In hindsight a very simple bug to fix, but it requires some
knowledge on the inner workings of our git storage. The changes
on merge of dive sites were simply not saved (completely) because
the git storage code has a cache that we need to invalidate
selectively (ie. for the dive we just gave a new dive site uuid)
to get things finally embedded in the overall commit.
The main reason this bug went unnoticed for more than 2 years is
that most people use the XML/SSRF format (where this problem is
non exsistent), and dive site merging is probably not a very
much used feature either.
Fixes: #939
Signed-off-by: Jan Mulder <jlmulder@xs4all.nl>
A bogus key/value pair was introduced in the cylinder,
consisting of a lonely "m" without value. This is caused
by commit 46004c39e2 and fixed in 48d9c8eb6e. See referenced
commits for more info.
Just ignore this key/value pair. No processing is broken
due to this, as the git storage stores only metric SI type data.
In fact, the m unit is superfluous anyway.
Signed-off-by: Jan Mulder <jlmulder@xs4all.nl>
In a session with the profile I saw that the planner spends
a lot of time waiting to obtain the lock for the factor cache.
Most of the time we are only reading that cache and that
is save to do in parallel (according to the Qt IRC channel).
So we can use a QReadWriteLock instead of a QMutex. This
appears to be quite a performance boost, in particular
for VPM-B
Signed-off-by: Robert C. Helling <helling@atdotde.de>
I never realized that my hashes weren't written, because it only
outputs a debug instead of a warning message.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Commit 46004c39e2 introduces a new field in the logbook outputs
(depth of a cylinder). While in XML the depth unit is stored with a space
between value and unit (m), in our git storage, the unit m is without
space. As the git storage parser uses a space to separate individual
key/value pairs, the erroneously saved space results in parsing warnings
when opening the logbook.
The unwanted space is normally saved just after download of a new dive
from the dive computers, so all desktop-git-storage uses are affected,
and more worrying, mobile beta users.
Signed-off-by: Jan Mulder <jlmulder@xs4all.nl>
Internal floating point (FP) calculations should be performed using double
unless there is a very good reason. This avoids headaches with conversions.
Indeed, the vast majority of FP calculations were already done using double.
This patch adapts most remaining calculations. Not converted where things
that were based on binary representations and variables which weren't used
anyway.
An analysis of all instances follows:
core/plannernotes.c, l.404:
This was a comparison between two floats. On the left side, first an integer
was cast to float then multiplied with and integer and divided by a constant
double. The right hand side was an integer cast to a float. Simply divide by
1000.0 first to convert to double and continue with calculations. On the right
hand side, remove the cast, because the integer will be implicitely cast to
double for comparison. This conversion actually emits less instructions,
because no conversion to double and back is performed.
core/planner.c, l.613:
Same analysis as previous case.
subsurface-desktop-main.cpp, l.155:
A local variable representing the version OpenGL version. Turn this into
integer logic. Not only does this avoid dreaded FP rounding issues, it also
works correctly for minor version > 10 (not that such a thing is to be
expected anytime soon).
abstractpreferenceswidget.[h/cpp]:
A widget where the position is described as a float. Turn into double.
desktop-widgets/divelogexportdialog.cpp, l.313:
total_weight is described as float. Use double arithmetics instead. This
instance fixes a truncation warning emitted by gcc.
The function isCloudUrl() was only called in one place, parse_file().
But, isCloudUrl() could only return true if the filename was of the
git-repository kind (url[branch]). In such a case, control flow would
never reach the point where isCloudUrl() is called, since
is_git_repository() returns non-NULL and the function returns early.
Therefore, remove this function. Moreover, adapt the affected if-statement
by replacing "str && !strcmp(str, ...)" with the more concise
"same_string(str, ...)".
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
See 2182167b53. Keep the dupicated code in sync.
Originally-by: Salvador Cuñat <salvador.cunat@gmail.com>
Signed-off-by: Jan Mulder <jlmulder@xs4all.nl>
See commit 6f42ab46da. Unfortunately, this code is duplicated
(and an obvious candidate for code cleanup). So replicate the mentioned
commit here. In fact, the mentioned issue #666 talkes about the mobile
app, and the fix was only done for the desktop.
Originally-by: Salvador Cuñat <salvador.cunat@gmail.com>
Signed-off-by: Jan Mulder <jlmulder@xs4all.nl>
In libdivecomputer, a new divemode is added (DC_DIVEMODE_SCR) useful
for dive computers that have specfic functionality for semi-closed
rebreathers. At this moment, only the HW computers seem to provide
this.
This commit takes care of proper recognition of this new divemode
when importing data from a dive computer.
Tested on an actual import from an OSTC3 that contained
dives in this new mode.
Signed-off-by: Jan Mulder <jlmulder@xs4all.nl>
This is just code cleanup. Jef renamed the CCR divemode constant
in libdivecomputer, but added a define to be backward compatible as
as well (so this rename did not break our Subsurface build).
Obviously, this breaks the build for people that build against an older
libdivecomputer, but I see no reason to do that.
Signed-off-by: Jan Mulder <jlmulder@xs4all.nl>
This gets us the first merge with the upstream iostream implementation.
This requires a small change for serial_ftdi.c to build.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
copy_string() does the same as the current code, but in one instead
of four lines. Strictly speaking, it does not exactly the same thing
because the empty string ("") case is handled differently. copy_string()
returns NULL instead of a copy of "", which is probably preferred anyway.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Current values (1000m, 10 min) may be too long to choose an accurate fix
while automatically applying gpsfixes to dives. They are fine if we are
diving from a static position, but will give wrong positions e.g. while
drift diving.
Reducing the default values to shorter 100m, 5min won't hurt most dives
from shore or static boats, but will make other diving styles get more
accurate gpsfixes.
signed-off-by: Salvador Cuñat <salvador.cunat@gmail.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
If the credential functions return GIT_EUSER, a call to git_remote_fetch
fails, but giterr_last() may return NULL. This led to a crash in
verbose mode.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The list iteration in dive_remove_picture() was buggy and would
crash if handled a picture that is not in the list.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
In this function, a repository is created, but the returned object
is not used. Might just as well free it.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
On iOS save all discovered devices. Later qt_ble_open queries this
list in order to actually connect to the remove device.
The Desktop code stores this data with the list items and only saves
when the "Save" button is clicked. This is not supported with the
current ConnectionListModel implementation.
Signed-off-by: Murillo Bernardes <mfbernardes@gmail.com>
Currently, in is_remote_git_repository(), git URLs of the form
"file://..." are recognized as local and the "file://" prefix is
removed. The shortened URL is then processed as if it was a remote
URL, which of course has to fail. So far so good - this is not
a remote repository after all. But the removal of the prefix is
not propagated to the calling is_git_repository() function and
handling as a local git repository therefore fails likewise.
To fix this issue, move removal of the "file://" prefix one level
up to the is_git_repository() function.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Calculating variations when in recreational mode doesn't make sense, and can
prevent variations from being calculated when switching back to Buhlmann or
VPM-B modes.
Signed-off-by: Rick Walsh <rickmwalsh@gmail.com>
This reenables the computation of plan variations but now in a separate
thread. Once finieshed, a signal is sent to update the notes.
Signed-off-by: Robert C. Helling <helling@atdotde.de>
before we run out of memory. Diving deep with air and small GFhigh
can cause those (try GF 30/70 at 75m with 25+min bottom time)
Signed-off-by: Robert C. Helling <helling@atdotde.de>
Change the merging behavior for the following information:
Divemaster, buddy, suit:
From "(a) or (b)" to "a, b"
Notes:
From "(a) or (b)" to "a\n--\nb"
Signed-off-by: Stefan Fuchs <sfuchs@gmx.de>
This should help us to move parsing that is not XML related to other
files, hopefully making the code cleaner.
Signed-off-by: Miika Turkia <miika.turkia@gmail.com>
Store cylinder.depth in XML files and in git storage.
This info is in fact the gas switch depth of a specific gas/cylinder
in the planner.
This change avoids the need of typing in a user specific depth value
again when replanning an existing planned dive.
Signed-off-by: Stefan Fuchs <sfuchs@gmx.de>
Pass the planner state struct to the profile computation so it can use
deco_time and first ceiling to display VPM-B ceiling.
Signed-off-by: Robert C. Helling <helling@atdotde.de>
For UI responsiveness, we need to be able to run the planner in the background. This needs the
planner state to be localized (and we need to pass a pointer around).
In order to not let too many lines overrun (and to save typing in the future)
I have renamed instances of struct deco_state to ds. Yes this should have gone
to a separate commit but I accidentally commit --amend'ed it.
Computing of planner variations is temporarily disabled.
Unlock the planner when returning early
So we don't deadlock in add dive and recreational mode (which
use the planner without actually planning).
Signed-off-by: Robert C. Helling <helling@atdotde.de>
Add the capability to select the location name from a list, constructed
from the known dive sites in the logbook.
Fixes: #546
Signed-off-by: Jan Mulder <jlmulder@xs4all.nl>
Each preferences object owns its string members. In three cases, pointers
were copied instead of strings, leading to (in the best case) dangling
pointers if the user edited values:
1) In the GET_TXT macro in core/prefs-macros.h
2) In the PreferencesDialog::defaultsRequested() method
3) In main() of the mobile version
This patch fixes these issues, by using copy_string() or copy_prefs()
as appropriate.
The only reason that the old code didn't crash regularly is that the
default_prefs object was only used at startup and defaultsRequested()
is (at the moment?) dead code.
This patch also aligns the backslashes in core/pref.h and fixes a typo.
The declaration of copy_prefs() is moved to the core/prefs.h header.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Currently not mandatory in our code because we never include
prefs.h from a C file today but for the future this could avoid confusion.
Signed-off-by: Stefan Fuchs <sfuchs@gmx.de>
Bool is the correct choice for this option.
int was used before because it was not clear to me how and if I can use
bool in this C file.
Signed-off-by: Stefan Fuchs <sfuchs@gmx.de>
In the old implementation there were two static C-style strings, filename
and path, which were initialized to NULL and filled on first call of
the function (i.e. singletons).
There is no sense in having two static variables indicating whether
this function was called previously. Moreover, there is no point
in remembering filename accross function calls, because it is not
used once path is set to a non-NULL value.
Therefore, make the filename variable non-static and calculate it only on
first invocation (as indicated by a NULL path). Moreover, free() the filename
variable after its use to fix a memory leak of the old code.
The windows code is slightly different in that the temporary filename is
not dynamically allocated.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
This tries to remove subtle ownership issues. When copying preferences
structures, the default filename is copied. But the default preferences
struct simply takes a pointer to a global string which is free()d in main().
Now, this is not strictly a bug because the free()ing of preferences
resources is not implemented. Yet, let's try to make this consistent.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Boolean settings were declared in pref.h randomly as bools and shorts.
Since the code relied anyway on bool being well-defined and identical
on the C- and C++-sides, turn all of them into bools. They use less
space and express intent more clearly.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
In general, the C++-side of the preferences code consistently uses
the bool data type for boolean settings. There are five exceptions,
which use short instead:
showPo2
showPn2
showPhe
saveUserIdLocal
displayInvalidDives
This patch attempts to make the code more consistent by turning
these into bools as well.
Tests showed that writing as short and reading as bool is handled
gracefully by the Qt variant code. Therefore, an upgrade should not
cause user-visible changes to their settings.
As a bonus, two extern declarations of the set_save_userid_local()
function, which is not defined anywhere, were removed.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Make all char * pointers in pref.h const to make it clear that these
strings are not mutable. This meant adding a number of (void *) casts
in calls to free(). Apart from being the right thing to do, this commit
makes the code more consistent, as many of the strings in pref.h were
already const.
While touching core/qthelper.cpp turn three instances of (void*) into
(void *).
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Commit d6c013f303 introduced a cast to avoid a signed/unsigned
comparison warning for all translation units that included
core/dive.h.
Commit 1f8506ce64 then changed the definition of duration_t from
unsigned to signed, inverting the effect of d6c013f303.
Thus, revert d6c013f303 to allow compilation with -Wall without
flooding.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Free a bunch of C-style strings before assigning newly copied strings.
One case was particularly buggy:
prefs.locale.use_system_language = copy_string(qPrintable(value));
Note that prefs.locale.use_system_language is a bool, which of course
always evaluates to true! Probably nobody noticed because a restart
is required when changing locale.
Moreover remove a few double-semicolons.
Stefan Fuchs points out that sometimes you get cylinder duplication when
you merge dives, particularly with a planned dive. For example, if we
had different manual pressures in the two different dives, the cylinders
will be kept separate.
But that also means that we don't want to plot the pressures from those
other cylinders that came from another dive and are now associated with
another dive computer.
Change the "seen" logic for the cylinder to ignore cylinders that are
only mentioned by other dive computers than the active one.
Reported-by: Stefan Fuchs <sfuchs@gmx.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
It's already used in core/gaspressures.c where it was declared
privately, and we'll have a new user in the profile code, so just
declare it in a proper header file like it should have been.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The "prev" cylinder can never be negative since commit 56c206d19f
("For more manual gas pressure details"), so remove stale code that
checks for a case that cannot happen any more.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Some BT devices support both, classical and LE, modes. Users could
choose either by prepending or removing "LE:" in the device address
field. After commit d23bd46a1b, the
device field is always disabled in Bluetooth mode.
Therefore, add a mode combo box to the Bluetooth device selection
dialog. In the default mode (auto), the old code path (based on
the Qt device flags) is used. The two other modes (force LE, force
classical) allow the user to force the preferred behavior.
This feature is meant as a stop-gap measure until a more refined
transport choice is implemented. Therefore, the value of the new
combo box is not saved in the settings, to avoid cluttering of
the preferences with soon to be obsolete entries.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The Windows auto-verbose + log file creation if starting
from a non-terminal has the problem that the print_version()
call is never made becase 'verbose' is updated programatically
in windows.c and not by the user (by passing -v).
To work around the issue:
- move the windows console creation call before *everything* else
- then immediatelly install the message handler
- then see if 'verbose' is set and explicitly call print_version()
print_version() now also has a flag (version_printed), to avoid
printing the version multiple times, if the user decided to add
an extra -v to the Desktop shortcut.
Signed-off-by: Lubomir I. Ivanov <neolit123@gmail.com>
If the user has not started Subsurface from a terminal
make sure that verbosity is enabled (verbose = 1), so that
the log files are populated with information useful for
debugging.
Signed-off-by: Lubomir I. Ivanov <neolit123@gmail.com>
When filling samples with values during DC import fill sticky values
like CNS, NDL, stoptime,... immediately into current sample.
Otherwise we will not fill the sticky values into the last sample
created.
Add two new sticky values: heartbeat and bearing
Signed-off-by: Stefan Fuchs <sfuchs@gmx.de>
Fixup the NDL value to '-1' at the very beginning of a dive.
Some dive computer report a NDL of 0 at the very beginning of a dive
and then only some 10 seconds later they report the correct value
like 240 min for the first time.
Translate this 0 at the beginning of a dive into our internal '-1'
for no info available.
Signed-off-by: Stefan Fuchs <sfuchs@gmx.de>
Type duration_t changed from uint to int.
Default value of '-1' introduced for some of the values in struct sample:
NDL used -1 as default.
Bearing uses -1 as default (no bearing set).
Display pXX, EAD, END, density, MOD only if values are larger than 0.
In profile don't display data from two first and two last plot_data
entries in info box.
Signed-off-by: Stefan Fuchs <sfuchs@gmx.de>
Currently, on Linux, after selecting a Bluetooth device the name of the
device is shown. On reopening the download dialog, on the other hand,
the address is shown. In the device selection dialog both are shown.
This patch changes the download dialog such that both, name and address,
are shown. The bulk of the patch introduces the name of the device in
the preferences and DCDeviceData. It has to be noted that DCDeviceData
is an encapsulation of the libdivecomputer device_data_t. Nevertheless,
the new Bluetooth-name field is, at the moment, not passed through to
libdivecomputer.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
On BLE connection timeout a weird error-message was shown, because
the controller was still in connecting state and no error string was
set. Therefore, handle the timeout case with a special case label.
Moreover, remove three unnecessary calls to disconnectFromDevice(),
which is called in the destructor of the controller anyway (verified
by looking at Qt source).
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Commit 9771255919 introduces a compiler warning due to mismatched
pointer types. Fixed here.
Reported-by: Stefan Fuchs <sfuchs@gmx.de>
Signed-off-by: Jan Mulder <jlmulder@xs4all.nl>
I'm sure this bug has heen here forever, but the CNS clock is
not very relevant for most divers, and even some technical divers
do not care about this value.
However, doing long decompression dives, the value can easily
grow over 100%, and a lot further. For example, the OSTC computers
use 2 bytes to store the CNS value in the profile data, and I
have multiple dives in my logbook going way over 255%.
This all said. Just store the CNS value in an unsigned 16 bit.
Signed-off-by: Jan Mulder <jlmulder@xs4all.nl>
