This results in identical XML if the session did not change, allowing
to keep track of state-changes easily (e.g. session file in git).
Thanks to deva (again).
Remove the directory recursively including itself. The function
is used in two places only:
* LV2Plugin::add_state() -- no change, remove tmp. state
* Session::save_as() -- on error, remove target
In both cases removing the folder itself is correct.
This is part of a larger fix for #6450. Previously new route groups were
created with different default properties depending on which code created
the new group. The group tab method had all properties set and that is now
going to be the default.
It looks as if the default timer resolution for applications running under wine
is different than Windows so just test that the minimum timer resolution is
below a certain amount rather than checking before and after calling
timeBeginPeriod
Iterating over a const Midi-Sequence calls Evoral::Sequence::set_event(),
which in turn used Evoral::Event::operator=() which always created
a new event-ID (create copy of the event).
Issues fixed:
- Saving *unmodified* MIDI produced new event-IDs on every save;
files changed with every save. - greetings to Deva.
- all [GUI] operations that use IDs to refer to notes e.g. undo.
invalid undo-history.
Also clarify assignment operator name. Prefer explicit assign() over =.
AFAICT this could happen if a region's end time (on the timeline) was earlier than the end time in the actual recording. This could cause a situation where the last block of detected silence would have an end time greater than the end tome for the region being processed. Strip Silence would create its new regions - but the last one it created would usually come out with a negative duration.
GStatBuf is not usable on 32 bit Windows without the redefinition in
pbd/gstdio_compat.h so add a test to check for the correct behavior of
g_stat and g_utime on all platforms now that the issue is fixed.
