The "virial" form of the Z compression factor is of the form
Z = 1.0 + A*p + B*p^2 + C*p^3 + ..
and it's considered the "right" polynomial form to use. It happens to
also make for one constant less per gas (since the 1.0 can be added
later), and can be used to simplify the expression and avoid a few
floating point operations.
However, in order for that kind of expression simplification to make
sense, we need to make sure that we don't calculate the powers of the
pressure multiple times either, and that means we have to inline all the
actual calculations.
Our compiler options still mean that the generated code isn't optimal,
but that's a separate issue. And it is a lot better than it used to be.
Being clever about this does potentially make the code a tiny bit less
legible, but maybe that's not too bad for something that we'd expect to
not ever touch once we get it right.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Lubomir found better compressibility data for the pure gases that we
need for scuba, making the air table superfluous: we get good values
from just regular linear mixing of the Oxygen, Nitrogen and Helium
calculations.
Also, rather than using a quintic polynomial, a cubic one does
sufficiently well, making for smaller code and fewer coefficients.
And judging by the reactions from people on G+ (as well as just looking
at how good the fit is with the air data), this is all the right way to
do this, and this thus removes the Redlich-Kwong equation.
All-credit-goes-to: Lubomir I. Ivanov <neolit123@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
.. and use a linear mix of them for arbitrary gas mixes.
For the special case of air, we continue to use the air-specific
polynomial.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
This goes back to just doing air compressibility, but using the
least-squares quintic polynomial equation that Lubomir generated based
on the Wikipedia table for air at 300K in the 1-500 bar range.
We might be able to do similar things for mixed gases..
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
The gas compressibility is such a specialized thing that I really prefer
having it separate.
This keeps Robert's Redlich-Kwong equation as-is, but let's experiment
with other models soon...
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
The dive list might contain dives in the future, don't add the new dive to
then end but instead add it at the correct spot in the list
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
As it turns out, the van der Waals equation gives results that are
numerically not really useful, so we use the Redlich Kwong equation
which is, according to Wikipedia, much more accurate, which can be confirmed
given the empirical values for air.
As opposed to the previous approach with a look-up table, this takes
into account the actual gasmix. This always assumes the gas to be at
20 degrees Centigrade.
Signed-off-by: Robert C. Helling <helling@atdotde.de>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
"weights" and "cylinders" are QStringList Q_PROPERTIES, and Grantlee
should be able to render them, but it doesn't.
To be able to print the whole list of weights and cylinders we
introduce two new QString properties "weightList" and "cylinderList".
The variable replacement in the previous patch deals with the
conversation of the user side HTML, e.g.:
USER -> INTERNAL
"{{ dive.weights }} -> {{ dive.weightList }}"
"{{ dive.cylinders }} -> {{ dive.cylinderList }}"
Signed-off-by: Lubomir I. Ivanov <neolit123@gmail.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
There was a reported case of an import of a dive that gave a salinity of
35g/l. This is an actual salinity (an amount of salt in the water) but
for subsurface the salinity is actually the density of the water. So for
too small values of the salinity add the density of fresh water.
Signed-off-by: Robert C. Helling <helling@atdotde.de>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
We could in theory make this dependent on the gasmix, but for now let's
just assume (incorrectly) that everything we breathe acts like air.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
We had two totally different usage cases for "get_volume_string()": one
that did the obvious "show this volume as a string", and one that tried
to show a cylinder size.
The function used a magic third argument (the working pressure of the
cylinder) to distinguish between the two cases, but it still got it
wrong.
A metric cylinder doesn't necessarily have a working pressure at all,
and the size is a wet size in liters. We'd pass in zero as the working
pressure, and if the volume units were set to cubic feet, the logic in
"get_volume_string()" would happily convert the metric wet size into the
wet size in cubic feet.
But that's completely wrong. An imperial cylinder size simply isn't a
wet size. If you don't have a working pressure, you cannot convert the
cylinder size to cubic feet. End of story.
So instead of having "get_volume_string()" have magical behavior
depending on working pressure, and getting it wrong anyway, just make
get_volume_string do a pure volume conversion, and create a whole new
function for showing the size of a cylinder.
Now, if the cylinder doesn't have a working pressure, we just show the
metric size, even if the user had asked for cubic feet.
[Dirk Hohndel: added call to translation functions for the units]
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
This actually didn't make a difference for the common case, since our
simplified gas compressibility model had a compressibility factor of 1.0
up to 200 bar, and increased smoothly from there. As a result, the
common 2400 and 3000 psi workpressures didn't really see an effect from
this.
Not taking compressibility into account does kind of make sense for
cylinder naming, since the cylinder may be used for different gases with
very different compressibility characteristics.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
This marks "surface_volume_multiplier()" static in preparation for
changing it to use an actual honest-to-goodness compressibility
estimation. Without that, it wasn't obvious that the function wasn't
used in other random places.
Also, remove the "wet_volume()" function. It was unused, but more
importantly, it was wrong. Yes, it was the inverse of "gas_volume()",
but when you calculate wet volumes from the imperial sizes, you don't
actually use the "real" gas volume, you use the idealized one.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
The code was wrong (and in the case of metric display for weights >= 20kg,
spectacularly wrong) in more or less all cases.
Rounding. It's good for the sole.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Instead of re-calculating all the interpolation data for each plot entry
(which means that we have a quadratic algorithm that walks over all the
plot-info points for each plot-info point), we can just update it
incrementally within any particular interpolation segment.
The previous cleanups made the code sane enough to understand, and makes
it trivial to see how you don't have to recalculate the full thing.
This gets rid of the O(n**2) algorithm, and it instead becomes O(n*m)
where 'n' is the number of plot entries, and 'm' is the number of gas
segments (which is usually a much smaller numer, typically "1").
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
With the two bigger simplications, this just re-organizes the code to do
the "interpolate.pressure_time" update that is shared among all the
"after segment start" cases in just one place.
That leaves the get_pr_interpolate_data() much simpler, and makes it
much clearer what it actually does.
In particular, it becomes very obvious that "interpolate.pressure_time"
is constant for one particular segment (it's the total pressure time),
and that "interpolate.acc_pressure_time" is the one that gets updated
for every entry.
The next step is to only call this for the first entry, and then update
just the "acc_pressure_time" in the caller.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Getting rid of the pointless always-zero pressure now makes it obvious
how some of the remaining code can just be removed too: there is no
point in re-initializing the pressure_time entries to zero at the
segment start, because they started out zero and we just checked that we
don't do anything to them before we hit the segment start.
Similarly, now that the silly pressure testing is gone, it is obvious
that the code for "i < cur" and "i == curr" cases is identical, and the
two cases can just be collapsed.
Signed-off-by: Linus Torvalds <torvalds@ linux-foundation.org>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
In the function fill_missing_tank_pressures(), we only ever call
get_pr_interpolate_data() if "pressure" is zero. So passing it in as an
argument, and then testing whether it is zero or not, is just totally
pointless, and only obfuscates things.
This whole thing seems to be due to people editing the code over time,
with the tests becoming superfluous as the code around it changed, and
nobody looking at whether it actually made sense any more.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
We (ab)use fake_dc() to create a pleasing profile for a manually added
dive. Based on it's intended use, fake_dc() simply handed back a dc
structure that pointed at staticly allocated samples - that's obviously
(now that I think about it) going to blow up in my face if I edit a
manually added dive more than once.
So now we have an option for fake_dc() to actually allocate the samples -
this way the rest of the code can treat these samples as we would treat
samples created any other way. We can free them and replace them with a
new set.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
- coding style (ugh - I should have fixed that when I first committed them)
- remove redundant variables
- add similar code to the length and temperature helpers
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Right now this just tests for zero duration, but maybe this should also
return true for positive duration and max depth of 0.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
If cylinder does not have start and end pressures assigned, attempt to
grab them from the samples instead.
Signed-off-by: Miika Turkia <miika.turkia@gmail.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
It is better to use the proper function to test if cylinder is in use
than just checking the description.
Signed-off-by: Miika Turkia <miika.turkia@gmail.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Now it is possible to enter a specific unit that is different from the
unit stored in the preferences. If only numbers are inputed the unit will
be the same as specified by the users preferences.
Signed-off-by: Joakim Bygdell <j.bygdell@gmail.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Once we identified that our filename is actually a git designator (as seen
by the fact that it ends in a [branchname] surrounded by '[]'), we
shouldn't try to open that filename in order to try other ways of parsing
the data; instead we should just return an error to the caller.
This way the calling code can tell that an error occured.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Since we only show the first cylinder we can also only edit the first cylinder.
Signed-off-by: Joakim Bygdell <j.bygdell@gmail.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
When we detect a redundant DC we free the memory reserved for the model.
Thus we need to malloc that memory here.
Fixes#1002
Signed-off-by: Miika Turkia <miika.turkia@gmail.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
We used to mark CCR dives by having "SP change" event at time 0:00.
As we nowadays mark CCR dives by setting dc->divemode appropriately,
better to convert the old dives to this format as well. This way we do
not have to take the special old format into account on multiple places
in the source.
Signed-off-by: Miika Turkia <miika.turkia@gmail.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
As per Tomaz recomendation the helper functions from 19588ce and e072596
are moved from qmlmanager to DiveObjectsHelper.
[Dirk Hohndel: merged with the latest code]
Signed-off-by: Joakim Bygdell <j.bygdell@gmail.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
The way sectioning of the dive list works is by watching for different strings
in the section.property. In order to be able to tell different trips apart we
combine the address of the dive trip variable with the location (which will
create a new section for a new trip, even if the location text is the same) and
then strip that information out before showing the trip header.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Cloud Storage is a non-gui based class, we currently use
two different approaches for cloud storage, one on the desktop
target and other on the mobile target, we should use only one.
Signed-off-by: Tomaz Canabrava <tomaz.canabrava@intel.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
The shorts where being used on the preferences since a long
while and we cannot just simply change them to bool since this
could break the preferences files, so work around that by
changing them to booleans, since it's the correct type for a
true / false answer.
Also, move some plot curves to the new settings style
Signed-off-by: Tomaz Canabrava <tomaz.canabrava@intel.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
start of the QSettinsg Object Wrapper usage on the code
this first patch removes two macros that generated around
200 lines in runtime for something like a quarter of it
Basically, whenever we changed anything we called the
PreferencesDialog::settingsChanged and connected everythign
to that signal, now each setting has it's own changed signal
and we can call it directly.
The best thing about this approach is that we don't trigger
repaints for things that are not directly profile related. (
actually we still do, but the plan is to remove them in due time)
this commit breaks correct atualization of the profile (because
everything was connected to PreferencesDialog::settingsChanged)
and now I need to hunt a bit for the correct connections
Signed-off-by: Tomaz Canabrava <tomaz.canabrava@intel.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
This is not hoocked up on Subsurface code yet, but it's already
being compilled. now I just need to hoock things up.
Signed-off-by: Tomaz Canabrava <tomaz.canabrava@intel.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
