You wight as well compile two binaries, one that uses ALSA, one that uses JACK. Let us know what comes out so we can approve the package. > I think we can try to ask FPC to make the guidelines for this situation more > I asked FE-Legal, quoting the link you gave above, and they told me to list So I think we are safe with having it built into vmpk.
They just distribute the source code to be included in other software (as is the case with universalchardet library). Unlike rtaudio, rtmidi does not provide any libraries (static or dynamic). Another example was the java DAW frinika. We had to package them one by one, and upstream all the clementine's patches before we released clementine under Fedora. For instance, clementine was bundling libraries Yes, but the common convention is to get the library first into Fedora, then link to the library dynamically or statically. > Fedora, we can consider it just part of this very work.
> copy of a library that *exists* on a system.' Since RtMidi doesn't exist in The guide line says 'A package should not include or build against a local Do you have an argument why compiling with jack support will be bad jack is pretty much the standard sound server in Linux audio production > jack-supporting programs, and this will allow "vmpk" to communicate with all of > Compiling with jack support is important, since we have a large collection of DRTMIDI_DRIVER=JACK -DPROGRAM_NAME=vmpk-jack
Please see the README and CMakeLists.txt files. > to build another binary with jack support, you can even call the second binary > After you build the "vmpk" binary, you can pass some additional flag to cmake I think we can try to ask FPC to make the guidelines for this situation more explicit and detailed. > I asked FE-Legal, quoting the link you gave above, and they told me to list all
> all the licenses separately in the license tag of a package that all code was > thinking exactly the same way you are, and a package reviewer asked me to list > I know about that guideline, which contradicts what FE-Legal says. I will ask RtMidi upstream for this issue later. We should not choose a shared library name for RtMidi upstream, who doesn't intend to make it a shared library. The guide line says 'A package should not include or build against a local copy of a library that *exists* on a system.' Since RtMidi doesn't exist in Fedora, we can consider it just part of this very work.Ģ. I think we are not necessary to talk about the issue of this library here:ġ. > not sure if bundling is the correct solution. > 2- Even if it should remain static, should we not package it separately? I am > 1- It can be made dynamically linkable. Do you have an argument why compiling with jack support will be bad for Fedora users? jack is pretty much the standard sound server in Linux audio production software. DRTMIDI_DRIVER=JACK -DPROGRAM_NAME=vmpk-jackĬompiling with jack support is important, since we have a large collection of jack-supporting programs, and this will allow "vmpk" to communicate with all of them. > server used in audio production type applications.Īfter you build the "vmpk" binary, you can pass some additional flag to cmake to build another binary with jack support, you can even call the second binary vmpk-jack. > ? Should we build this package with jack support? Jack is the most common sound I asked FE-Legal, quoting the link you gave above, and they told me to list all the licenses separately: I was thinking exactly the same way you are, and a package reviewer asked me to list all the licenses separately in the license tag of a package that all code was compiled into a single binary. I know about that guideline, which contradicts what FE-Legal says. >, since MIT is compatible with GPL, so only GPL is needed listed. Depending on the unbundling situation we might need to add MIT to the > * The license tag as GPLv3+ is correct, except the bundled rtmidi library is It should better be asked upstream.Ģ- Even if it should remain static, should we not package it separately? I am not sure if bundling is the correct solution. But this might conflict with upstream's intentions. > rtmidi library is still not dynamically linkable.ġ- It can be made dynamically linkable. > (1-2 years ago?), the rtmidi library was not packagable, so we allowed this as However as far as I remember, the last time we checked
? The package bundles the rtmidi library.