Re: [fluid-dev] [ANNOUNCE] FluidSynth 1.1.4 released!

[Orcan Ogetbil]

On Thu, Aug 4, 2011 at 11:27 AM, Pedro Lopez-Cabanillas wrote:
> On Thursday 04 August 2011, Orcan Ogetbil wrote:
>> On Thu, Aug 4, 2011 at 3:54 AM, David Henningsson wrote:
>> > In short, version 1.1.4 of FluidSynth brings us:
>> >
>> >  * Several improvements and fixes to the CMake build system,
>> >   especially for Mac OS X
>> >  * Several bug fixes to the engine, notably quite a few which could
>> >   cause FluidSynth to sound bad in some cases.
>> >  * API additions, that make it possible to
>> >   - load MIDI files from memory
>> >   - to inspect/modify MIDI events as they are being played from a
>> >     MIDI file
>> >   - to change channels between melodic and drum mode
>> >   - and to silence all notes in one command.
>> >  * Improvements to the CoreAudio driver
>> >
>> > A more detailed list of changes is available at
>> > https://sourceforge.net/apps/trac/fluidsynth/wiki/ChangeLog1_1_4
>> >
>> > Looking back at the changes, I'm feeling a little proud today. One of the
>> > reasons for feeling proud, is that we have more than one new contributor:
>> > people who have taken the step to actually work with the code, improve it,
>> > and create non-trivial patches with new functionality in it. That is a good
>> > foundation for the future and I hope all contributors feel welcomed and
>> > properly attributed.
>> >
>> > Also, the improvements/fixes to the build system is worth a special note
>> > (thanks Pedro!), I was surprised to see so many of them! And finally, we've
>> > managed to close several long standing bugs, which makes me feel that this
>> > release will be stable and enjoyable for a lot of people, with less
>> > annoyances than ever. (Thanks to the jOrgan folks for the test run!)
>> >
>> > Now, enjoy the release, and last but not least - make sure you make some
>> > really great music with it! :-)
>> >
>> > On behalf of the FluidSynth developer team,
>> >  David Henningsson
>> >
>>
>> Hi all, thanks for releasing 1.1.4. Sorry I did not have time to test
>> the prerelease tarball you made the other day. I was planning to do
>> this today.
>>
>> Anyhow, I saw in the CMakeLists.txt, a couple lines such as
>>            ${CMAKE_INSTALL_PREFIX}/${LIB_INSTALL_DIR}${LIB_SUFFIX} )
>>
>> Unfortunately this is wrong. The variable ${LIB_INSTALL_DIR} is
>> already expected to have the lib suffix in it, i.e. it is set to
>> /usr/lib64 on multilib 64bit systems. This is the cmake standard we
>> have with hundreds of packages in Fedora.
>
> This may be the standard in Fedora, but it is not a CMake standard at all. 
> You can ask in the CMake mailing list, or search for similar questions in the 
> archives, for instance:
> http://www.cmake.org/pipermail/cmake/2011-March/043180.html
>
> I've adopted the variable name "LIB_INSTALL_DIR" because i'm used to it in my 
> KDE projects, but it is not a standard outside KDE. The FluidSynth's 
> LIB_INSTALL_DIR is a constant defined in cmake_admin/DefaultDirs.cmake, with 
> the value "lib". The other variable LIB_SUFFIX is initialized in the main 
> CMakeLists.txt with the value "64" or empty. You can override the 
> initialization value of these variables in the command line, if you want:
>
> $ cmake .. -DLIB_SUFFIX=""
>
> This can be easily added to the RPM spec file too. I find this method much 
> more comfortable than applying patches.
>

Hi Pedro,

Sure the behavior can be overridden that way. However, when building
RPMs, there are cmake macros we use that pass all the standard flags
to all packages that use cmake. I am sure other RPM based
distributions, or even DEB ones use some sort of standardization in
cmake flags too.

I really do believe that what I claimed is a  (maybe unwritten) cmake
standard, since we do not have to do such overrides you suggested for
any other package (we have a lot of them). I also adopt this
convention in my projects where I am the upstream. Of course,
fluidsynth can be a special case, but does it really have to be?

Thank you for your time,

Orcan