Missing PortMixer in 1.3.7-beta
|
|
|
||||||||
|
|
|
||||||||
|
|
On Sat, 28 Feb 2009 22:42:49 +0100, Michael wrote:
> On Sat, 28 Feb 2009 21:52:28 +0100, Michael wrote: > > > Is the PortMixer combo-box visible for anyone else in Audacity 1.3.7-beta > > with Linux? > > > > Audio Device Info debug dialog says "Unable to open Portmixer", too. > > The new AudioIO code is responsible for this. One must enter and save the > Preferences a first time, or else the AudioIO code tries to come up > with defaults that result in a -1 card index later on. for the Playback/Recording devices, *if* the config file does not define the "PlaybackDevice" and "RecordingDevice" values yet. The dialog defaults to displaying the names for I/O device #0 which doesn't match the internal defaults. Attached is one patch to fix this. ------------------------------------------------------------------------------ Open Source Business Conference (OSBC), March 24-25, 2009, San Francisco, CA -OSBC tackles the biggest issue in open source: Open Sourcing the Enterprise -Strategies to boost innovation and cut costs with open source participation -Receive a $600 discount off the registration fee with the source code: SFAD http://p.sf.net/sfu/XcvMzF8H _______________________________________________ audacity-devel mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/audacity-devel |
||||||||
audacity-1.3.7-audiodevdefaults.patch (1K)
|
|
||||||||
|
|
||||||||
|
|
In reply to this post
by Markus Meyer
|
||||||||
|
|
|
||||||||
|
|
On Mon, 02 Mar 2009 12:26:53 +0100, Markus wrote:
> I would have expected the problem appears on all platforms but there > must be something different going on on Linux regarding the device > detection. Maybe. If the default I/O device index on Windows is 0, the problem won't be visible. Only if the default device index is non-zero, you should be able to reproduce it. Audacity >= 1.3.7-beta doesn't save the default preferences. However, the GUI implementation retrieves configuration values from the preferences backend (gPrefs). This creates a conflict between actual defaults and what the GUI displays as defaults in case of missing/unset preferences. E.g. the Audio I/O dialog (prefs/AudioIOPrefs.cpp -> ShuttleGUI.cpp) relies on /AudioIO/PlaybackDevice and /AudioIO/RecordingDevice being set, or else it falls back to its own default value (zero or empty string) that differs from the defaults in the Audio I/O frontend code (AudioIO.cpp). > > The real fix would make the Audio I/O Prefs retrieve the device name > > for the default device numbers instead of saving the defaults. > > Thanks for the step-by-step directions to reproduce. I'm afraid I > currently cannot reproduce the problem because I do not have a working > Linux install right now. Anyone else? > > As you say you have an idea how a "real fix" would look like, do you > think you can make up a patch for that? That would be great! I've attached a different patch, which doesn't write/save the defaults. Instead, it follows the footsteps of the current AudioIOPrefs.cpp code, which queries PortAudio to learn about devices. It's a question of design. Could still be improved by moving much more into the AudioIO.cpp wrapper instead of querying PortAudio directly. 1) It's good that defaults are _not_ saved to the cfg file. Fully agreed here. That way defaults can continue to change automatically at run-time (depending on hardware changes and/or software updates) till the user enters the Preferences dialogue and confirms/saves the settings explicitly in order to override defaults. 2) It's not good that the GUI relies on preferences being set, and if not being set, it falls back to its own defaults. As in the case of AudioIOPrefs.cpp where this fails. ShuttleGUI.cpp does not care about the actual default audio I/O device numbers as determined in AudioIO.cpp. As outlined above, it retrieves the device names from the Preferences (gPrefs). If the Preferences don't set the rec/play device, either ShuttleGUI or the AudioIOPrefs code must learn about the actual defaults as determined by AudioIO.cpp. That's what the new patch implements, except that it AudioIOPrefs queries PortAudio instead of AudioIO (which is private for most of its useful methods). ------------------------------------------------------------------------------ Open Source Business Conference (OSBC), March 24-25, 2009, San Francisco, CA -OSBC tackles the biggest issue in open source: Open Sourcing the Enterprise -Strategies to boost innovation and cut costs with open source participation -Receive a $600 discount off the registration fee with the source code: SFAD http://p.sf.net/sfu/XcvMzF8H _______________________________________________ audacity-devel mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/audacity-devel |
||||||||
|
|
|
||||||||
|
|
||||||||
|
|
|
||||||||
|
|
|
||||||||
|
In reply to this post
by Richard Ash
|
||||||||
|
|
|
||||||||
|
|
|
||||||||
|
|
||||||||
| Free Embeddable Forum Powered by Nabble | Help |
