
structures and then keep copying content around. It seems like we should
be able to reduce that.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
We don't want to do this when calling plan() from createTemporaryPlan() -
we only want to record the dive at the end of createPlan().
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
This patch paints the dive red if the user is breaking ceiling
on the planner - it's quite fast, it analizes the depth over the
max(tissue_1 .. tissue_16) and changes the color of the profile.
Signed-off-by: Tomaz Canabrava <tomaz.canabrava@intel.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Gas consumption calculation fixed. Pressure difference still needs cylinder size to be set.
Signed-off-by: Robert C. Helling <helling@atdotde.de>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
In commit bcdd6192fe ("Show translated event names in tooltip") I was
too aggressive in replacing the checking for event names with checking for
event types. It turns out that we are abusing an existing event type in
the planner (and use a different event name to mark the difference). By
just checking for the type this now caused incorrect information to be
displayed in the info box (a simply "PO2 warning" on a Suunto D9 could
turn into a "Bailing out to OC" notice).
The correct fix is to get our own range of SAMPLE_EVENT_xxx numbers from
libdivecomputer. Once we have those, we can do this the right way. For now
we just fall back to also checking the event name (which is what I wanted
to get away from so translated names don't trip us up).
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
In order for this to work we need to compare against the event type
instead of the event name - which makes much more sense to do, anyway.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Rewrite of the actual planner logic. Now ascend to the next potential stop
depth. There the state is cached and we try to ascend to the next stop
depth. If we hit the ceiling while doing that we go back to the cached
state and wait there for a minute. Then we try again. Then loop.
Converted all depth related variables from unsigned int to int. During
planning, in a time step the current depth can temporarily be negative and
comparisons of a negative int with an unsigned it have not the result I
expected ( (int) -2 < (unsigned int) 3 turns out to be false). And we
don’t really need the 32nd bit that unsigned buys us for depths.
Deco stops are now shown in the same table as manually entered stops in
boldface (I removed the second table to save screen estate).
The gas shown in the table is still misleading as it means the gas used on
the segment leading up to that event.
The update of the profile only works partially upon changes in the list of
available gases.
Treatment of various gases is basically there but needs some more love.
The ascent velocity is now provided by a function that takes the current
depth as argument. Currently it always returns 10m/min but that will later
be variable (and hopefully user configurable).
The profile is not redrawn while deco is computed (avoiding an infinite
recursion).
The table got a new column for the duration of a segment while the old
“duration” column was renamed “Runtime” to reflect what it actually shows.
Currently, only the run time but not the duration are editable.
All deco gases are used from the depth where their pO2 is 1.4bar. This
should become more flexible.
Calculation of the pressure drop in cylinders without configured volumes
is suppressed. This solves a problem with the planner crashing when saving
a dive where not all cylinders had been manually given a volume.
[Short rant break: Treating 0/0 as air bites back at so many places. E.g.
Cylinder data is initialized with memsetting the whole structures to 0.
Then later suddenly this totally unconfigured cylinder is being treated as
it would contain air. Maybe at some point this was a feature. But it lead
to a naughty bug which took me over an hour to resolve. We should
seriously reconsider this choice and better move to 209/0 being air if
changing this everywhere is not too much trouble]
Signed-off-by: Robert C. Helling <helling@atdotde.de>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
We used to fall back to an AL80 default cylinder, but that meant that a
user who doesn't want a default cylinder at all had no way to indicate
that.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
The "report_error()" interface is a lot simpler, although some of the
C++ code uses QStrings which make them a bit annoying, especially for
the varargs model. Still, even with the explicit conversion to UTF8 and
"char *", the report_error() model is much nicer.
This also just makes refreshDisplay() do the error reporting in the UI
automatically, so a number of error paths don't even have to worry. And
the multi-line model of error reporting means that it all automatically
does the right thing, and reports errors for each file rather than just
for the last file that failed to open.
So this removes closer to a hundred lines of cruft, while being a
simpler interface and doing better error reporting.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
This shouldn't happen, but in case we run out of gases we shouldn't use
the negative gas index (which is the error return of get_gas_idx()) for
the array. Let's fall back to the (incorrect) first gas.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
I know everyone will hate it.
Go ahead. Complain. Call me names.
At least now things are consistent and reproducible.
If you want changes, have your complaint come with a patch to
scripts/whitespace.pl so that we can automate it.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Earlier we converted the C++ code to using true/false, and this converts
the C code to using the same style.
We already depended on stdbool.h in subsurfacestartup.[ch], and we build
with -std=gnu99 so nobody could build subsurface without a c99 compiler.
[Dirk Hohndel: small change suggested by Thiago Macieira: don't include
stdbool.h for C++]
Signed-off-by: Anton Lundin <glance@acc.umu.se>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
More to get clang-scan to quiet down that for the unlikely event that
unsigned int and int is different sizes.
Signed-off-by: Anton Lundin <glance@acc.umu.se>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Some code in plan() left from the gtk days introduced a safety stop in the
plan. It created a un-editable diveplanpoint.
Fixes#349
Signed-off-by: Anton Lundin <glance@acc.umu.se>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
The code would have leaved gasidx undefined if it doesn't find a correct
gas, so this asserts instead of using uninitialized variables as array
index.
Signed-off-by: Anton Lundin <glance@acc.umu.se>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
The previous code checked against the current depth to find the next
deco stop, not the ceiling we actually should head for.
Signed-off-by: Anton Lundin <glance@acc.umu.se>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
With this every cylinder downloaded from a divecomputer that doesn't
provide cylinder data, and every cylinder manually added anywhere will
default to the default cylinder that is set in the preferences.
For people who most of the time dive with the same equipment (always on
dive boats with AL80, or almost always diving their personal HP119) this
should be a nice improvement.
If you don't like this behavior, simply leave the default cylinder setting
in the preferences empty.
This commit also fixes the incorrect s->value call (should be
s->setValue). I wonder what this did to the default filename before...
Fixes#145
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
The *-clang* selector doesn't appear to work correctly in my build environment
(or I just don't understand how it is supposed to work). Either way, making
this conditional on !mac works.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
There is a add_gas_switch_event, so don't duplicate the code.
Signed-off-by: Anton Lundin <glance@acc.umu.se>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
This allows to add missing gas change events to the currently shown dive
computer. Only gases defined in the Equipment section are offered.
Fixes: #250
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
- get_gas_from_events does NOT always set o2/he. It only updates them IFF
a matching event is found; so we need to make sure we start out with a
valid gas mix
- the way we tried to restore the edited dive in case of an edit to a
manually added that is cancelled was completely bogus. Way too complex
when we can simply and reliably simply store the dive and then copy it
back
Fixes#270
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Right now hardcoded to AL80. This way in the future we'll have a volume of
gas that's available. And this makes much more sense then a random string
in the description field.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
This partially reverts changes in commit 1b655d5c806b ("Correctly track
gases when manually adding and then editing dives") as it turns out this
did NOT help us correctly track gases (which is ironic, given the title of
that commit). I didn't actually want to revert that commit as
infrastructure has changed since then and this made the patches look even
more incomprehensible.
So we are back to tracking the "gas on which we arrive at this spot" in
each dive plan node as this makes the rest of our planning so much easier
- I had forgotten about the reasons why we did things this way when I made
the above mentioned commit.
Instead we now make sure that our available tanks are added the correct
way, that such entries are ignored when planning and when drawing the
editable profile, and that at the end it all gets assembled correctly.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
This is a corner case in the planner that was exposed by the recent
changes to the way the dive plan reflects the gases during the dive.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
The code had quite a few odd special cases that may have been left-overs
from the old Gtk algorithm. With this the gas is actually in the dive plan
node where it's use starts. And we maintain the gas correctly between
multiple edit sessions.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Since the model name is written into the XML file it has to be a literal
string that isn't translated. Otherwise a datafile written in one locale
behaves differently when opened by Subsurface under a different locale.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
This gets rid of compiler warnings "format not a string literal and no
format arguments [-Wformat-security]". E.g. when building distribution
packages these warnings are often treated as errors preventing the
build (with good reason).
Signed-off-by: Miika Turkia <miika.turkia@gmail.com>
This may seem like a really odd change - but with this change the Qt tools
can correctly parse the C files (and qt-gui.cpp) and get the context for
the translatable strings right.
It's not super-pretty (I'll admit that _("string literal") is much easier
on the eye than translate("gettextFromC", "string literal") ) but I think
this will be the price of success.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
I always worry if these are worth following up on - but these seem pretty
clear and obvious to me. As far as the planner is concerned, depth is
unsigned.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
