Re: [Lmuse-developer] [Lmuse-user] More fixes and changes to muse 0.9.x Plus instrument editor finished. Enjoy !

[Tim <ter...@ro...>]

On September 3, 2008 02:43:10 pm Frank Neumann wrote:
> Hi list,
> On Sun, 17 Aug 2008 21:18:33 -0400
> Tim <ter...@ro...> wrote:
>
> [..]
>
> > So, please try it out and let me know if there are any problems.
>
> I tried that now, and on a 64bit AMD system under Ubuntu 8.04 I get these
> compiler errors:
>
> make[4]: Entering directory `/home/franky/src/muse-cvs/build/muse'
> if g++ -DHAVE_CONFIG_H -I. -I. -I..  -Imidiedit -Iarranger -Iliste
> -Iwidgets -Imixer -Idriver -Iwaveedit -Implugins -Iinstruments
> -DINSTPREFIX=\"/usr/local\"  -g -fno-exceptions -Wall -W -D_GNU_SOURCE
> -D_REENTRANT   -DQT_CLEAN_NAMESPACE -DQT_NO_COMPAT   -I.. -I../muse/widgets
> -I/usr/share/qt3/include -O3 -fomit-frame-pointer -ffast-math
> -fstrength-reduce -funroll-loops -mcpu=athlon64 -march=athlon64 -I..
> -I../synti -I../muse/widgets -DQT_SHARED -DQT_THREAD_SUPPORT -DQT_PLUGIN  
> -g  -MT audiotrack.o -MD -MP -MF ".deps/audiotrack.Tpo" -c -o audiotrack.o
> audiotrack.cpp; \ then mv -f ".deps/audiotrack.Tpo" ".deps/audiotrack.Po";
> else rm -f ".deps/audiotrack.Tpo"; exit 1; fi `-mcpu=' is deprecated. Use
> `-mtune=' or '-march=' instead.
> [..]
> audiotrack.cpp: In function ‘void cacheJackRouteNames()’:
> audiotrack.cpp:40: error: cast from ‘AudioInput*’ to ‘unsigned int’ loses
> precision audiotrack.cpp:52: error: cast from ‘AudioOutput*’ to ‘unsigned
> int’ loses precision audiotrack.cpp:1085: error: cast from ‘const
> AudioTrack*’ to ‘unsigned int’ loses precision audiotrack.cpp:1105: error:
> cast from ‘const AudioTrack*’ to ‘unsigned int’ loses precision make[4]:
> *** [audiotrack.o] Error 1
Ooops! Hold the phone... That shouldn't have happened. 
Sorry I think that's my fault, NOT yours. Now that I look at the code 
 I think I remember why I cast a pointer to int like that.
I think it was because I couldn't figure out how to get the STL map to work 
 with pointers (in this case AudioInput*), so I had to cast as unsigned int !

Otherwise Line 25, for example, would be like this: 
std::map <AudioInput*, jackRouteNameMap > jackRouteNameCache;
 instead of this:
std::map <unsigned int, jackRouteNameMap > jackRouteNameCache;

Frank, I will try again. It's not a nice way of coding that...
Even if your fixes (changing to unsigned long) work, it's not nice.
Although we might just have to go with it in the end...

>
> The problem here is that on 64 Bit, unsigned int is 32 bit, while a pointer
> to a type is 64 bit. I can fix this locally by replacing some occurrances
> of "unsigned int" or "int" with "unsigned long", but since (as usual) I am
> not sure whether this is the correct solution, I'd rather leave it up to
> you to fix&commit that.
>
> After that, it at least compiled&linked. Now I'll test it a bit :-).
>
> The nice thing is that in order to understand a bit of your code, I had to
> read some tutorials about STL. Nice side-effect :-).
Yes, you might see some attempts in there at fixing a nagging problem of
 'erase-while-iterating' STL code. I recently read up on the problem and tried
 the solutions given, which I was not aware of, but I still had some problems.

>
> Greetings,
> Frank
>
> PS: Tim, for a long while I have been wondering - everywhere on this list,
> in source files, and in your user profile on SourceForge, I always just see
> "Tim" or "T356". Do you have a real name (particularly a last name)? :-)
I was thinking about that the other day. I guess ever since I got on the net
 in '94 I've always been paranoid about giving details, even my last name.
I've always been a bit 'passive' rather than 'active' on the net (no chat, 
 facebook etc.). 
Well, that is, until I started contributing to projects ! 
But you are right. It's a bit rude of me not to introduce myself.

Tell you what. I will put my full name in the AUTHORS file next commit.
I have to admit, the file does look kind of silly without my last name.
Those of you who are curious can see it next time...

Tim.