Currently, we just skipped the notification comparison, if the
notification had a raw icon attached. This is a bit counterintuitive.
Calculating a checksum of the raw icon's data is the solution.
For that we cache the pixel buffer and introduce a field, which saves
the current icon's id. The icon_id may be a path or a hash.
So you can compare two notifications by their icon_id field regardless
of their icon type by their icon_id field.
This implies, that memory will only get freed, if the whole test method
is gone through. In case of failure, memory leaks aren't important. But
it's crucial not to fail with a segfault just for an assertion.
Fixes#580
Before, it only did test for `ASSERT_FALSE` both times. So a raw_image
wasn't guaranteed to be a condition to falsify the return.
Using the single field function, will test for both cases.
To assert static objects, we either have to add a method into the
object under test itself and recompile on a test run with the activated test
method (but also recompile after tests, so that method is gone on
release builds).
Alternatively we can include the whole .c file in our test
infrastructure and save the object in the test directory. So there's no
necessity to clean it up prior to a release build and it's contained
away.
This requires, that the test folder isn't excluded in coveralls.
To replace all occuring format strings inside the msg, replace_all had
been used previously. This leads to buggy behavior, if a format string
occurs in the replaced text, as the format string will get replaced
again under certain conditions.
Introducing a pointer, which skips the already replaced parts, will
prevent doubly replacing format strings from content.
Fixes#322
