The merge_events() function was subtly and not-so-subtly broken in a
couple of ways:
- in commit 8c2383b49 ("Undo: don't modify source-dives on merge"), we
stopped walking the event list after we merged the first event from a
dive when the other dive computer had run out of events.
In particular, this meant that when merging consecutive dives, the
second dive only had the first event copied over to the merged dive.
This happened because the original code just moved the whole old list
over when there was nothing left from the other dive, so the old code
didn't need to iterate over the event list. The new code didn't
realize that the pointer movement used to copy the whole rest of the
list, and also stopped iterating.
In all fairness, the new code did get the time offset right, which
the old code didn't. So this was always buggy.
- similarly, the "avoid redundant gas changes" case was not handled for
the "we ran out of events for the other dive computer" case.
This fixes both issues.
Cc: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Smtk2ssrf has a segfault which matches quite well glibc's
CVE-2019-6488 (except for the x32 part).
It came from a call to memcpy in mdb_ole_read() func, used to get the
header and the profile of a dive from the database.
May be it could be fixed in libmdb but Mdbtools project has been stalled
for the past 5 years so ...
The segfault seems to be triggered by an empty profile in the first dive
in the database (a pretty common case in older Aladin DCs due to their
little memmory). The only special thing here is the fact it's the first
dive in the database structure (not the first by its index).
We can avoid the crash if we don't call mdb_ole_read_full() func on zero
sized profile field.
The problem here is we can't get the size of the fields and build the
MdbColumn in the same roud. Happily we just need the MdbColumn struct
for the dive profile and header. So, we can change the previous approach
using MdbColumns through almost all functions to a simpler one using the
already bounded strings by smtk_open_table() and just using the
col[n]->bind_pointer in the main function where the columns are built to
be used by mdb_ole_read_full().
Reported-by: Robert C. Helling <helling@atdotde.de>
Signed-off-by: Salvador Cuñat <salvador.cunat@gmail.com>
So far, if the conversion process failed, we simply returned
an empty file. Now, we report that something's wrong.
Signed-off-by: Robert C. Helling <helling@atdotde.de>
It appears that multi line attributes silently fail. Without this change, the Download
button is enabled if vendor and product are chosen, even if there isn't a connection
selected. With this change (having all three conditions on the same line) the code
works as expected.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
gettimezoneoffset() returns incorrect values when called with a time_t.
Since we only accept the value here if it is within 5 minutes of 'now',
using the current timezone offset is a fair approximation.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
For some reason we suddenly started logging the GPS fixes in UTC instead
of local time. Which caused the matching algorithm to fail (unless you
happened to be diving in UTC). Unclear what broke this, but this seems
like an easy enough fix, since the GPS fix being reported is by
definition "right around now". So using gettimezoneoffset() with the
current time seems "good enough".
I don't know when gettimezoneoffset() with an argument got broken, TBH.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
A while ago, we introduced a preference whether O2 should
be considered narcotic. We used this when computing
best mix or when entering the He content via MND. But
we forgot to make the displayed MND depend on this
preference. This patch add this.
Fixes#2895
Signed-off-by: Robert C. Helling <helling@atdotde.de>
When removing the MAX_CYLINDERS restriction, the layout of the
pressure readings was changed from a (cylinder,sample) to a
(sample,cylinder) scheme. I.e. previously there were one cylinder
block for each sample, then one sample block for one cylinder.
However, after populating the samples, the array size was reduced
to the actual number of used samples. With the new layout this
breaks indexing. Therefore, restore the old layout.
Fixes#2887
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
We did something really horribly wrong when merging cylinders. It's
been broken since commit 7c9f46a ("Core: remove MAX_CYLINDERS
restriction"), and used some really strange logic.
This rewrites the logic to be (I think) a bit more easy to understand.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
This updates libdivecomputer to support the Oceans S1 and McLean Extreme
divecomputers.
It also adds the Oceans S1 to the list of dive computers we reconize by
bluetooth name.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
- use hidapi grantlee and mdbtools from MXE
- update MXE version to use QT 5.15, and pull in libzstd and CMake 3.17.3
- fix linking of winmm on windows build with new mxe
- add some instructions on building the container
- add some new dependancies from QT 5.15 to the packaging
- add a patch to MXE to Build qtconnectivity with native-win32-bluetooth
[Dirk Hohndel: small refactor]
Signed-off-by: Paul Buxton <paulbuxton.mail@googlemail.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Having the full list of all members in the exact order should be enough to get
g++ to accept the named initializers.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
The last time those changed, we forgot to update serial_ftdi. In that change
set_latency had been removed from libdivecomputer and poll and ioctl had been
added. This caused the callbacks to no longer be aligned correctly and the
functions were called with the wrong arguments through the wrong function
pointers, leading to crashes.
Instead of the fragile assumptions about order and type of function pointers,
use named initializers. And while we are at it, fix that for the bluetooth
implementation as well.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
For unknown reasons, the dive site and trip to be parsed into were
passed as pointers to pointers. A simple pointer seems to be enough,
since the object is not allocated by the function.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
This seems like the easiest way to show the state without disrupting the UI
elsewhere. Directly below the email address used for cloud storage.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
This allows the mobile UI to reflect the three states that the dive list can be
in:
- changes that haven't been written to local storage
- there potentially are changes in local storage that were not synced with the cloud
- dive list is in sync with cloud storage
The last state could be misleading if the user access the cloud from a
different device and makes changes to the cloud storage from there, but from
the point of view of this device, the states are consistent.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
If we haven't connected at all to the cloud server we assume that there are
local changes. And whenever we save changes only locally, we also set that
flag.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
This may seem like a bit heavy handed as it adds more global state, but given
the number of ways in which attempts to sync with the cloud can fail it seems
much more reliable to claim success in the spots where we actually know that we
have successfully synced with the remote server. Transporting that information
back through the various call chains turned out to be very disruptive and ugly,
so I went with global state instead.
Whenever we access cloud storage (or any git repo), we always first check if it
actually is a git repo by calling is_git_repository() - so this is the perfect
spot to initialize the variable to false.
And there are only two spots where we either clone the remote repo
(create_local_repo()) or update the remote with the (potentially merged) local
changes (check_remote_status()). So those are the two places where we set the
variable to true.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
In many cases we did not log the issues the code ran into to stderr which made
remote debugging user problems much harder. This hopefully will help with that.
Since I was looking at the code, I also made the existing messages more
consistent.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Apparently libdivecomputer can return DC_GASMIX_UNKNOWN when
fetching tank info with
dc_parser_get_field(parser, DC_FIELD_TANK, i, &tank);
This caused emission of a warning, which was annoying users.
Disable the warning in that case.
Fixes#2866
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Having a lot of tags (or more precisely, a tags string that is very long) could
cause the width of the dive details view to extend past the width of the the
page. The txtTags label was missing a maximum width, and to make the result
more useful, I also added correct wrapping and elide to the mix (stupidly, we
had the wrap and width for the fixed name of the field ('Tags'), but not for
the user determined content of that field).
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
The UI is ugly, and of course this is hidden in the developer options that have
to first be enabled in the advanced settings. As I mentioned in the previous
commit, I believe the actual risk that something gets damaged here is very low,
but still, explaining this so it makes sense to the casual user may be a bit...
difficult.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
If you pass in the repo name, it isn't treated as cloud storage, but simply as
local git storage and imported (i.e., added) to the current dive list.
This allows the user to work around failed no-cloud->cloud transitions, merge
different accounts, and most importantly deal with situations were conflicts on
the server caused us to move a cache out of the way and potentially make dives
that were on the mobile device inaccessible to the user.
Once a UI is added, this allows the user to recover those dives (realistically
this is not really all that potentially 'dangerous' to do, but it's definitely
something that would best be done after talking to someone who understands the
cloud storage and can guide the user...).
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Every time I edit main.qml, QtCreator fixes this for me. And then I filter it
from the commit to not mix white space and actual changes. So let's just get
this fixed and move on.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
When the dive that is used to generate the header, we need to print
empty cylinders for the dives that had less cylinders than the one used
for header printing.
Signed-off-by: Miika Turkia <miika.turkia@gmail.com>
With XSLT we have to be specific if we want to apply cylinder iteration
operation to one dive only. Thus using the last dive as reference when
printing cylinder header.
Signed-off-by: Miika Turkia <miika.turkia@gmail.com>
This is a partial revert of commit 99438121c4 ("mobile/dive-edit: use template
components and theme colors")
Clearly the information given in the Qt documentation on how to theme ComboBox
is flat out broken. The trade-off between 'better dark theme' and 'broken user
experience' is fairly easy to make.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
This function was globalized in be462ae1a6 to be used for the calender
widget, but that never came to be. Therefore, for now unglobalize it
until it is needed. That said, there probably is a helper function to
turn pictures into gray-scale.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
