Actually, I was wrong regarding the relationship with the resources framework. The empty.wav is an old and hackish way to fix the problem. Toby and I agree that the way to fix this is simply filtering out the empty.wav reference during our upgrade() routine which is always run on project-load. When confronted with an empty path, the current code already creates an empty sample, instead of loading one from disk. However.. I do agree there is a problem with hanging-references to files in triple-osc. We could add a manual "clear" option. But, perhaps what we can do is just not complain if triple-oscillator fails to load a sample for a Non-User-Wave oscillator? This isn't a problem in instruments that only use one sample (like AFP) because exactly one generator is critical. It is not a problem with plugins like bit-invader because the shape data is stored in the preset.
Also, I've been considering adding a logging-console to LMMS for a long time. We would be able to report failed resources, file i/o, audio i/o, and various messages through it and it wouldn't be as intrusive as a message box. A junior programmer could easily add this, or I'll probably eventually add it since I'm particularly bothered by message boxes. -Paul On Mon, Dec 7, 2009 at 6:27 AM, Tan Zong Xuan <[email protected]> wrote: > Perhaps. The simplest way to fix this particular problem would be to > recreate this 'empty.wav', which was present in earlier versions, so that > using newer versions of LMMS to open older LMMS files will no longer throw > an error. Alternatively, add a button labelled 'Clear' in the required > places to remove any previously defined user waveforms/soundfonts/etc . > > On Mon, Dec 7, 2009 at 3:05 AM, Paul Giblock <[email protected]> wrote: >> >> The problem is caused by the new resources stuff I bet. I wonder if the >> resources plugin itself should have a facility for loading up a default >> (usually dummy) resource per-resource-type? >> >> Empty for audio-samples, default-sf2 for soundfont, new project for >> projects, etc? >> >> Paul >> >> On Dec 5, 2009 5:25 AM, "Tan Zong Xuan" <[email protected]> wrote: >> >> It seems that it is impossible to 'undefine' user-defined waveforms for >> Triple Oscillator instruments once they have been defined (i.e. when the >> file browser comes up, you can't select nothing). This results in irritating >> error message boxes popping up when instruments from older LMMS versions are >> loaded using the development version, as the instruments sometimes have the >> user-defined waveform set to 'empty.wav', a file which is missing in the >> development version of LMMS. The only way to get rid of previously defined >> waveforms is to manually edit the uncompressed project file using a text >> editor. Could this feature be coded such that it is possible to select no >> sound file as the user-defined waveform (thereby 'undefining' it)? >> >> On another note, when using LMMS with JACK, every time LMMS loads a new >> file, JACK xruns and disconnects LMMS. LMMS automatically reconnects after >> that, but is disconnected a 2nd time. After the 2nd reconnection, everything >> is fine, but it is rather unpleasant to have 2 error messages warning you >> that JACK kicked out LMMS every time you load a new file. This is on Ubuntu >> 9.10, with LMMS grabbed from git. >> >> >> ------------------------------------------------------------------------------ >> Join us December 9, 2009 for the Red Hat Virtual Experience, >> a free event focused on virtualization and cloud computing. >> Attend in-depth sessions from your desk. Your couch. Anywhere. >> http://p.sf.net/sfu/redhat-sfdev2dev >> _______________________________________________ >> LMMS-devel mailing list >> [email protected] >> https://lists.sourceforge.net/lists/listinfo/lmms-devel >> > > ------------------------------------------------------------------------------ Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev _______________________________________________ LMMS-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/lmms-devel
