Instead of passing a uuid, pass a pointer to the dive site.
This is small step in an effort to remove uuids.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
This changes more of the dive-site interface to return pointers
instead of UUIDs. Currently, most call sites directly extract
UUIDs afterwards. Ultimately, the UUIDs will be generally replaced
by pointers, which will then simplify these callers.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The uemis downloader uses a cache for location to divesite id.
Trivially, the divesite-uuid can be replaced by a pointer. This
is a tiny step to remove divesite UUIDs.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
As a first step in removing dive-site uuids, change the interface
of the get_dive_site_*() functions to return pointers instead
of uuids. This makes code a bit more complicated in places where
the uuid is extracted afterwards (needed NULL check). Nevertheless,
these places should disappear once pointers instead of uuids are
stored in the dive-structures.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Including ssrf-version.h in an include file that a lot of files depend on
caused a ton of unnecessary recompiles with every commit. This should reduce
that problem.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
The reverseGeoLookup() function defined in qthelper.cpp has long
ago moved to its own compilation unit. It is not even defined in
the headers anymore. Remove it and the now unnecessary <QNetwork*>
includes.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
When creating a trip header (as it is used in the mobile app right now), we need
to show the number of dives shown, not the total number of dives in that trip.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
We assume that any dive that gets added to a trip initially gets shown. The
filter logic then needs to make sure it adjusts this number (which then makes
it easy to tell the user how many dives of that trip are visible with the
current filter).
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
This feels a bit like cheating, but if we need to be able to make modifications
to the underlying dive and only have the helper object (for example inside of a
view model), doing everything through the helper object can turn into a real
performance issue.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
With this we can make it configurable if the search should include the notes field
and if the search should be case sensitive or not.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Otherwise we could match the end of one string and the beginning of the next,
so having a buddy name Linus and a dive master named Alvin would be matched
by USA.
Also add Notes to the full text search (I had forgotten those earlier).
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
A static variable was used to collect two parts of the location
string. Instead, explicitly store these two parts in two different
strings and concetenate them. One fewer variable and the program
flow is thus hopefully more clear.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Introduce a parser_state structure, which describes (most) of the
global parser state. Create such a structure in the entry routines
to the parser and pass it down to the individual functions. The
parser state is initialized and freed with the init_parser_state()
and free_parser_state() functions.
The main benefits are:
1) Isolation of parser state.
2) Keeping the global name space tidy.
3) Prevent memory leaks which could happen in truncated files by
freeing all the parser state after parse.
A somewhat controversial point might be that the individual
parsing functions are split in those that need parser-state and
those that don't. This means that there are now two versions of
the MATCH macro, viz. one for the former and one for the latter.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
This is consistent with most other free_*() functions in the core
code and will make cleanup of parser state less verbose.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Export the free_picture() function from dive.c. The parser may need
this in case of truncated files to free its temporary resources.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Recently, the subsurface webservice was removed. Remove the corresponding
code in the parser. This removes a static variable, which was used
to generate unique dive-site ids.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The variables country and city used in divinglog_place()
were never freed. Free them when the pointers are reset.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The utf8_string() function is used to extract whitespace-trimmed
strings. The function would happily overwrite the pointer to
the old string, which could therefore leak (suppose an XML has
redundant attributes).
Therefore preemtively free the string output parameter. This makes
it of course necessary to only pass in NULL-initialized pointers
or pointers to owned string.
The code survives the current set of parser-tests.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Instead of having people treat latitude and longitude as separate
things, just add a 'location_t' data structure that contains both.
Almost all cases want to always act on them together.
This is really just prep-work for adding a few more locations that we
track: I want to add a entry/exit location to each dive (independent of
the dive site) because of how the Garmin Descent gives us the
information (and hopefully, some day, other dive computers too).
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The deco parameters need special treatment to wire them up to the
underlying deco model code. And with the new preferences setup this
is a lot of boilerplate.
Signed-off-by: Jan Mulder <jlmulder@xs4all.nl>
On parsing of dive computer extra data, key/value pairs are stored
in global state. They are added to the dive computer with
add_extra_data(), which makes a copy of the string. The local
copies of the strings are never freed.
free() the strings after storing them. The data still leaks in case
of unfinished parsing of extra_data tags, but this will be
taken care of in a subsequent commit.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
delete_current_divecomputer() had some duplicate code to release
dive-computer resources. Use the free_dc_contents() function instead.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The serial and fw_version strings of struct divecomputer were copied
by pointer. This worked because they were never freed or modified.
Instead, do a deep copy of the strings and free them when appropriate.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
On dive computer copy, the extra data (a list of key/value pairs)
was simply copied as a pointer. This worked because the list was
never freed nor modified. Copy and free the list on dive computer
copy and free, respectively.
This made it necessary to move the STRUCTURE_LIST_* macros up in
the dive.c file.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
In commit e5dca8228e a fixed order
of the arguments to merge_dives() was introduced: first dive old,
second dive downloaded. This made the dl variable, which pointed
to the downloaded dive useless. One instance was forgotten, which
led to a null-dereference.
Remove.
Reported-by: Dirk Hohndel <dirk@hohndel.org>
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
xml_parsing_units stores the units of the currently parsed XML
file. It is not used outside of parse-xml.c. Therefore, make
it of static linkage and remove the declaration from dive.h.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
sanitize_cylinder_type(), which is indirectly called from fixup_dive(),
had ft^3 -> mliter conversion code, which was executed on the
condition "xml_parsing_units.volume == CUFT".
But nowhere in the code base would xml_parsing_units.volume ever be
set to non-metric. Moreover, xml_parsing_units reflects the units
of the latest parsed XML file, but fixup_dive() is called in numerous
contexts not related to XML parsing. Therefore, the whole piece of
code seems highly questionable.
Remove this code.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Dive site data was collected in "cur_dive_site", which was then
merged into an existing or a new dive site. But only the struct
dive_site pointed to by "cur_dive_site" and the taxonomy data
were freed, not the textual data such as name or description.
Therefore, split out the approrpriate free-ing from the
delete_dive_site() function and call that instead of a simple
free().
A similar situation occured for dives that would not be added
to the dive-table because they were deemed incomplete. Use
free_dive() here instead of a simple free() too.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
I incorrectly thought that 'ci_pointing_to_guiding_tissue' was the only
missing initialization, because that is the only one valgrind pointed at.
... that is, until I started looking at a few more dives, which showed
that there were other parts tht weren't initialized either, like
double tolerated_by_tissue[16];
double tissue_inertgas_saturation[16];
double crushing_onset_tension[16]; // total inert gas tension in the t* moment
so just make sure to clear the whole data structure, to avoid any random
behavior due to uninitialized deco state.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
In select_dive(), the selected dive would only be made the
current dive, if it wasn't previously selected. If multiple
dives were selected and the user clicked on one of them which
is not the current dive, then the current dive would be
deselected and thus not be the current dive anymore. The
only remaining dive would not be made the current dive,
because it was already selected. End result: null dive shown.
Therefore, always make the selected dive the current dive,
even if it is already selected.
Fixes#1792
Reported-by: Jan Mulder <jlmulder@xs4all.nl>
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
copy_dc_renumber() is an internal function to copy dive computers
and renumber the cylinders. Since only the structure was copied,
in the case of multi-dc dives, the merged dives shared the same
computer. If one of them was freed, use-after-free crashes would happen.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
This was used to track whether we had selected the native BT mode in the
download dialog. But the information is redundant as we can tell from the
device name whether this is a BT/BLE download or not.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
process_imported_dives() might delete the currently selected
dives. This could lead to use-after-free problems. Therefore,
reset the currently selected dive to the last dive.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
On dive import, dives to be added may be merged into already
existing dives. In such a case, the dive to be added is deleted.
Before doing so, it must be removed from the trip is belongs to
to avoid corruption of the trip-list.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
On merging make a deep copy of the picture list, to avoid a use-after-free
crash after the orginal dive is deleted.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
In commit 8c2383b495 dive merging was
changed to not modify the original dive. On import, dives were then
merged and the original deleted. The merge_weightsystem_info() was
not adapted accordingly (deep copy of string instead of pointer),
leading to a use-after-free crash.
Resolve this by doing a deep copy.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Instead of calling free() on all dives, call free_dive() which also
frees additional allocated data, not only the dive struct.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
There were two declaration of clear_table(), one in dive.h and one
in parse.h. The definition was in parse.c. Since the parser doesn't
even use the function, move the function and its declaration to
divelist.[ch] and remove the redundant declaration in dive.h.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
This variable was only used in the divinglog_dive() function. There,
it was initialized right at the beginning and therefore there seems
to be no point in conserving its value across function-calls.
Make the variable local and remove the global version.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Comits f427226b3b and 43c3885249 of the undo series introduced 2 calls
of autogroup_dives() without checking the autogroup global boolean.
This is a bug. An import from DC (for example) then triggers an
autogrouping, the divelist is autogrouped, and the UI button
is off.
This commit solves this. I've chosen for a guard in the autogroup_dives()
that now is a no-op when called when the user did not select autogrouping.
In additon, simplified the other calls to this function, as we do
not need to check before calling any more.
Signed-off-by: Jan Mulder <jlmulder@xs4all.nl>
The dive-site-edit widget uses a copy of the to-be-edited site
to compare with old values. Generally, this seems overkill
(the original dive-site can be used for such a comparison).
But one place where it can't simply be removed is the taxonomy,
because the widget needs a place to store the unsaved data.
Change the code to use an explicit taxonomy structure instead
of the one provided in the copy. This should ultimately allow
removal of the latter.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The global object displayed_dive_site is used a a backing-store
by the dive-site-edit widget. All external accesses were removed,
therefore make the object local to the widget.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The global object "displayed_dive_site" is used to store the
old dive site data for the edit-dive-site widget. The fields
of the widget were initialized from this object in the show
event. Therefore the object was updated in numerous parts of
the code to make sure that it was up-to-date. Instead, move
the initialization of the object to the function that also
initiatlizes the fields. Call this function explicitly before
showing the widget.
This makes the data-fow distinctly easier to understand.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The name seems crazy until you realize that FQ is 0x4651 which is the model
number of the i770R. And the six digits are the serial number of the device.
Still crazy, but at least now you understand WHY.
Thanks to Jef for decoding that.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
This makes no sense, but apparently we need to start a fresh scan in order to be able
to talk to a different BLE dive computer on the Mac.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
On macOS, we cannot connect to a BT/BLE device until we have scanned it. Right
now this just sits quietly and waits, which given how long this can take is
rather unsatisfying and might look like Subsurface is hung.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Since we trigger a scan even without the dialog to pick the right
device, we need to remember all devices that we find.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Instead of only starting the scan when explicitly asked to do so in the BT
dialog, create the discovery agent when the download dialog opens, since on
macOS we cannot connect to a device without having scanned for it first.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
The reverseGeoLookup() fetches dive-site data via GPS coordinates.
The coordinates and the result were passed via the global
"displayed_dive_site" object. To make data-flow more clear,
pass data as in and out parameters instead.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
By making reply a std::unique_ptr<>, the function can be quit
from any point and the reply will be freed. This is valid according
to Qt's documentation as we're not deleting during signal processing.
This commit fixes a leak: reply was overwritten with the address of
a new object without freeing the old object.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
According to Qt's docs, the QNetworkAccessManager is supposed
to be a long-living object. Therefore, don't create one on
every geo-lookup, but a single object for all geo-lookups.
By making the object function-local it is only initiaized
on first use. Morover this limits the amount of concurrent
geo lookups.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Fetching the taxonomy from GPS coordinates was implemented in
a QThread. But the only access to the main function was a
direct call to run(). Thus, the thread was *never* started.
The function call was always asynchronous [it was using an
event loop though, so the UI doesn't hang]. Notably this
means that the signals connected to the thread would never
fire. And the spinner would never be activated.
Thus:
1) Turn the thread into a simple function.
2) Remove the spinner.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Ultimately, we want to use a single dive-list and not replicate
it in the Qt-model code. To this goal, let's start with using
the same sort function.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
For undo, it is crucial that commands don't modify existing dives.
Unfortunately, dive merging would write into the data-structures
of the to-be-merged dives. To prevent it from doing so, make the
input dives const-pointers.
This led to a whole cascade of functions that had to take const
and significant churn.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Select the proper dives after the add, remove, split and merge
dives commands on undo *and* redo. Generally, select the added
dives. For undo of add, remember the pre-addition selection.
For redo of remove, select the closest dive to the first removed
dive.
The biggest part of the commit is the signal-interface between
the dive commands and the dive-list model and dive-list view.
This is done in two steps:
1) To the DiveTripModel in batches of trips. The dive trip model
transforms the dives into indices.
2) To the DiveListView. The DiveListView has to translate the
DiveTripModel indexes to actual indexes via its QSortFilterProxy-
model.
For code-reuse, derive all divelist-changing commands from a new base-class,
which has a flag that describes whether the divelist changed. The helper
functions which add and remove dives are made members of the base class and
set the flag is a selected dive is added or removed.
To properly detect when the current dive was deleted it
became necessary to turn the current dive from an index
to a pointer, because indices are not stable.
Unfortunately, in some cases an index was expected and these
places now have to transform the dive into an index. These
should be converted in due course.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
In DiveListView, we have a very fundamental problem: When
On the one hand, we get informed of user-selection in the
DiveListView::selectionChanged() slot. This has to set the
correct flags in the C-backend.
On the other hand, sometimes we have to set the selection
programatically, e.g. when selecting a trip. This is done
by calling QItemSelectionModel::select().
But: this will *also* call into the above slot, in which
we can't tell whether it was a user interaction or an
internal call. This can lead to either infinite loops or
very inefficient behavior, because the current dive
is set numerous times.
The current code is aware of that and disconnects the
corresponding signal. This is scary, as these signals are
set internally by the model and view. Replace this
by a global "command executing" flag in DiveListNotifier.
The flag is set using a "marker" class, which resets the flag
once it goes out of scope (cf. RAII pattern).
In DiveListView, only process a selection if the flag is not
set. Otherwise simply call the QTreeView base class, to reflect
the new selection in the UI.
To have a common point for notifications of selection changes,
add such a signal to DiveListNotifier. This signal will be
used by the DiveListView as well as the Command-objects.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Don't delesect dives, when unregistering them from the backend.
If a previously selected dive is added, select it in the dive-list.
For this purpose introduce a SELECTED_ROLE to query the DiveTripModel
for selected dives.
Unfortunately, when adding multiple selected dives, current_dive_changed
is called for each of them, making this very slow. This will have
to be fixed in subsequent commits.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
If the autogroup flag is set, search for appropriate trips in
DiveAdd() and add the dive to this trip. If no trip exists, add
a new trip.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Previously, each dive-list modifying function would lead to a
full model reset. Instead, implement proper Qt-model semantics
using beginInsertRows()/endInsertRows(), beginRemoveRows()/
endRemoveRows(), dataChange().
To do so, a DiveListNotifer singleton is generatated, which
broadcasts all changes to the dive-list. Signals are sent by
the commands and received by the DiveTripModel. Signals are
batched by dive-trip. This seems to be an adequate compromise
for the two kinds of list-views (tree and list). In the common
usecase mostly dives of a single trip are affected.
Thus, batching of dives is performed in two positions:
- At command-level to batch by trip
- In DiveTripModel to feed batches of contiguous elements
to Qt's begin*/end*-functions.
This is conceptually simple, but rather complex code. To avoid
repetition of complex loops, the batching is implemented in
templated-functions, which are passed lambda-functions, which
are called for each batch.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
insert_trip() adds a trip to the backend, but merges trips if
there exists a trip with the same date. This is a disaster
for the MergeTrips command, because this command adds a new
trip and removes the previous two. Of course if the added trip
is merged, this cannot work.
Therefore, add an insert_trip_dont_merge() function, which
adds the trip, but doesn't merge.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
AddDivesToTrip, CreateTrip, AutogroupDives, RemoveAutogenTrips
and MergeTrips basically all did the same thing as RemoveDivesFromTrip,
which was already implemented. Thus, factor our the common functionality
and hook it up to make all these functions undo-able.
Don't do the autogroup-call everytime the dive-list is rebuilt
(that would create innumberable undo-actions), but only on dive-load /
import or if expressly asked by the user [by switching the autogroup
flag].
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
For this, an output-parameter was added to the backend merge_dives()
function. When non-zero, instead of adding the merged dive to
the preferred trip, the preferred trip is returned to the caller.
Since the new UndoObject, just like the delete-dives UndoObject,
needs to remove/readd a set of dives, the corresponding functionality
was split-off in a helper function.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
For this, the core functionality of the split_dive() and
split_dive_at_time() functions were split out into new
split_dive_dont_insert() and split_dive_at_time_dont_insert(),
which do not add the new dives to the log. Thus, the undo-command
can take ownership of these dives, without having to remove them
first.
The split-dive functionality is temporarily made desktop-only
until mobile also supports "UndoObjects".
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Play manual addition of dives via an UndoCommand. Since this does in
large parts the same thing as undo/redo of dive deletion (just the
other way round and only a single instead of multiple dive), factor
out the functions that add/delete dives and take care of trips.
The UI-interaction is just mindless copy&paste and will have to
be adapted.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The original undo-code was fundamentally broken. Not only did it leak
resources (copied trips were never freed), it also kept references
to trips or dives that could be changed by other commands. Thus,
anything more than a single undo could lead to crashes.
Two ways of fixing this were considered
1) Don't store pointers, but unique dive-ids and trip-ids.
Whereas such unique ids exist for dives, they would have to be
implemented for trips.
2) Don't free objects in the backend.
Instead, take ownership of deleted objects in the undo-object.
Thus, all references in previous undo-objects are guaranteed to
still exist (unless the objects are deleted elsewhere).
After some contemplation, the second method was chosen, because
it is significantly less intrusive. While touching the undo-objects,
clearly separate backend from ui-code, such that they can ultimately
be reused for mobile.
Note that if other parts of the code delete dives, crashes can still
be provoked. Notable examples are split/merge dives. These will have
to be fixed later. Nevertheless, the new code is a significant
improvement over the old state.
While touching the code, implement proper translation string based
on Qt's plural-feature (using %n).
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Disable the Download button when one of the fields vendor, product,
connection is not filled in. The app will crash when trying.
In addition, make the underlying core code to actual download
more safe by checking this, and silently fail instead of crash.
And, yes, this is a double fix in this scenario, but the core code
is used in more places, so better safe than sorry.
Signed-off-by: Jan Mulder <jlmulder@xs4all.nl>
Since a known DC will have the name prepended to the BT/BLE addresss
we need to substring match the BT address.
Signed-off-by: Joakim Bygdell <j.bygdell@gmail.com>
By saving the device address together with the vendor and product we fix the
corner case where a user with two DCs would not get quick select buttons if they
where the same vendor and model.
Signed-off-by: Joakim Bygdell <j.bygdell@gmail.com>
The LocationInformationModel added two dummy sites to the front
of the list (add new dive site). This was never used - desktop
uses its own model, mobile only extracts the list of dive site
names with a custom function. Remove this functionality.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
With removal of the git_local_only from the preferences (see
ae653703a5), the users choice, in the mobile app, was not
stored any more in between sessions. This resulted in issue
1725.
So, in order to store that user preference, we need a new
preference. This is added here, but its not yet hooked up
in the app yet. This deals only with the preference handling.
And adapted tests are included.
Signed-off-by: Jan Mulder <jlmulder@xs4all.nl>
We bundle our version of libdivecomputer and don't expect Subsurface to work
with a different version, certainly not with something older than 0.5.
I kept the checks for SAMPLE_EVENT_STRING and DC_FIELD_STRING and DC_SAMPLE_TTS
because maybe there's a situation where being able to compile with a current
upstream version is useful.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Commit 810903bdb9 ("Import: pass a dive table to process_imported_dives()")
introduced the variables
struct dive *old_dive, *merged;
into process_imported_dives(), but never used them. It seems to be an
artifact of having split the function to use the try_to_merge_into()
helper function (that has those same variable names and _does_ use
them), but forgetting the original variables from the pre-split case.
Gcc understandably warns about it:
core/divelist.c: In function ‘process_imported_dives’:
core/divelist.c:1351:26: warning: unused variable ‘merged’ [-Wunused-variable]
struct dive *old_dive, *merged;
^~~~~~
core/divelist.c:1351:15: warning: unused variable ‘old_dive’ [-Wunused-variable]
struct dive *old_dive, *merged;
^~~~~~~~
and the trivial fix is to just remove that line that declares the stale
and unused variables.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
bperrybap reported on github that the ftdi timeouts can be excessive:
"the timeout period while waiting for read data to be 10x or even 100x
longer than it should be when there are read issues on the data cable
particularly when using Android and USB OTG cables. i.e. a 5 second
read timeout for not receiving data can be as long as 7 minutes"
and the reason is that the code at one point tried to use the regular
"gettimeofday()" to handle timeouts, but that doesn't exist in Windows.
We already have Windows-specific code to sleep for a number of
milliseconds in "ftdi_serial_sleep()", let's just extend that same
concept and add a "ftdi_serial_get_msec()" that returns the number of
msec's since some arbitrary point in time.
On Windows, that's just "GetTickCount()", and in sane environments it's
just a trivial wrapper around gettimeofday() to turn sec/usec into msec.
NOTE! The actual msec value doesn't have any meaning. Only the
difference between two calls to ftdi_serial_get_msec() is meaningful.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Some divecomputer backends (ok, right now really only the Aqualung i770R
and i300C) want to know the bluetooth name of the dive computer they
connect to, because the name contains identifying information like the
serial number.
This just adds the support for that to our Qt BLE code.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
If only selected dives were exported into HTML, the statistics would
nevertheless cover all dives. A counter-intuitive behavior. Fix by
adding a selected_only flag to calculate_stats_summary().
Reported-by: Jan Mulder <jlmulder@xs4all.nl>
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The statistics of the selected dives were calculated
a) into a global objects and
b) at a completely different place than where they're used.
There's no plausible reason for either. There fore render
into a caller-provided structure at the place of use.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Statistics were calculated into global variables every time the
current dive was changed.
Calculate statistics only when needed and into a structure
provided by the caller.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
This flag had two distinct uses:
- signal that dives were downloaded, not imported
- use to mark imported dives
Both are not used anymore, therefore remove the flag.
The uemis downloaded misused the flag to mark deleted
dives. Instead misuse the "hidden_by_filter" flag.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
process_imported_dives() is more efficient for downloaded than for
imported (from a file) dives, because it checks only the divecomputer
of the first dive.
This condition is checked via the "downloaded" flag of the first
dive. Instead, pass an argument to process_imported_dives().
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Dive importing is now performed via a distinct table which is
merged into the main dive table. Thus, it is known which of the
dive is new and which is old. This information can now be
implicitely encoded in the parameter-position of merge_dive()
[i.e. pass old as first and new as second dive].
This makes marking of downloaded dives via a flag unnecessary.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Dives are now in all cases imported via distinct dive_tables.
Therefore the "preexisting" marker is useless. Remove.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Dives were directly imported into the global dive table and then
merged in process_imported_dives(). Make this interface more flexible,
by passing an independent dive table.
The dive table of the to-be-imported dives will be sorted and merged.
Then each dive is inserted in a one-by-one manner to into the global
dive table.
This actually introduces (at least) two functional changes:
1) If a new dive spans two old dives, it will only be merged to the
first dive. But this seems like a pathological case, which is of
dubious value anyway.
2) Dives unrelated to the import will not be merged. The old code
would happily merge dives that were not even close to the
newly imported dives. A surprising behavior.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The old surface interval calculation had fundamental issues:
1) process_all_dives(), which calculates the statistics over *all*
dives was used to get the pointer to the previous dive.
2) If two dives in the table had the same time, one of those would
have been considered the "previous" dive.
3) If the dive, for which the surface interval is calculated is
not yet in the table, no previous dive would be determined.
Fix all this by creating a get_surface_interval() function and
removing the "get previous dive" functionality of process_all_dives().
Remove the process_all_dives() call from TabDiveInformation::updateData().
Reported-by: Jan Mulder <jlmulder@xs4all.nl>
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Because some BLE operations can be very slow (device and service
discovery etc), we have some rather excessive default timeout for BLE
(currently set to 12 seconds).
But once we actually have started doing IO, that long timeout can be a
big performance problem, when the libdivecomputer backend has support
for retry and packet loss.
For that reason, libdivecomputer has a 'set_timeout()' function that
allows the divecomputer backend to say how quickly it expects the dive
computer to answer before the backend will start resending packets.
Let's just implement that for the actual IO side of BLE too. The
default timeout value remains the general BLE timeout, and this only
affects the actual IO phase, but it improves things enormously for the
case where there is packet loss at that point.
For example, on the Aqualung i770R, the timeout for packet loss ends up
now being just one second rather than the full 12 seconds of default BLE
timeout. Which gets the retry going much faster.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
When we enable notifications, we actually want to make sure to wait for
that write to have completed before we start communicating with the
device, because otherwise we might lose notification events.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
In commit 30fb7bf35c ("qt-ble: set up infrastructure for better
preferred service choice") I moved the service filtering from the
addService() callback into the "select_preferred_service()" function
that picks the right service for the device.
That was nice for debugging, since it meant that we showed the details
of _all_ services, but it also meant that we ended up starting service
discovery on _all_ services, whether they looked at all interesting or
not.
And that can make the BLE device discovery process quite a bit slower.
The debugging advantage is real, but honestly, service discovery can
generally be better done with specialized tools like the Nordic nRF app,
so the debugging advantage of just listing all the details of all the
services is not really worth the discovery slowdown in general.
So move the basic "filter by uuid" back to the service discovery phase,
and don't bother starting service detail discovery for the services that
we can dismiss immediately just based on the service UUID.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The error handling was incorrect for the case where we successfully
opened the libdivecomputer iostream in divecomputer_device_open(), but
the dc_device_open() call failed.
When the dc_device_open() failed, we would (correctly) not do the
dc_device_close() but we would _also_ not do the dc_iostream_close() to
close the underlying file descriptor, which is wrong.
Normally this isn't all that noticeable, partly because the common case
is that dc_device_open() succeeds if you actually do have a dive
computer connected, but also because most of the time it just leaked a
file descriptor or something like that.
However, particularly for the POSIX serial device case, libdivecomputer
does a
ioctl(device->fd, TIOCEXCL, NULL)
call to make serial opens exclusive. This is what we want - but if we
then fail at closing the serial file descriptor, we won't be able to
retry the import at all because now the next open will fail with EBUSY.
So the error handling was incorrect, and while it doesn't usually matter
all that much, it can be quite noticeable particularly when you have
transient errors.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
If no dives were downloaded in do_libdivecomputer_import(), an
error message would be produced. To check for downloaded dives,
the function would access the global downloadTable instead of
the actual table the dives are imported to (at the moment the
same - but the interface allows for a different table).
Move the error-creation to the caller to avoid this situation.
An alternative option would be to check the actual table the
dives were supposed to be downloaded to. But from a program-logic
point of view "no dives" does not seem like an error condition.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The Uemis downloader determines the dive-number to be downloaded
by either checking the download-table [interrupted connection] or
the global dive table [fresh download].
The downloadTable is passed in the device data structure, but
in the function to determine the latest dive, the global
downloadTable is accessed directly [thus supposing that this
table was passed in device data].
Instead, use the table from device data to avoid funny surprises
should we change to a non-global download table.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
On DAN-file import after each dive except the first, the dive-list
was processed. This seem bogus and inefficient. An artefact from
old code? In any case, remove.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Since we now keep track of up to 4 DCs we don't want to display the last used one
but rather the one that is connected.
Signed-off-by: Joakim Bygdell <j.bygdell@gmail.com>
Currently, we can only delete dives that are indexed in the main
dive table. In the future, we will have to delete dives outside
of this table (e.g. for undo). Therefore, split out the free_dive()
function from delete_single_dive(), which takes an index into
the main dive table.
In the process, adopt the dive freeing-code from clear_dive(),
which frees more data than the code in delete_single_dive().
This potentially fixes a memory-leak.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
This reverts commit 1c4a859c8d,
where the override modifiers were removed owing to the noisy
"inconsistent override modifiers" which is default-on in clang.
This warning was disabled in 77577f717f,
so we can reinstate the overrides.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
In d815e0c947 a dive_table pointer
was added to the parsing functions to allow parsing into tables
other than the global dive table. This will be necessary for undo of
import and implementation a cleaner interface. A few cases, notably
CSV and proprietary formats were forgotten.
Implement parsing into arbitrary tables also for these cases.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
A few of these prototypes were already in import-csv.h.
Put them in an 'extern "C" { ... }' block.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The existing BLE dive computers treat BLE as the packetized protocol it
is, and read whole packets at a time.
However, the Mares BlueLink backend treats it as just a basic "serial
over BLE" transport, and for historical reasons reads the reply packets
in smaller chunks.
This allows that kind of IO behavior, where if the divecomputer backend
reads just a part of a packet, we'll split the packet, return the part
the user asked for, and push back the leftover packet onto the received
packet queue.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
We traditionally only allow samples to have a time format of 'mm:ss', so
if you have a dive over an hour, you would just have a minutes field
larger than 60 minutes.
But Matthew Critchley is trying to import some dives from his VMS
Redbare CCR, and the sample timestamp format he has is of the type
'hh:mm:ss'.
That could be fixed by a xslt translation, but there's no real reason
why we couldn't just support that format too.
Reported-by: Matthew Critchley <matthew.s.critchley@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This is perhaps overly verbose, but the timing details helped figure out
some EON Core download issues, and it's nice to see when things actually
happen.
It's also good to see when the data actually enters our queues, and when
we read and write the packets. That might help debug the issues Fabio
is seeing with the Mares Bluelink.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
We used to just find all services and connect the characteristics change
signal etc to them all, but we really only care about the actual
preferred service that we'll be using.
So move the qt ble signal connection to after we've selected the
preferred service that we will actually be enabling notifications on and
do the writes to.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
XMP is a media-metadata standard based on XML which may be used
across a variety of media formats. Some video-processing software
writes XMP data without updating the native metadata fields.
Therefore, we should aim at reading XMP metadata and give priority
of XMP data over native fields.
Pros:
- Support for *all* common media formats.
Cons:
- XML (complex, verbose, chaotic).
- Does not even come close to fulfilling its promise of being
well defined (see below).
Implement a simple XMP-parser using libxml2. Connect the XMP-parser to
the existing Quicktime/MP4 parser.
First problem encountered: According to the spec, XMP data supposed
to be put in the 'XMP_' atom. But for example exiftools instead
writes an 'uuid' atom with a special 16-byte uid. Implement both,
more options will probably follow.
Second problem: two versions of recording the creation date were found
1) The content of a <exif:DateTimeOriginal> tag.
2) The xmp::CreateDate attribute of a <rdf:Description> tag.
Here too, more versions are expected to surface and will have
to be supported in due course (with an obvious priority problem).
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
We use that in the mobile app to scale the whole app, as all sizes there
are relative to the default font.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
We used to just blindly pick "first" and "last" characteristic from the
preferred service, and that was stupid but happened to work for the dive
computers we supported. Note that for some of them, "first" and "last"
was actually the *same* characteristic, since it could be a single one
that supported both.
However, this first/last hack definitely doesn't work for the Mares
BlueLink BLE dongle, and it's really all pretty wrong anyway.
So re-organize the code to actually look at the properties of the
characteristics. I don't have a BlueLink to test with, but my EON Core
and Shearwater Perdix AI are still happy with this, and the code
conceptually makes a lot more sense.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
And remove some includes and defines that are not used any more after
removal of the GPS webservice code.
Signed-off-by: Jan Mulder <jlmulder@xs4all.nl>
Fix multiple run-time errors in connect call introduced in 504e912512.
1) Set the proper signature of the signal. 2) make the used slot
a real slot (so move it to the proper section in the header) and
3) set the proper signature for the slot.
Highly unlikely that normal users notice the runtime errors and
possibly unwantend behavior, as this all deals with the subtile GPS
service update threshold.
Signed-off-by: Jan Mulder <jlmulder@xs4all.nl>
We used to just pick the first non-standard service we found (with a
special case for the Heinrichs Weikamp dive computers that have an
actual registered standard service).
We then waited for that service to finish discovery, and started using
it.
This changes the logic to wait for _all_ services to finish discovery,
and then after that we pick the one we like best. Right now the rule
for picking a preferred service is the same one we had before, but the
difference is that we now have the full discovery data, so we *could* do
something better.
Plus this makes our debug messages a lot more legible, when we don't
have the mix of overlapping service discovery with the actual IO we do
to the preferred service.
NOTE! This doesn't much matter for most of the dive computers that we
currently support BLE for. They don't tend to have a lot of odd
services.
But at least both the Mares BlueLink and the Garmin Descent both have
multiple services and it's not obvious which one to use, and this will
make it not only easier to debug those, it will make it easier to pick
the right preferred service descriptor to use.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This is not only much clearer (and smaller code), but it also lowers the
latency for the waiting, since we don't always wait for the full 100ms.
Get rid of the now unused "waitfor()" function that just unconditionally
waited for 100ms.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
process_dives() is used to post-process the dive table after loading
or importing. The first parameter states whether this was after
load or import.
Especially in the light of undo, load and import are fundamentally
different things. Notably, that latter should be undo-able, whereas
the former is not. Therefore, as a first step to make import undo-able,
split the function in two versions and remove the first parameter.
It turns out the the load-version is very light. It only sets the
DC nicknames and sorts the dive-table. There seems to be no reason
to merge dives.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
We only store the address part of the connection name, so don't try to find an
exact match, try to find the sub-string.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
This does feel clumsy and complicated. This is a lot of special case
handling and a lot of boilerplate for something that really should be
quite simple.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
In visit_on_node() in core/parse-xml.c the name is extracted into
a static buffer. There seems to be no need for this being static,
as the name is only passed to the entry() function which (hopefully)
does not store a reference to the name anywhere.
If it does, this would need a *big* *fat* comment.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The existing code creates a deterministic ID (not exactly "unique") in order to
help us avoid merge conflicts in git-storage mode. But as a side effect, if we
re-download the same dive twice from a dive computer that supports GPS (right
now only the Garmin Descent Mk1) we are guaranteed to create the same dive site
uuid when we do this. So when we download a dive - whether we will actually
*use* that dive later or not - we will be filling in the dive site information
with the data we got from the dive computer.
... and in the process we will be overwriting any data that was filled in
manually. The name of the dive site, but also possibly even the GPS of the dive
site (maybe the user decided to edit that using the map, because while the
automatically downloaded GPS data was "correct", maybe the user wanted to
change it to be the actual under-water location using the satellite data,
rather than the place where you started the dive or where you surfaced).
In order to avoid this collision, this patch just makes the libdivecomputer
download not use the dive time, but "time of download" for the dive site time,
and thus effectively generate a new uuid for every download.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
On import of dive media, the timestamp is read from the
metadata to check if the image belongs to the selected dives.
The pictures are then listed in a dialog.
Currently, the metadata is read twice if images are outside
of a dive: once in picture_check_valid() and if it turns
out that the picture is not valid again in picture_get_time()
to display the proper timestamp.
Even though metadata-extraction is reasonably fast, this is
a bit of an embarrassment.
Instead, read the timestamps only once in the constructor of
the dialog and from then on only used these timestamps. Keep
the timestamps in a QVector. Rename the picture_check_valid()
function to picture_check_valid_time() and pass a timestamp
instead of a filename.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The merge_one_sample() function adds a sample to the destination
dive if dives are merged. For long periods between samples at surface
depths, it adds a surface interval.
To decrease the number of global objects, make the sample structure
non-static. Of course, initialization of an on-stack structure is
slower. Therefore move it into the corresponding if. Thus, the
structure will be initialized only once per surface-interval.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Unit_system is read from git storage but units are set from locale when the
app starts. To prevent a miss-match between unit and unit_system we have to
always update the preferences variable when set_unit_system is called so that
the user doesn't end up with imperial units when the preferences say metric.
Signed-off-by: Joakim Bygdell <j.bygdell@gmail.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Using dc_serial_open as a fallback to ftdi_open is just wrong, and will
never work, just mask the real error and introduce read herrings.
Signed-off-by: Anton Lundin <glance@acc.umu.se>
Add a checkbox that triggers replacement of all English characters by
x's in notes, buddy, dive guide and (while we are at it) suit.
This is ment for people sharing logs for debugging that are concious
about privacy issues. It leaves the lenth of strings in tact as well
as special charcters as those might be needed to track down a particular
parsing problem.
Signed-off-by: Robert C. Helling <helling@atdotde.de>
Automate calling of setGpsTimeThreshold when qPref..::time_threshold changes
and thus avoiding the need to call setGpsTimeThreshold directly.
Signed-off-by: Jan Iversen <jani@apache.org>
Windows doesn't have nanosleep() unless libwinpthread is used.
Since the nanosleep() usage in serial_ftdi_sleep():
- does not break in case of EINTR
- has input in milliseconds
the WINAPI Sleep() should be a good alternative.
Signed-off-by: Lubomir I. Ivanov <neolit123@gmail.com>
1) Add a missing <div>
2) More importantly: recognize html content via <div>-tags instead of
<table>-tags.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
QML is quite special, the variables needs to start with lower case in C++
and are used starting with upper case in QML
Signed-off-by: Jan Iversen <jani@apache.org>
