build for mac os x
|
|
| 1 2 3 |
lasconic
|
Reading the Cmakefile a little bit more, resfile.o seems to be a file generated for windows only with wine utilities. Should not be needed for macOS. So I guess you can comment the line : set (resource_file ${PROJECT_BINARY_DIR}/resfile.o)
Keep us posted :) Lasconic 2008/12/7 Lasconic <[hidden email]> I think that the .o file will be generated on compilation so in the next step... You can try to remove the APPLE flag in the IF ELSE ENDIF statement I guess. ------------------------------------------------------------------------------ SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas, Nevada. The future of the web can't happen without you. Join us at MIX09 to help pave the way to the Next Web now. Learn more and register at http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix.com/ _______________________________________________ Mscore-developer mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/mscore-developer |
Frye
|
In reply to this post
by lasconic
If I just remove the line with resfile.o I can get the compilation to run quite far, but I do not have the slightest clue what may be missing because of that. Could someone tell me a bit of that resfile.o. When it is generated and what is is used for?
|
||||||||||||||||
Johan Pauwels-2
|
In reply to this post
by lasconic
Some javascript/style in this post has been disabled (why?)
Hi all,I haven't given up yet, just being a bit slow with the write-up. I'm getting closer, but still not there. On 30 Nov 2008, at 19:03, Lasconic wrote:
You generate those files using "cmake -G <generator name>" or via cmake-gui, more info can be found in the usual spots. That's how I made an Xcode project, which has already greatly helped me debugging the build process.
******************* Previous revision ******************* One of the problems indeed seemed that the moc preprocessor didn't seem to be called. I was able to get rid of half the linking errors by replacing "QT4_WRAP_CPP" with "QT_WRAP_CPP", which does not make a lot of sense to me, but seems to work. Also I found in CMakeLists.txt in trunk/mscore/mscore the "QT4_WRAP_CPP" appearing twice for "mocs" and "mocs2", which just seemed to duplicate functionality. The result for "libawlplugin.dylib" were 69 link errors (coming from twice that amount), mostly unrelated to moc-processing, I believe, but I have no idea what causes them. A random sample: "QListData::shared_null", referenced from: __ZN9QListData11shared_nullE$non_lazy_ptr in awlplugin.o "QFrame::QFrame(QWidget*, QFlags<Qt::WindowType>)", referenced from: Awl::ColorLabel::ColorLabel(QWidget*)in awlplugin.o "QPainter::QPainter(QPaintDevice*)", referenced from: Awl::Knob::paint(QRect const&)in knob.o Awl::ColorLabel::paintEvent(QPaintEvent*) in colorlabel.o Anyway, this was all just specific for the code checked out last week, so let's not waste time and concentrate on the conclusions that are still valid for the build process of the latest revision. **************** Latest revision **************** Apparently the "libawlplugin.dylib" target dissapeared from the current release (checked out today), so with it all the linking errors. Consequently, the "QT4_WRAP_CPP" function also didn't give any problems. After applying the changes below, only the "clock_gettime" not-portable-issue which I reported in a previous mail remained, hopefully it's the last deal-breaker. Necessary changes: select.cpp: 47 replace "#include values.h" by "limits.h" (as in "all.h", as said in previous mail) mscore/mscore/CMakeLists.txt: 95 I see you added a check for APPLE to if (APPLE OR MINGW) set (resource_file ${PROJECT_BINARY_DIR}/resfile.o) set (AUDIO pa.cpp pm.cpp) but resfile.o cannot be found on my machine, so I changed the if-statement back to "if(MINGW)". I have no idea what this line is supposed to do. mscore/CMakeLists.txt: 289 & 295 change lupdate and lrelease to ${QT_LUPDATE_EXECUTABLE} and ${QT_LRELEASE_EXECUTABLE} respectively. Now the lupdate and lrelease target can be build on my machine too (unfortunately, this doesn't bring me any closer to building the main program) Addition to previous comment: You just commented the "#include <QtGui/QX11Info>" statement, so I guess it has no drawbacks on building the Linux version either, but in the meantime I found out that the proper way of checking for its presence is including it like this: #ifdef Q_WS_X11 #include <QtGui/QX11Info> #endif A portable version of that functionality can be found in QtGui/QDesktopWidget http://doc.trolltech.com/4.4/qdesktopwidget.html, but if its not necessary anyway, all the better of course. Ok, now that I'm finished writing, I just see some new input about the Mac build by Joachim Ganseman, which I haven't read yet, so all my comments might by irrelevant, but I just wanted to share them anyway. He's a much better Mac developer than I am, so I guess all problems will be solved soon. Johan ------------------------------------------------------------------------------ SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas, Nevada. The future of the web can't happen without you. Join us at MIX09 to help pave the way to the Next Web now. Learn more and register at http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix.com/ _______________________________________________ Mscore-developer mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/mscore-developer |
||||||||||||||||
|
Frye
|
In reply to this post
by lasconic
Ok,
I am able to get it to compile, but linking is giving me the next problems. There were few changes I had to make, but nothing big. However I still cannot get anything out of the build process I could run. Linker is giving the following output: Undefined symbols: "_Pm_GetDefaultInputDeviceID", referenced from: PortMidiDriver::init() in pm.o "_Pm_SetFilter", referenced from: PortMidiDriver::init() in pm.o "_Pm_OpenInput", referenced from: PortMidiDriver::init() in pm.o "_Pt_Time", referenced from: _Pt_Time$non_lazy_ptr in pm.o "_Pt_Stop", referenced from: PortMidiDriver::~PortMidiDriver()in pm.o PortMidiDriver::~PortMidiDriver()in pm.o PortMidiDriver::~PortMidiDriver()in pm.o PortMidiDriver::init() in pm.o "_Pm_GetDefaultOutputDeviceID", referenced from: PortMidiDriver::init() in pm.o "_Pm_Read", referenced from: PortMidiDriver::init() in pm.o PortMidiDriver::read() in pm.o "_Pm_Poll", referenced from: PortMidiDriver::init() in pm.o PortMidiDriver::read() in pm.o "_Pm_GetErrorText", referenced from: PortMidiDriver::init() in pm.o "_Pt_Start", referenced from: PortMidiDriver::init() in pm.o "_Pm_Close", referenced from: PortMidiDriver::~PortMidiDriver()in pm.o PortMidiDriver::~PortMidiDriver()in pm.o PortMidiDriver::~PortMidiDriver()in pm.o ld: symbol(s) not found collect2: ld returned 1 exit status make[2]: *** [mscore/mscore] Error 1 make[1]: *** [mscore/CMakeFiles/mscore.dir/all] Error 2 make: *** [all] Error 2 Any ideas? Portaudio is installed, but I had to manually specify it's header location to get the build process to find portaudio.h so it may be that there is something else wrong in my setup. -Samuli
|
||||||||||||||||
Johan Pauwels
|
In reply to this post
by Johan Pauwels-2
> > mscore/CMakeLists.txt: 289 & 295 > change lupdate and lrelease to ${QT_LUPDATE_EXECUTABLE} and $ > {QT_LRELEASE_EXECUTABLE} respectively. Now the lupdate and lrelease > target can be build on my machine too (unfortunately, this doesn't > bring me any closer to building the main program) > > > Ok, now that I'm finished writing, I just see some new input about > the Mac build by Joachim Ganseman, which I haven't read yet, so all > my comments might by irrelevant, but I just wanted to share them > anyway. He's a much better Mac developer than I am, so I guess all > problems will be solved soon. Hey, I just wanted to say that I now read the posts on the mailinglist of last week (which I should have done before posting) and that I'm a bit stupid for not having it done sooner, because most of my remarks are in Joachim's post too. I had no idea so many new tries would happen in a week (and I even know Joachim). He also provided the solution for my the clock_gettime issue, which seems to be present after all. So the only really valid remark in my mail is the one I cut out, about the lupdate/lrelease. Now I'm stuck at the same point as Frye, portaudio seems to give problems. Johan ------------------------------------------------------------------------------ SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas, Nevada. The future of the web can't happen without you. Join us at MIX09 to help pave the way to the Next Web now. Learn more and register at http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix.com/ _______________________________________________ Mscore-developer mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/mscore-developer |
||||||||||||||||
|
lasconic
|
I'm not sure if it will help but the APPLE you remove in the if (APPLE OR MINGW)
means that the following line is not included right ? maybe it should ? set (AUDIO pa.cpp pm.cpp) Also, for linking for windows, portaudio.dll needs to be in a directory specified by the CmakeList file. For linux, the line 173: PKGCONFIG (portaudio-2.0 19 PORTAUDIO_INCDIR PORAUDIO_LIBDIR PORTAUDIO_LIB PORTAUDIO_CPP ) Seems to find portaudio lib. Maybe you should change the way portaudio lib is detected ? or hardcoded it just like for windows ? Again not sure it helps :) but it's great some people try :) lasconic 2008/12/7 Johan Pauwels <[hidden email]>
------------------------------------------------------------------------------ SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas, Nevada. The future of the web can't happen without you. Join us at MIX09 to help pave the way to the Next Web now. Learn more and register at http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix.com/ _______________________________________________ Mscore-developer mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/mscore-developer |
||||||||||||||||
Toby Smithe-2
|
In reply to this post
by Frye
On Sun, Dec 7, 2008 at 5:21 PM, Frye <[hidden email]> wrote:
> > Ok, > > I am able to get it to compile, but linking is giving me the next problems. > There were few changes I had to make, but nothing big. However I still > cannot get anything out of the build process I could run. I'd just like to quickly point out that those errors are issues with linking against portmidi not portaudio, so check the availability of that library, too. ------------------------------------------------------------------------------ SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas, Nevada. The future of the web can't happen without you. Join us at MIX09 to help pave the way to the Next Web now. Learn more and register at http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix.com/ _______________________________________________ Mscore-developer mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/mscore-developer |
||||||||||||||||
|
Johan Pauwels-2
|
In reply to this post
by lasconic
On 07 Dec 2008, at 19:10, Lasconic wrote: > I'm not sure if it will help but the APPLE you remove in the if > (APPLE OR MINGW) > means that the following line is not included right ? maybe it > should ? > set (AUDIO pa.cpp pm.cpp) You're absolutely right, my mistake that I didn't report it, but I did that, otherwise it wouldn't even compile. So to be clear, that particular line now looks like this with me, which of course breaks the windows build. if (APPLE OR MINGW) #set (resource_file ${PROJECT_BINARY_DIR}/resfile.o) set (AUDIO pa.cpp pm.cpp) else (APPLE OR MINGW) > Also, for linking for windows, portaudio.dll needs to be in a > directory specified by the CmakeList file. > For linux, the line 173: PKGCONFIG (portaudio-2.0 19 > PORTAUDIO_INCDIR PORAUDIO_LIBDIR PORTAUDIO_LIB PORTAUDIO_CPP ) > Seems to find portaudio lib. Maybe you should change the way > portaudio lib is detected ? or hardcoded it just like for windows ? I'm quite sure it's something like that, which shouldn't be too hard to solve, but I just stopped trying after I reached the same point as Samuli/Frye. Don't know if I'll have time to look at it again before next weekend, but there are at least 3 pair of eyes looking at it now. Johan ------------------------------------------------------------------------------ SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas, Nevada. The future of the web can't happen without you. Join us at MIX09 to help pave the way to the Next Web now. Learn more and register at http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix.com/ _______________________________________________ Mscore-developer mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/mscore-developer |
||||||||||||||||
|
Johan Pauwels-2
|
In reply to this post
by Frye
Some javascript/style in this post has been disabled (why?)
On 07 Dec 2008, at 18:21, Frye wrote:
I got a tiny bit further, but not much. Now I have linking errors not against portmidi, but something in the inwards of portmidi as far as I understand. What I've done is (all settings in Xcode) In target "mscore" info > build > Search paths Library Search Path set to /opt/local/lib (just as you add /opt/local/include to Header Search Path to be able to compile) In target "mscore" info > build > Linking add -lportmidi to Other Linker Flags (-lportaudio is added by CMake, but -lportmidi not, this should probably be done by CMake too) The resulting linking errors are: Undefined symbols: "_kMIDIPropertyConnectionUniqueID", referenced from: _kMIDIPropertyConnectionUniqueID$non_lazy_ptr in libportmidi.a(pmmacosxcm.o) "_MIDIGetNumberOfSources", referenced from: _pm_macosxcm_init in libportmidi.a(pmmacosxcm.o) "_MIDIPacketListAdd", referenced from: _send_packet in libportmidi.a(pmmacosxcm.o) _send_packet in libportmidi.a(pmmacosxcm.o) "_MIDIObjectGetStringProperty", referenced from: _EndpointName in libportmidi.a(pmmacosxcm.o) _EndpointName in libportmidi.a(pmmacosxcm.o) _EndpointName in libportmidi.a(pmmacosxcm.o) _ConnectedEndpointName in libportmidi.a(pmmacosxcm.o) "_kMIDIPropertyName", referenced from: _kMIDIPropertyName$non_lazy_ptr in libportmidi.a(pmmacosxcm.o) "_MIDIPacketListInit", referenced from: _send_packet in libportmidi.a(pmmacosxcm.o) _midi_write_short in libportmidi.a(pmmacosxcm.o) _midi_begin_sysex in libportmidi.a(pmmacosxcm.o) "_MIDIInputPortCreate", referenced from: _pm_macosxcm_init in libportmidi.a(pmmacosxcm.o) "_MIDIGetNumberOfDevices", referenced from: _pm_macosxcm_init in libportmidi.a(pmmacosxcm.o) "_MIDIPortConnectSource", referenced from: _midi_in_open in libportmidi.a(pmmacosxcm.o) "_MIDIOutputPortCreate", referenced from: _pm_macosxcm_init in libportmidi.a(pmmacosxcm.o) "_Pt_Time", referenced from: _Pt_Time$non_lazy_ptr in pm.o _Pt_Time$non_lazy_ptr in libportmidi.a(portmidi.o) _Pt_Time$non_lazy_ptr in libportmidi.a(pmmacosxcm.o) "_Pt_Stop", referenced from: PortMidiDriver::~PortMidiDriver()in pm.o PortMidiDriver::init() in pm.o PortMidiDriver::~PortMidiDriver()in pm.o PortMidiDriver::~PortMidiDriver()in pm.o "_MIDIEntityGetDevice", referenced from: _EndpointName in libportmidi.a(pmmacosxcm.o) "_MIDIPortDisconnectSource", referenced from: _midi_in_close in libportmidi.a(pmmacosxcm.o) "_MIDIGetDestination", referenced from: _pm_macosxcm_init in libportmidi.a(pmmacosxcm.o) "_MIDIClientCreate", referenced from: _pm_macosxcm_init in libportmidi.a(pmmacosxcm.o) "_MIDIObjectFindByUniqueID", referenced from: _ConnectedEndpointName in libportmidi.a(pmmacosxcm.o) "_MIDIEndpointGetEntity", referenced from: _EndpointName in libportmidi.a(pmmacosxcm.o) "_MIDIGetNumberOfDestinations", referenced from: _pm_macosxcm_init in libportmidi.a(pmmacosxcm.o) "_MIDIDeviceGetNumberOfEntities", referenced from: _EndpointName in libportmidi.a(pmmacosxcm.o) "_MIDIPortDispose", referenced from: _pm_macosxcm_term in libportmidi.a(pmmacosxcm.o) _pm_macosxcm_term in libportmidi.a(pmmacosxcm.o) "_Pt_Start", referenced from: PortMidiDriver::init() in pm.o _Pm_OpenOutput in libportmidi.a(portmidi.o) _midi_in_open in libportmidi.a(pmmacosxcm.o) "_MIDIClientDispose", referenced from: _pm_macosxcm_term in libportmidi.a(pmmacosxcm.o) "_MIDIGetSource", referenced from: _pm_macosxcm_init in libportmidi.a(pmmacosxcm.o) "_Pt_Started", referenced from: _Pm_OpenOutput in libportmidi.a(portmidi.o) _midi_in_open in libportmidi.a(pmmacosxcm.o) "_MIDIObjectGetDataProperty", referenced from: _ConnectedEndpointName in libportmidi.a(pmmacosxcm.o) "_MIDISend", referenced from: _midi_write_flush in libportmidi.a(pmmacosxcm.o) ld: symbol(s) not found Do these look familiar to any of the linux/windows guys? Johan ------------------------------------------------------------------------------ SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas, Nevada. The future of the web can't happen without you. Join us at MIX09 to help pave the way to the Next Web now. Learn more and register at http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix.com/ _______________________________________________ Mscore-developer mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/mscore-developer |
||||||||||||||||
|
lasconic
|
Where did you get portmidi ? precompiled ?
The first linking error "kMIDIPropertyConnectionUniqueID", seems to be part of MacOSX xmidi http://xmidi.com/docs/coremidi47.html Maybe the portmidi lib or object file (pmmacosxcm) is not compiled with the same xmidi lib you are trying to link ? Just trying to guess again :) Lasconic 2008/12/8 Johan Pauwels <[hidden email]>
------------------------------------------------------------------------------ SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas, Nevada. The future of the web can't happen without you. Join us at MIX09 to help pave the way to the Next Web now. Learn more and register at http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix.com/ _______________________________________________ Mscore-developer mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/mscore-developer |
||||||||||||||||
|
Frye
|
I got a tiny bit ahead with specifyin CoreMidi in the Other linker flags section under Linking in Xcode. However now my attempts crash on portmidi compilation for nonexistent windows.h. I'll try some more tomorrow. There might also be a way around this by using the local version of portmidi instead of the one under mscore.
-Samuli
|
||||||||||||||||
wschweer
|
Whithout changing the cmake configuration file you are compiling the
windows part of portmidi. Look at portmidi/CMakeLists.txt. The pm_win/xxx .c files must be removed and replaced by files needed for the mac os x driver. Sometthing like: if (MINGW) add_library ( portmidi STATIC pm_common/pmutil.c pm_win/pmwinmm.c pm_win/pmwin.c pm_common/portmidi.c porttime/porttime.c porttime/ptwinmm.c ) else (MINGW) add_library ( portmidi STATIC pm_common/pmutil.c pm_mac/pmmac.c pm_mac/pmmacosxcm.c pm_common/portmidi.c porttime/porttime.c porttime/ptwinmm.c ) endif (MINGW) On Monday 08 December 2008 17:44:02 Frye wrote: > I got a tiny bit ahead with specifyin CoreMidi in the Other linker flags > section under Linking in Xcode. However now my attempts crash on portmidi > compilation for nonexistent windows.h. I'll try some more tomorrow. There > might also be a way around this by using the local version of portmidi > instead of the one under mscore. > > -Samuli > > lasconic wrote: > > Where did you get portmidi ? precompiled ? > > The first linking error "*kMIDIPropertyConnectionUniqueID"*, seems to be > > part of MacOSX xmidi http://xmidi.com/docs/coremidi47.html > > Maybe the portmidi lib or object file (*pmmacosxcm)* is not compiled with > > the same xmidi lib you are trying to link ? > > Just trying to guess again :) > > > > Lasconic > > > > 2008/12/8 Johan Pauwels <[hidden email]> > > > >> On 07 Dec 2008, at 18:21, Frye wrote: > >> > >> I am able to get it to compile, but linking is giving me the next > >> problems. > >> There were few changes I had to make, but nothing big. However I still > >> cannot get anything out of the build process I could run. > >> > >> Any ideas? Portaudio is installed, but I had to manually specify it's > >> header > >> location to get the build process to find portaudio.h so it may be that > >> there is something else wrong in my setup. > >> > >> > >> I got a tiny bit further, but not much. Now I have linking errors not > >> against portmidi, but something in the inwards of portmidi as far as I > >> understand. What I've done is (all settings in Xcode) > >> > >> In target "mscore" info > build > Search paths > >> Library Search Path set to /opt/local/lib > >> (just as you add /opt/local/include to Header Search Path to be able to > >> compile) > >> > >> In target "mscore" info > build > Linking > >> add -lportmidi to Other Linker Flags > >> (-lportaudio is added by CMake, but -lportmidi not, this should probably > >> be > >> done by CMake too) > >> > >> The resulting linking errors are: > >> > >> *Undefined symbols:* > >> * "_kMIDIPropertyConnectionUniqueID", referenced from:* > >> * _kMIDIPropertyConnectionUniqueID$non_lazy_ptr in > >> libportmidi.a(pmmacosxcm.o)* > >> * "_MIDIGetNumberOfSources", referenced from:* > >> * _pm_macosxcm_init in libportmidi.a(pmmacosxcm.o)* > >> * "_MIDIPacketListAdd", referenced from:* > >> * _send_packet in libportmidi.a(pmmacosxcm.o)* > >> * _send_packet in libportmidi.a(pmmacosxcm.o)* > >> * "_MIDIObjectGetStringProperty", referenced from:* > >> * _EndpointName in libportmidi.a(pmmacosxcm.o)* > >> * _EndpointName in libportmidi.a(pmmacosxcm.o)* > >> * _EndpointName in libportmidi.a(pmmacosxcm.o)* > >> * _ConnectedEndpointName in libportmidi.a(pmmacosxcm.o)* > >> * "_kMIDIPropertyName", referenced from:* > >> * _kMIDIPropertyName$non_lazy_ptr in libportmidi.a(pmmacosxcm.o)* > >> * "_MIDIPacketListInit", referenced from:* > >> * _send_packet in libportmidi.a(pmmacosxcm.o)* > >> * _midi_write_short in libportmidi.a(pmmacosxcm.o)* > >> * _midi_begin_sysex in libportmidi.a(pmmacosxcm.o)* > >> * "_MIDIInputPortCreate", referenced from:* > >> * _pm_macosxcm_init in libportmidi.a(pmmacosxcm.o)* > >> * "_MIDIGetNumberOfDevices", referenced from:* > >> * _pm_macosxcm_init in libportmidi.a(pmmacosxcm.o)* > >> * "_MIDIPortConnectSource", referenced from:* > >> * _midi_in_open in libportmidi.a(pmmacosxcm.o)* > >> * "_MIDIOutputPortCreate", referenced from:* > >> * _pm_macosxcm_init in libportmidi.a(pmmacosxcm.o)* > >> * "_Pt_Time", referenced from:* > >> * _Pt_Time$non_lazy_ptr in pm.o* > >> * _Pt_Time$non_lazy_ptr in libportmidi.a(portmidi.o)* > >> * _Pt_Time$non_lazy_ptr in libportmidi.a(pmmacosxcm.o)* > >> * "_Pt_Stop", referenced from:* > >> * PortMidiDriver::~PortMidiDriver()in pm.o* > >> * PortMidiDriver::init() in pm.o* > >> * PortMidiDriver::~PortMidiDriver()in pm.o* > >> * PortMidiDriver::~PortMidiDriver()in pm.o* > >> * "_MIDIEntityGetDevice", referenced from:* > >> * _EndpointName in libportmidi.a(pmmacosxcm.o)* > >> * "_MIDIPortDisconnectSource", referenced from:* > >> * _midi_in_close in libportmidi.a(pmmacosxcm.o)* > >> * "_MIDIGetDestination", referenced from:* > >> * _pm_macosxcm_init in libportmidi.a(pmmacosxcm.o)* > >> * "_MIDIClientCreate", referenced from:* > >> * _pm_macosxcm_init in libportmidi.a(pmmacosxcm.o)* > >> * "_MIDIObjectFindByUniqueID", referenced from:* > >> * _ConnectedEndpointName in libportmidi.a(pmmacosxcm.o)* > >> * "_MIDIEndpointGetEntity", referenced from:* > >> * _EndpointName in libportmidi.a(pmmacosxcm.o)* > >> * "_MIDIGetNumberOfDestinations", referenced from:* > >> * _pm_macosxcm_init in libportmidi.a(pmmacosxcm.o)* > >> * "_MIDIDeviceGetNumberOfEntities", referenced from:* > >> * _EndpointName in libportmidi.a(pmmacosxcm.o)* > >> * "_MIDIPortDispose", referenced from:* > >> * _pm_macosxcm_term in libportmidi.a(pmmacosxcm.o)* > >> * _pm_macosxcm_term in libportmidi.a(pmmacosxcm.o)* > >> * "_Pt_Start", referenced from:* > >> * PortMidiDriver::init() in pm.o* > >> * _Pm_OpenOutput in libportmidi.a(portmidi.o)* > >> * _midi_in_open in libportmidi.a(pmmacosxcm.o)* > >> * "_MIDIClientDispose", referenced from:* > >> * _pm_macosxcm_term in libportmidi.a(pmmacosxcm.o)* > >> * "_MIDIGetSource", referenced from:* > >> * _pm_macosxcm_init in libportmidi.a(pmmacosxcm.o)* > >> * "_Pt_Started", referenced from:* > >> * _Pm_OpenOutput in libportmidi.a(portmidi.o)* > >> * _midi_in_open in libportmidi.a(pmmacosxcm.o)* > >> * "_MIDIObjectGetDataProperty", referenced from:* > >> * _ConnectedEndpointName in libportmidi.a(pmmacosxcm.o)* > >> * "_MIDISend", referenced from:* > >> * _midi_write_flush in libportmidi.a(pmmacosxcm.o)* > >> *ld: symbol(s) not found* > >> * > >> * > >> Do these look familiar to any of the linux/windows guys? > >> > >> Johan > >> > >> > >> ------------------------------------------------------------------------ > >>------ SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las > >> Vegas, Nevada. > >> The future of the web can't happen without you. Join us at MIX09 to > >> help pave the way to the Next Web now. Learn more and register at > >> > >> http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix. > >>com/ _______________________________________________ > >> Mscore-developer mailing list > >> [hidden email] > >> https://lists.sourceforge.net/lists/listinfo/mscore-developer > > > > ------------------------------------------------------------------------- > >----- SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas, > > Nevada. > > The future of the web can't happen without you. Join us at MIX09 to help > > pave the way to the Next Web now. Learn more and register at > > http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix.c > >om/ _______________________________________________ > > Mscore-developer mailing list > > [hidden email] > > https://lists.sourceforge.net/lists/listinfo/mscore-developer ------------------------------------------------------------------------------ SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas, Nevada. The future of the web can't happen without you. Join us at MIX09 to help pave the way to the Next Web now. Learn more and register at http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix.com/ _______________________________________________ Mscore-developer mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/mscore-developer |
||||||||||||||||
|
Frye
|
In reply to this post
by Frye
Sorry to bug you with maybe trivial questions, but I'm a real beginner with this mac stuff. Now it seems that these two files are missing:
mscore-build/mscore/portmidi/pm_mac/pmmacosxcm.c:32:19: error: pmmac.h: No such file or directory mscore-build/mscore/portmidi/pm_mac/pmmacosxcm.c:33:24: error: pmmacosxcm.h: No such file or directory Should these be provided by the portmidi package? Or are these something that are generated during the build process? -Samuli
|
||||||||||||||||
|
lasconic
|
Both seems to be a part of portmidi source : http://portmedia.svn.sourceforge.net/viewvc/portmedia/portmidi/trunk/pm_mac/
2008/12/8 Frye <[hidden email]>
------------------------------------------------------------------------------ SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas, Nevada. The future of the web can't happen without you. Join us at MIX09 to help pave the way to the Next Web now. Learn more and register at http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix.com/ _______________________________________________ Mscore-developer mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/mscore-developer |
||||||||||||||||
|
Frye
|
Ok, I need to grab the latest portmidi files and try with those. As there seem to be quite a few mac files missing from the current one.
Thanks for pointing that out. I really need to get more familiar with svn too =)
|
||||||||||||||||
|
wschweer
|
Ther files are already there but the compiler does not find them.
You should try to add another include path to portmidi/CMmakeLists.txt: include_directories( ... ${PROJECT_SOURCE_DIR}/portmidi/pm_mac ... On Monday 08 December 2008 21:53:49 Frye wrote: > Ok, I need to grab the latest portmidi files and try with those. As there > seem to be quite a few mac files missing from the current one. > > Thanks for pointing that out. I really need to get more familiar with svn > too =) > > lasconic wrote: > > Both seems to be a part of portmidi source : > > http://portmedia.svn.sourceforge.net/viewvc/portmedia/portmidi/trunk/pm_m > >ac/ ------------------------------------------------------------------------------ SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas, Nevada. The future of the web can't happen without you. Join us at MIX09 to help pave the way to the Next Web now. Learn more and register at http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix.com/ _______________________________________________ Mscore-developer mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/mscore-developer |
||||||||||||||||
|
Frye
|
Weird as my source tree seems to lack them. After I copied the new pm_mac directory over everything seems alot brighter. Currently only 1 error. And that is about the CoreMidi I put to the config earlier. So it somehow fixes those linker errors, but still causes some trouble. And removing the CoreMidi from the linker flags pops back tose pm linker errors. I'm getting too tired today, but maybe I can find more time tomorrow to see what is going on.
Thanks for all the help so far. -Samuli
|
||||||||||||||||
|
Frye
|
Nah, could not give up just yet, but no I have to.
Below is what I'm getting at the moment, if anyone have any Ideas where to look feel free to give me a hint. -Samuli Undefined symbols: "_Pm_GetDefaultInputDeviceID", referenced from: PortMidiDriver::init() in pm.o "_Pm_SetFilter", referenced from: PortMidiDriver::init() in pm.o "_Pm_OpenInput", referenced from: PortMidiDriver::init() in pm.o "_Pt_Time", referenced from: _Pt_Time$non_lazy_ptr in pm.o "_Pt_Stop", referenced from: PortMidiDriver::~PortMidiDriver()in pm.o PortMidiDriver::~PortMidiDriver()in pm.o PortMidiDriver::~PortMidiDriver()in pm.o PortMidiDriver::init() in pm.o "_Pm_GetDefaultOutputDeviceID", referenced from: PortMidiDriver::init() in pm.o "_Pm_Read", referenced from: PortMidiDriver::init() in pm.o PortMidiDriver::read() in pm.o "_Pm_Poll", referenced from: PortMidiDriver::init() in pm.o PortMidiDriver::read() in pm.o "_Pm_GetErrorText", referenced from: PortMidiDriver::init() in pm.o "_Pt_Start", referenced from: PortMidiDriver::init() in pm.o "_Pm_Close", referenced from: PortMidiDriver::~PortMidiDriver()in pm.o PortMidiDriver::~PortMidiDriver()in pm.o PortMidiDriver::~PortMidiDriver()in pm.o ld: symbol(s) not found collect2: ld returned 1 exit status "_Pm_GetDefaultInputDeviceID", referenced from: PortMidiDriver::init() in pm.o "_Pm_SetFilter", referenced from: PortMidiDriver::init() in pm.o "_Pm_OpenInput", referenced from: PortMidiDriver::init() in pm.o "_Pt_Time", referenced from: _Pt_Time$non_lazy_ptr in pm.o "_Pt_Stop", referenced from: PortMidiDriver::~PortMidiDriver()in pm.o PortMidiDriver::~PortMidiDriver()in pm.o PortMidiDriver::~PortMidiDriver()in pm.o PortMidiDriver::init() in pm.o "_Pm_GetDefaultOutputDeviceID", referenced from: PortMidiDriver::init() in pm.o "_Pm_Read", referenced from: PortMidiDriver::init() in pm.o PortMidiDriver::read() in pm.o "_Pm_Poll", referenced from: PortMidiDriver::init() in pm.o PortMidiDriver::read() in pm.o "_Pm_GetErrorText", referenced from: PortMidiDriver::init() in pm.o "_Pt_Start", referenced from: PortMidiDriver::init() in pm.o "_Pm_Close", referenced from: PortMidiDriver::~PortMidiDriver()in pm.o PortMidiDriver::~PortMidiDriver()in pm.o PortMidiDriver::~PortMidiDriver()in pm.o ld: symbol(s) not found collect2: ld returned 1 exit status
|
||||||||||||||||
Joachim Ganseman
|
Oops, by accident sent this only to Samuli (my apologies) - now to the
whole list: Hello, I gave the mac build a new shot, but unfortunately did not get any further. For the current checkout (rev 1418), the following changes are still needed to make the cmake xcode build work from scratch: select.cpp: values.h -> limits.h utils.cpp: change line 358 to: #if defined (__MINGW32__) || defined (__APPLE__) score/cmakelists: comment away the lines containing "resfile.o" and "pa.cpp" and "pm.cpp" then: install QT and CMake for mac (I skip any sound library bindings for the moment), run ./gen-qt-projectfile . >> mscore.pro run cmake -G Xcode And I still don't get the font working properly. I'm really beginning to suspect this is some kind of bug in Qt itself, not in MuseScore. As far as I can tell, all resources, images, etc, are imported properly through the Qt .qrc file. All images are displayed perfectly, font is also rendered as long as it is normal text. But somehow, the Mscore fonts, which are used for display of the score and generation of some knobs in the GUI, won't display properly. I've been digging through quite some code trying to figure this out. There are a few things that seem quite strange to me: - what is the difference between the array pSymbols[], defined in sym.cpp, and the array symbols[], which is being defined in the same file on line 364? I find it very strange that in pSymbols[] the "note2sym" gets the hex code "0xe104" while in symbols[] that same code is tied to "Rest_M3". And why are all these numbers chosen so high? - in that same file, I see the syntax foreach() appear regularly. As far as I know, that is not standard C++ syntax - there is however a for_each() function in the C++ STL. My compiler does not complain, so I guess there is no problem with it, but just out of curiosity, where does it come from? - just to check: calling the CTable() function as a debug measure, generates me a file containing these lines as first: { 56, 2, 8, -91, 51, 92, 67 }, { 49, 2, 15, -91, 35, 91, 67 }, { 53, 2, 7, -89, 50, 90, 67 }, { 57600, 0, 11, -239, 189, 239, 211 }, { 57601, 0, 11, -239, 189, 239, 211 }, { 57602, 0, 11, -239, 189, 239, 211 }, { 57603, 0, 11, -239, 189, 239, 211 }, { 57605, 0, 11, -239, 189, 239, 211 }, can anyone confirm whether this is correct or not? - in the Qt docs is stated that QPixMap can be hardware dependant (see http://doc.trolltech.com/4.4/qpixmap.html#details). I tried painting the GUI icons based on fonts (done in symIcon() in icons.cpp) on QImages instead and converting those to QPixMaps later, but that had no effect. I tried digging deeper in the Qt sources themselves to see how an image is drawn ( see http://www.koders.com/cpp/fid042633577810F7D295CD1C5A772BBC1F7C702A19.aspx?s=drawTextItem#L4364 ), but I eventually gave up. As a last resort I installed the Qt 4.5.0 beta, but that had no effect either (it runs and compiles exactly the same as previously). I installed Eclipse with CDT and made the project file for that editor (see http://www.cmake.org/Wiki/Eclipse_CDT4_Generator ) but that had no effect either. It does build flawlessly, too, but still no fonts. I'll keep anyone updated on any new progress, but as for now, none has been made by me (yet) Best wishes to everyone for the new year, of course, too :) Joachim ------------------------------------------------------------------------------ Check out the new SourceForge.net Marketplace. It is the best place to buy or sell services for just about anything Open Source. http://p.sf.net/sfu/Xq1LFB _______________________________________________ Mscore-developer mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/mscore-developer |
||||||||||||||||
|
wschweer
|
The score font is loaded with the qt function
QFontDatabase::addApplicationFont(). I looked into the qt task tracker and found the bug 210200: When adding a font with QFontDatabase::addApplicationFont() it fails on Leopard, while it works on Tiger. (in qt4.4.0 scheduled for 4.5.0) You can try to install the two fonts in question mscore-20.ttf and mscore1-20.ttf manually and hope that mscore finds them. On Thursday 08 January 2009 18:25:38 Joachim Ganseman wrote: > Oops, by accident sent this only to Samuli (my apologies) - now to the > whole list: > > Hello, > > I gave the mac build a new shot, but unfortunately did not get any further. > > For the current checkout (rev 1418), the following changes are still > needed to make the cmake xcode build work from scratch: > select.cpp: values.h -> limits.h > utils.cpp: change line 358 to: #if defined (__MINGW32__) || defined > (__APPLE__) score/cmakelists: comment away the lines containing "resfile.o" > and "pa.cpp" and "pm.cpp" > > then: install QT and CMake for mac (I skip any sound library bindings > for the moment), > run ./gen-qt-projectfile . >> mscore.pro > run cmake -G Xcode > > And I still don't get the font working properly. I'm really beginning > to suspect this is some kind of bug in Qt itself, not in MuseScore. As > far as I can tell, all resources, images, etc, are imported properly > through the Qt .qrc file. All images are displayed perfectly, font is > also rendered as long as it is normal text. But somehow, the Mscore > fonts, which are used for display of the score and generation of some > knobs in the GUI, won't display properly. > > I've been digging through quite some code trying to figure this out. > There are a few things that seem quite strange to me: > > - what is the difference between the array pSymbols[], defined in > sym.cpp, and the array symbols[], which is being defined in the same > file on line 364? I find it very strange that in pSymbols[] the > "note2sym" gets the hex code "0xe104" while in symbols[] that same > code is tied to "Rest_M3". And why are all these numbers chosen so > high? there are two different fonts, one for score typesettings and another for use in normal text context. The code numbers use some free space in the unicode tables. > > - in that same file, I see the syntax foreach() appear regularly. As > far as I know, that is not standard C++ syntax - there is however a > for_each() function in the C++ STL. My compiler does not complain, so > I guess there is no problem with it, but just out of curiosity, where > does it come from? foreach is part of the qt library and works with qt container classes. I changed from the STL container to qt container because they are much easier to use. > > - just to check: calling the CTable() function as a debug measure, > generates me a file containing these lines as first: > { 56, 2, 8, -91, 51, 92, 67 }, > { 49, 2, 15, -91, 35, 91, 67 }, > { 53, 2, 7, -89, 50, 90, 67 }, > { 57600, 0, 11, -239, 189, 239, 211 }, > { 57601, 0, 11, -239, 189, 239, 211 }, > { 57602, 0, 11, -239, 189, 239, 211 }, > { 57603, 0, 11, -239, 189, 239, 211 }, > { 57605, 0, 11, -239, 189, 239, 211 }, > can anyone confirm whether this is correct or not? its not correct. It looks like the mscore font is not found and substituted by another one (codepoints 57600-57605 are all the same). Werner ------------------------------------------------------------------------------ Check out the new SourceForge.net Marketplace. It is the best place to buy or sell services for just about anything Open Source. http://p.sf.net/sfu/Xq1LFB _______________________________________________ Mscore-developer mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/mscore-developer |
||||||||||||||||
| Free Embeddable Forum Powered by Nabble | Help |
