Key palette is empty on Mac PPC (Big-endian)

Could you try to Revert to factory settings ?

My iMac is the same condition too.
>key palette is empty in rc1 3070, Mac OsX 10.4.11 G5,

But the below error occurred.

$ /Applications/MuseScore.app/Contents/MacOS/mscore -F
in stat: : No such file or directory
in stat: : No such file or directory
Init midi driver failed
link SVGID_3_ hasn't been detected!
link SVGID_3_ hasn't been detected!
registerPlugin: add action idx 6
Load plugin: no menu property
preferences.checkUpdateStartup: 0
lastupdate: 11.05.2010 00:24:43.474
link SVGID_3_ hasn't been detected!

A duplicate can be found here #5996: Key signatures are missing in palette: R3070-0.9.6-RC1 with screenshot attached
http://musescore.org/en/node/5996#comment-16606

Someone can reproduce with 0.9.6 stable?

Also finding Key Palette empty in 0.9.6.

My system is:
Mac G5 Quad Running OS 10.4.11.

The Drum Palette is also empty but the Key Palette should probably be made a Priority Issue - otherwise we can only create scores in C Major or A Minor!!! ;o)

Ay Carumba: for instructions on using the Drum palette see drum notation in the handbook. If you have further questions on the topic please use the forums rather than this bug report.

For the moment, the empty key signature palette has been reported only on Mac 10.4 and on G5. Someone with a Mac Intel can reproduce ? If not, it might be an issue of endianness

Also running 10.4.11 on a G5, having the same issue - and I foolishly replaced my previous version of Musescore when I installed the new one. Is there a way to download a copy of 0.9.5?

aoide, for past versions see http://sourceforge.net/projects/mscore/files/
For further support questions please use the forums

I think the problem comes from :

union {
 int subtype;
  struct {
   int accidentalType:4;
   int naturalType:4;
   unsigned customType:16;
   bool custom : 1;
   bool invalid : 1;
   };
  };


The mapping between the int and the struct is different if the endianness changes apparently. Union are quite new for me. How will you set it up for a big endian system? Inversing the fields of the struct is not enough I guess?


 int subtype;
  struct {
   bool invalid : 1;
   bool custom : 1;
   unsigned customType:16;
   int naturalType:4;
   int accidentalType:4;
   };
  };


Maybe it's easier to be sure that the byte order is always the same in subtype?

Hi
The version 0.9.6 crashes at the very start on a MacBook Pro with an intel prcessoor under system 10.5.8...

21/06/10 12:26:55 com.apple.launchd[1467] ([0x0-0x59059].net.sourceforge.mscore[2258]) Exited abnormally: Bus error
21/06/10 12:27:48 [0x0-0x5b05b].net.sourceforge.mscore[2263] in stat: : No such file or directory
21/06/10 12:27:48 [0x0-0x5b05b].net.sourceforge.mscore[2263] in stat: : No such file or directory
21/06/10 12:27:48 [0x0-0x5b05b].net.sourceforge.mscore[2263] in stat: : No such file or directory
21/06/10 12:27:51 com.apple.launchd[1467] ([0x0-0x5b05b].net.sourceforge.mscore[2263]) Exited abnormally: Bus error

alatour, please use the forums to ask for help or report a new problem. This bug report is related to an empty key signature palette not a crash at startup.

In order to investigate I compile a version to print the following :
printf("Size of KeySigEvent = %d\n", sizeof(KeySigEvent) );
printf("Size of int = %d\n", sizeof(int) );

Thanks to Essington on IRC, I got size = 4 back for each line. So the struct has the size of the int.

Just because of a contradiction spirit, I tried the idea at #12 and changed the code for :


 int subtype;
  struct {
   bool nothing : 6;
   bool invalid : 1;
   bool custom : 1;
   unsigned customType:16;
   int naturalType:4;
   int accidentalType:4;
   };
  };


And it works! Now we need a way to do the distinction between PPC and Intel at runtime to deal with the different struct.

I have some idea to test to solve this issue. Unfortunatly I don't have a PPC Mac. Can someone with a PPC contact me to test a couple of ideas?

fixed in r3269. The patch is quite invasive and so I commited it only in the maintenance branch.
The patch has been tested by Essington. Thanks!

The beams are a separate issue that appears in 0.9.6 or later since it supports beaming across bar lines (but doesn't support beaming across systems). To fix the issue select the first note of each system and double click Beam . For any further discussion on the beaming topic see:
http://musescore.org/en/node/3970

Automatically closed -- issue fixed for 2 weeks with no activity.