| Age | Commit message (Collapse) | Author |
|---|
|
Division is expensive, so replace it with multiplication instead. But
don't multiply by 0.1 (inexact in floating point), multiply by 10 and
then do one division at the end.
Make sure the final division is at the very end, so that the result
isn't immediately used. That allow the division to overlap with the
function return overhead, hiding it further.
This is silly, but while thinking about different file formats and doing
profiling of loading big files, it turned out that "strtod_flags()"
actually showed up in profiles. Not very high, but at more than 1%.
This makes the common case (no exponent) use only addition and
multiplication until the very end, and makes the division be the very last
thing it does, which minimizes the data dependencies on the division.
For my stupid test-case, it cut the cost of strtod_flags() in half
according to the profile. The half a percent speedup on loading time isn't
really noticeable or even measurable outside of profiling startup costs,
but rather than carry this along in my tree or just throw it away, I'm
sending it out to see if anybody cares.
Note that we could avoid the final division by instead multiplying
"decimal" with 0.1 rather than multiplying by 10 (and switching the sign
test over), but that's a fundamentally inexact operation in binary floatig
point, so doing the "multiply by tens for decimals" ends up keeping
everything exact as long as possible.
For our use, we probably really don't care, but whatever. End result: this
should not only speed things up immeasurably, it *might* also make things
more precise at a level that we really don't care about :^p
I'm really selling this piece of crap, aren't I?
[Dirk Hohndel: sorry - had to pull the full email into the commit message
this is so good, you couldn't make it up]
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
Maybe not yet perfect, but better.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
Show decimal depth and make the line length more appropriate for the dive
profile shown.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
Unused variables, more initialization order complaints.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
Alternatively, we could use fabs() to determine the absolute value of
floating point arguments.
The author of commit b6a30dcdd3225 ("Improve floating point equality
test") clearly has a rather loose definition of "improve". And the
maintainer who accepted that patch shares the blame...
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
Fiz zvalue of the partialPressure Gas Graph over the TimeAxis, and also
corrects the position of the PartialPressure gas graph. It now no longer
appears to 'float' over the TimeAxis, now it's on the very top of it.
Signed-off-by: Tomaz Canabrava <tomaz.canabrava@intel.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
Signed-off-by: Tomaz Canabrava <tomaz.canabrava@intel.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
This patch creates better defaults for the grid lines, size, position and
z-level.
Signed-off-by: Tomaz Canabrava <tomaz.canabrava@intel.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
Create the grids, and the way that they move is fabulous <3
Signed-off-by: Tomaz Canabrava <tomaz.canabrava@intel.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
Change the axis update method to move the lines alongside the texts.
Signed-off-by: Tomaz Canabrava <tomaz.canabrava@intel.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
This function is for non-duplication of code for different QLists, since I
now have Texts and Lines, I needed to make the code in a way that it
worked without dupplication.
It's a pretty standard use of templates, I think that even the C guys will
not be offended by it. :)
Signed-off-by: Tomaz Canabrava <tomaz.canabrava@intel.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
This class was pratically a re-do of the axis, I'll do the axis to plot
the lines when it feels like it.
Signed-off-by: Tomaz Canabrava <tomaz.canabrava@intel.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
The events were static on the canvas even if the profile changed its size
because of a toggle of the partial pressure gas. This patch makes events
move on the canvas to their correct place.
Signed-off-by: Tomaz Canabrava <tomaz.canabrava@intel.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
The first depth label of its axis is always zero, do not show it.
Signed-off-by: Tomaz Canabrava <tomaz.canabrava@intel.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
That particular dive didn't have a temperature, and thus we got a crash
while accessing the last temperature text.
Signed-off-by: Tomaz Canabrava <tomaz.canabrava@intel.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
This way the partial pressure graph can have smaller pressure labels than
all the other labels on the profile.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
Previously all text in the new profile was deawn in the same font. With
this change the labels on all axes are smaller.
It might be even better to allow per-axis configuration of the label size
as along the time axis the bigger size looked better. But especially for
partial pressures this looks much better.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
Signed-off-by: Miika Turkia <miika.turkia@gmail.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
If the user never clicks on the log and dump checkboxes (and that's what
we expect to be the case by default), then the log / dump flags (and
filenames) were never initialized.
I am reasonably certain this will close the following three bugs, that all
show the same symptom: they behave as if libdivecomputer dump was set,
even if it wasn't.
Fixes: #426
Fixes: #431
Fixes: #435
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
Used gas mixes and gas changes are imported. Also po2, ndl, cns and
ceiling are added to profile samples. As far as I can tell, the Searwater
Desktop shows ceiling in 3 meter (or feet equivalent) steps, but stores in feet
(or probably meters). I just use the value reported, no conversion to 3 meter
steps.
Fixes #432
Signed-off-by: Miika Turkia <miika.turkia@gmail.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
Sqlite database from Shearwater Desktop log software is imported. Just
the basic information like location, buddy, notes and dive profile
(depth and temperature).
This is tested with a DB in Imperial units, thus metric input might
contain errors.
Signed-off-by: Miika Turkia <miika.turkia@gmail.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
Move the opening of DB connection to occur before DC dependent code.
This way we can try to detect log software before calling the DC
dependent import function. This prepares for adding support for
Shearwater sqlite database.
Signed-off-by: Miika Turkia <miika.turkia@gmail.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
Signed-off-by: Tomaz Canabrava <tomaz.canabrava@intel.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
The last temperature text used to have the same align flags
as all the other texts: Right. this makes it much more appealing.
Signed-off-by: Tomaz Canabrava <tomaz.canabrava@intel.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
The code shamelessy copied from the old profile introduced
a bug where the old temperature was not correctly shown.
I'v added a new member to the class that will store the
last valid temperature, and use that to calculate if there's
a reason or not to display it.
Signed-off-by: Tomaz Canabrava <tomaz.canabrava@intel.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
Just removed the unused tanks from the toolbar,
since it's just a removal of something on the ui
related code, it's good to see how much verbose
the xml to generate the UI is.
Signed-off-by: Tomaz Canabrava <tomaz.canabrava@intel.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
This commit was jointly authored by Pedro Neves and Willem Ferguson.
Pedro added a FAQ section to the manual. Many long lines of text were
broken into shorter lines.
Signed-off-by: Willem Ferguson <willem@willem-Precision-M4700.(none)>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
First attempt to use the tools on one of our source files.
So far so good. Not perfect, but a HUGE improvement.
Most importantly it effortlessly fixed the Allman style braces introduced
in commit 51220f26ef71 ("Add recent files to main menu.").
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
The struct / class regex was way too relaxed.
Having the parent class on the same line usually looks better.
clang-format appears to do something odd with continuation strings - even
with UseTab: Always it indents those with four spaces.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
Add(up to four) recent files to File main menu.
Signed-off-by: Boris Barbulovski <bbarbulovski@gmail.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
This hasn't been tested enought, but it seems to get really close.
It assumes that clang-format is in your patch.
Run
perl scripys/whitespace.pl FILENAME
and you'll get a diff of what it things is wrong with that file.
If you like what you see, simply pipe the output into patch -p0
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
rint() is "round to nearest integer", and does a better job than +0.5
(followed by the implicit truncation inherent in integer casting). We
already used 'rint()' for values that could be negative (where +0.5 is
actively wrong), let's just make it consistent.
Of course, as is usual for the messy C math functions, it depends on the
current rounding mode. But the default round-to-nearest is what we want
and use, and the functions that explicitly always round to nearest
aren't standard enough to worry about.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
We should *not* default to the incorrect "1 bar". Instead, we should
leave the resuling pressure at 0 mbar, which leaves visual entries empty
and uses the default surface pressure for calculations.
Reported-by: Pedro Neves <nevesdiver@gmail.com>
Cc: Patrick Valsecchi <patrick@thus.ch>
Cc: Dirk Hohndel <dirk@hohndel.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
This way the dialog is centered and the main window gets dimmed.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
This isn't broken, but it's confusing.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
C++ style of accessing single instance class object.
Signed-off-by: Boris Barbulovski <bbarbulovski@gmail.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
When we move, shrink or expand the depth axis, the
meandepth should also change it's position. this
patch adds that.
Signed-off-by: Tomaz Canabrava <tcanabrava@kde.org>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
The extra axis labels on the Temperature and Cylinder Pressure
graphs are not wanted in release mode as they create a lot of visual
clutter, but they are useful to have in debug mode since we can then
better understand what went wrong.
Signed-off-by: Tomaz Canabrava <tcanabrava@kde.org>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
Same as the last commit, but for cylinder pressure. Another bug
was spotted, where the mean depth line does not move to the correct
location after a axis-size-changed - fixing that on next commit.
Signed-off-by: Tomaz Canabrava <tcanabrava@kde.org>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
This patch adds shrinking of the Temperature axis when partial
pressures are shown. This adds an unwanted side effect however,
the axis started showing it's values - and we didn't do that on
the gtk version or on the old profile. While this is good for
debugging, it's not wanted for the software if it's on release
mode. I'll fix that in due time.
Signed-off-by: Tomaz Canabrava <tcanabrava@kde.org>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
When the user checked the option to show or hide the partial
presure graph the size of the profile graph was kept the same;
usually I do those changes on the objects themselves, on the
settingsChanged method, but since the calculation of the
size of the profile was done on the ProfileView class for
consistency with the other objects, I had to change it there
too.
Signed-off-by: Tomaz Canabrava <tcanabrava@kde.org>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
The coding style for switch isn't the prettiest, but let's at least keep
things consistent.
Also, those ugly break statements never got reached, anyway.
And there was no need to re-declare and re-generate currentString.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
Don't have nested loops with the same loop variable. Really. Even if it is
legal C++.
And don't declare local variables more than once. This will only cause
issues later.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
Otherwise, why pass in the parent to begin with?
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
If you need to use this->error to distinguish the local variable from the
object membe that should be a hint that maybe you didn't pick the best
name for the local variable.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
|
Since the variable was inside of an inner scope this was technically legal,
but it's just too annoying for words.
(The diff in the commit doesn't make this obvious, but outside the for
loop in the same function there is a divedatapoint *dp, so we had a
pointer to divedatapoint and a divedatapoint with the same name...)
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
|
Linus Torvalds
Dirk Hohndel
Tomaz Canabrava
Miika Turkia
Willem Ferguson
Boris Barbulovski
Tomaz Canabrava