We didn't enforce a limit on tag length, but we would crash on a tag
longer than 127 chars.
This uses the xml buffer as scratch space. Don't really know if this is
fair, but it looks like it works.
Signed-off-by: Anton Lundin <glance@acc.umu.se>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
First might be garbage if get_ranges would be called when
dive_table.nr == 0. This would rather signal that something else is
broken, but at least we shouldn't make it worse.
Signed-off-by: Anton Lundin <glance@acc.umu.se>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
If realloc moved the memory, we shouldn't try to access it. realloc
copied that memory so access it via the new function instead.
Signed-off-by: Anton Lundin <glance@acc.umu.se>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
We can position our old flag in a PNG with double the size of the
original flag - 48x48px, so that the pin is exactly at the center
of the PNG.
This way when the user clicks on a map location the pin should
be placed exactly there.
Patch also makes the flag bigger.
Fixes#335.
Signed-off-by: Lubomir I. Ivanov <neolit123@gmail.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
The forward declaration of 'struct device_data_t' is not
needed as it already comes from '../libdivecomputer.h'.
Reported by the program cppcheck.
Signed-off-by: Lubomir I. Ivanov <neolit123@gmail.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
next_segment():
The iterator limit check (i < size - 1) should precede
the indexing (buf[i]).
Reported by the program cppcheck.
Signed-off-by: Lubomir I. Ivanov <neolit123@gmail.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
ProfileGraphicsView::plot_depth_profile():
The iterator limit check (i < 7) should precede the indexing
(increments[i]).
Reported by the program cppcheck.
Signed-off-by: Lubomir I. Ivanov <neolit123@gmail.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
calculate_cns():
The check if prev_dive is NULL should be before calling
prev_dive->when.
Reported by the program cppcheck.
Signed-off-by: Lubomir I. Ivanov <neolit123@gmail.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Italics marker needed to come after the 's to work properly
Signed-off-by: Tim Wootton <tim@tee-jay.demon.co.uk>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Found the word "ter" below this para, assume it's not
supposed to be there.
Signed-off-by: Tim Wootton <tim@tee-jay.demon.co.uk>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Introduce some harness in ProfileGraphicsView::plot_one_event, so we
detect bad stuff and bail, instead of dereferencing undef pointers.
Signed-off-by: Anton Lundin <glance@acc.umu.se>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
If first sample is not a DC_SAMPLE_TIME, we would have bin dereferencing
a null pointer.
This might actually never happen, unless we talk to a really weird dc,
but this makes the static analyzer happier.
Signed-off-by: Anton Lundin <glance@acc.umu.se>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
try_to_xslt_open_csv will read the file, so we don't need to do it
before that and leak that memory.
Signed-off-by: Anton Lundin <glance@acc.umu.se>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
More to get clang-scan to quiet down that for the unlikely event that
unsigned int and int is different sizes.
Signed-off-by: Anton Lundin <glance@acc.umu.se>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
last_pressure is updated a couple of lines later, so no need to do it
here.
Signed-off-by: Anton Lundin <glance@acc.umu.se>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
prepare_dives_for_divelogs() did a silly thing, which I was
responsible for. When populating 'tempfile' we benefit
from QString, but then return a pointer to a local variable
(char *) without alocating it on the heap. This resulted
in undefined behavior, as we don't know the lifespan of that
local memory on the stack.
Patch fixes that by using strdup() and freeing the memory
when/if needed.
Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Lubomir I. Ivanov <neolit123@gmail.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Amazing progress. Bulgarian, Chinese (Taiwan), Slovak got completed - many
other translations added the hand full of new strings that showed up over
the last few days.
Thanks to all the translators.
I only pull these, authorship is tracked at Transifex.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Replaced graphics for Import from Divelogs.de to reflect latest UI schanges.
Signed-off-by: Willem Ferguson <willemferguson@zoology.up.ac.za>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Both JS and CSS resources need to be embedded into HTML
Signed-off-by: Sergey Starosek <sergey.starosek@gmail.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Some code in plan() left from the gtk days introduced a safety stop in the
plan. It created a un-editable diveplanpoint.
Fixes#349
Signed-off-by: Anton Lundin <glance@acc.umu.se>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
The notion of current path changes as we open files in the file system.
What we really want is the directory from where Subsurface was started.
That covers both the case of Windows and running Subsurface from the
install directory.
This worked before because all support files were opened before the first
user interaction. But opening the manual showed the flaw in the previous
logic.
Fixes#348
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
When trying to remember a newly manually created dive across the
sort_table() call I abused the dive->selected flag and didn't clear it out
afterwards (most likely thinking "hey, I'm calling selectDive() on this
anyway").
This caused the UI code to correctly mark the dive as selected, but when
it called down to the C code to keep the internal data structures in sync,
that code didn't update amount_selected as it thought this dive had
already been selected. And that caused other parts of the code to get
confused, which manifested for example in the failed upload to
divelogs.de.
So this clears out the flag after abusing it to remember a dive. Maybe we
should add a 'remembered' flag instead, but for now this fixes the issue.
Fixes#351
Initial-fix-by: Lubomir I. Ivanov <neolit123@gmail.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
It gets lost in the noise with all the line number changes, but this
introduces only a single new string.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
We use the same dialog for both the Subsurface WS and for the
divelogs.de one. Instead of the old title (which suggests download)
we can use a more generic one - "Webservice Connection"
This makes it easy for us to maintain a dialog that can both
upload and download data, while the dialog title is descriptive
enough.
Fixes#352
Signed-off-by: Lubomir I. Ivanov <neolit123@gmail.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
If rowCount() is 0 we get an assert:
ASSERT: "last >= first" in file kernel/qabstractitemmodel.cpp...
To solve that we wrap the beginRemoveRows() call in a bnrach:
if (rowCount() > 0) {
...
Signed-off-by: Lubomir I. Ivanov <neolit123@gmail.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Clang yelled about it, and it looks prettier. It also felt kinda strange
to explicit initialize the first element to zero and the rest to zero
implicit.
Signed-off-by: Anton Lundin <glance@acc.umu.se>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Pedantic Tim says "Barring waterproof device, most of us won't
be entering these fields during the dive." ;-)
Signed-off-by: Tim Wootton <tim@tee-jay.demon.co.uk>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Initialize diveplan.dp to NULL, so we know that we will bail in
drawProfile, when the initial settingsChanged is fired.
Signed-off-by: Anton Lundin <glance@acc.umu.se>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
