If a function that returns an allocated pointer is passed directly into
something that does not take ownership of the pointer, the allocation is
leaked. This can be fixed by assigning the pointer to a new variable
and freeing it after operation on the memory.
At least one user has managed to get it set on their book so even
though it was supposed to be unimplemented it got through somehow.
Restoring it allows books with it set to load.
- Renamed option to "Account Balance" to avoid confusion with running
total.
- Added helper function to ensure running balance and balance forward
are only shown when transaction are grouped by account and sorted as
in register. In that case column heading remains "Running Balance"
and balance forward is shown. Otherwwise column heading is renamed
"Account Balance" and balance forward is not shown.
- Also added missing code for Common Currency conversion.
Ensure that dialog resources stored in options are freed when the
dialog is destroyed.
The crash happened when a new dialog replaced the old one on the options
and the old one's destructors tried to access a dangling reference to
a GtkWidget.
Edit/Report Options
As a result of auto instead of auto& in a std::visit call which
resulted in a temporary option whose GncOwner* was destroyed before
we could use it.
This fixes memory leaks that are only present in testing code.
Not very useful on itself, but it does make it easier to fix memory
leaks and other AddressSanitizer problems in actual gnucash code later.
These were not used outside a test.
And that test was not leak free, as a result of the functions not doing
what they are supposed to do when the current value is not of the type
that is expected. (NULL is returned, but the value is not replaced)
It's better to use the default std::pair implementation
- spotted by clang-tidy bugprone-suspicious-string-compare
- will always return false, because type is the same
Also,
- remove redundant "using GncItem ..."
- spotted by clang-tidy bugprone
- probably no visible effects because of uniqueness, and other checks
libgnucash/engine/gnc-optiondb.cpp:149:5: warning: this call will remove at most one item even when multiple items should be removed [bugprone-inaccurate-erase]
m_options.erase(std::remove_if(m_options.begin(), m_options.end(),
^
libgnucash/engine/gnc-optiondb.cpp:358:5: warning: the value returned by this function should be used [bugprone-unused-return-value]
std::remove_if(m_callbacks.begin(), m_callbacks.end(),
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
libgnucash/engine/gnc-optiondb.cpp:358:5: note: cast the expression to void to silence this warning
for certain reports.
Those reports being ones using complex options, apparently because
the callbacks weren't protected from Guile's garbage collector.
So replace the anyway ugly hack of a void* with a std::any wrapping
a class holding a std::unique_ptr with a custom deleter. The
constructor calls scm_gc_protect_object on the SCM containing the
callback and the custom deleter calls scm_gc_unprotect_object. The
copy constructor, required for std::any, makes a new std::unique_ptr
and calls scm_gc_protect_object again ensuring that the protect and
unprotect calls are symmetrical.
Meanwhile std::any hides the Guile dependency from all the classes
that don't need to know about it. The only ugliness is that there's
no good place to put a common implementation of SCNCallbackWrapper so it's
repeated in gnc-optiondb.i and dialog-options.cpp.
The code is was only used by the csv import assistant and forced
gtk as a dependency on gnc-imp-props-tx
Part has been moved to Account, other bits have
been moved to the importer code where they were
used.
By deleting the function and using GncOptionDbImpl::lookup_string_option
directly. It returns a string that we don't have to worry about
memory-managing.
Also create a new GncOptionDbImpl::set_string_option to replace
gnc_option_db_set_string_value.
reverse transaction.
The proximate cause was that xaccTransBeginEdit put the KVP cache
variable of the new transaction in a state that prevented the value
from being copied. More generally the KVP cache variables didn't
handle any invalidating events.
With the change to GValue usage in qof_instance_get_kvp it's now
a simple memory dereference with no copying except for POD types
so caching is no longer useful. This commit removes caching from
Transaction, eliminating the notes problem.
John Ralls
Daniel Johnson
Maarten Bosmans
Christopher Lam
Muhammadyusuf
Vincent Dawans
Robert Fewell
Richard Cohen
Geert Janssens
Richard Cohen