Nayan Hajratwala over at Agile Shrugged managed to write a really great article describing how he got Subversion repos running on Mac OS X Server. Cheers Nayan!

I got mine working on Mac OS X 10.5.8 but had to make a few different changes in order to get things working. His doesn't deal with multiple SVN repositories, only with a single one. This isn't ideal and I highly doubt anyone would go to the trouble of installing a SVN server and only ever need ONE svn repository.

I prefer using /opt/repos/svn as the master path as the Open.collab.net developers recommend doing it this way on Mac OS X.

Step 1 - Create a root SVN folder

Open Terminal.app:

sudo mkdir -p /opt/repos/svn

Step 2 - Create your SVN repository

sudo svnadmin create /opt/repos/svn/test
sudo chown -R www:www /opt/repos/svn

The second line makes Apache own this path and everything inside it including your new repo.

Step 3 – Configure the Apache web server via Server Admin

Server Admin is located in /Applications/Server/Server Admin.

1. Click Web in the Sidebar
2. Click Settings in the top bar
3. Select the Modules tab
4. Check the following modules:

authz_svn_module
dav_module
dav_fs_module
dav_svn_module

Step 4 – Enable WebDAV on your site

Reading comments, it seems that some people don't think you need this step. However for my configuration I'm pretty sure you do need to do this.

1. Click Sites in the top bar
2. Select your site
3. Click the WebDAV checkbox

Step 5 - Create a Realm

This will set up the OpenDirectory based security. We will be using Basic authentication only.

1. Select the Realms tab
2. Click “+” under the Realms box and enter the following:

Realm Name: svn_realm
Authentication: Basic
Location: /svn (Note: You have to change the dropdown to 'Location', not 'Folder').

3. Click “+” under the Users & Groups box, and drag/drop the users/groups that you want to give access to. Also be sure to give them read/write privileges...then press the Save button.

Step 6 – Tweak the Apache config file

There is no graphical UI to do this. So you need to use Terminal.app.
Edit the file under /etc/apache2/sites that corresponds to your site.
This should be a file starting with 000x.
In the <location "/svn"> section, add the following three lines:

DAV svn
SVNParentPath /opt/repos/svn
SVNListParentPath On

This is what mine looks like:

Step 7 – Restart the web server

You can do this from Server Admin ('Stop Web', then 'Start Web').

You should be able to access your repository at http://myserver/svn or locally at: http://localhost/svn

Be careful not to change any other settings in the Web portion of Server admin.
If you do it will change DAV svn to DAV Off again.

Free free to leave any comments and suggestions

Macports doesn't install ODE correctly either. Macports installs a PPC binary on your system (even if you have an Intel Mac). So here is the correct way to compile ODE on a Mac.

Download the ODE Source Code from Sourceforge.

Extract the source to a directory of your choosing (I use: /lib/ode-0.11.1)

Open a Terminal:

cd build
premake4 gmake --platform=universal32
cd gmake
make

At this point you need to add in some flags to the makefile in order to get this to build on Intel correctly. So open ode.make and change the following 2 lines:

cd ode-0.11.1/build/gmake
vim ode.make

You need to add -m32 to the CXXFLAGS, which you can do by adding it to the CPPFLAGS (which uses CXXFLAGS. Find the line starting with CPPFLAGS and change it to this:

CPPFLAGS  += -MMD -m32 $(DEFINES) $(INCLUDES)

You also need to add -m32 to LDFLAGS:

LDFLAGS   += -m32 -dynamiclib -flat_namespace -framework Carbon

Save and quit.

Build Release Libraries:

make config=releasesingledll

Build Debug Libraries:

make config=debugsingledll

This will build the libraries to here:

ode-0.11.1/lib/ReleaseSingleDLL/libode_single.dylib
ode-0.11.1/lib/DebugSingleDLL/libode_singled.dylib

So there you go. Enjoy using ODE!

Install Bullet (Download Link):

cd bullet-2.75/
cmake . -G "Unix Makefiles" -DBUILD_EXTRAS=on -DBUILD_DEMOS=on -DCMAKE_BUILD_TYPE=Debug
make

Pretty simple. The demos will be built to bullet-2.75/demos/ and the Bullet libs are built to bullet-2.75/src/

1. Install Bullet (Download link):

You need to do a little bit of hackery on Mac OS X in order to get Bullet to build properly. (I'm running Snow Leopard 10.6). I have installed Bullet into /lib/bullet-2.75/

First, you need to get the CMakeLists.txt file here:

nano bullet-2.75/src/CMakeLists.txt

I originally got an error "ld: library not found for -lBulletMultiThreaded".
The solution (I found here) is to comment out the following lines using a pound symbol # like so:

#if (CMAKE_SIZEOF_VOID_P MATCHES "8")
#SUBDIRS( BulletSoftBody BulletCollision BulletDynamics LinearMath )
#else (CMAKE_SIZEOF_VOID_P MATCHES "8")
SUBDIRS( BulletMultiThreaded BulletSoftBody BulletCollision BulletDynamics LinearMath )
#endif (CMAKE_SIZEOF_VOID_P MATCHES "8")

Download and Install CMake.

You can now run CMake on the /lib/bullet-2.75 directory. CMake, configure, generate. CMake will generate an Xcode Project for you to run:

/lib/bullet-2.75/BULLET_PHYSICS.xcodeproj

Build the release and debug libraries and demos using AppAllBulletDemos, Debug/Release, i386, ALL_BUILD

Xcode builds the libraries here:

/lib/bullet-2.75/src

The Bullet Demo applications get built to here:

/lib/bullet-2.75/Demos

So you should have something like:

/lib/bullet-2.75/src/BulletCollision/Debug/libBulletCollision.a
/lib/bullet-2.75/Demos/AllBulletDemos/Debug/AppAllBulletDemos // <--Run this in the terminal

Now that Bullet is compiled, we can install and run PAL.

2. ODE (Open Dynamics Engine):

It no longer runs on Snow Leopard due to the DrawStuff API being removed in Snow Leopard.
UPDATE: I figure out how to compile ODE on OS X 10.6.

3. Install PAL:

When you run CMake on the PAL directory, you will find that it won't detect most of the Bullet library paths by itself. This requires some manual entries by yourself in order to get this to work. I will list my paths below, but note that I installed Bullet in a couple of places so these paths may not be right for you. Use this as a guideline only.

BULLET_LIBRARY_BulletDynamics_Debug
/lib/bullet-2.75/src/BulletDynamics/Debug/libBulletDynamics.a

One thing to note is that you will see BULLET_LIBRARY_BulletMath-NOTFOUND and BULLET_LIBRARY_BulletMath_DEBUG-NOTFOUND. I'm not sure why this happens but it didn't cause a problem for me, you can just ignore this.

Once you have setup all your paths, you can press Configure > Choose Xcode > then press Generate.
This will Generate the Xcode project file for you in /lib/PAL/PAL.xcodeproj

If you run the Xcode project, you can then build the libraries for PALBeginner binary I belive is Windows only as I get an error when I run it. Also, when you build you will need to change the build architecture to i386 as it doesn't seem that 64Bit is supported yet.

install these packages with synaptic:

swig
python-wxgtk2.8
libois-dev
libode-dev
libogre-dev ( Must be version1.4.9 )
libfltk1.1-dev
libxml2
libxml2-dev
scons
libboost-dev
libboost-signals-dev
libboost-signals1.40-dev
openal
libavformat
libtool
libxft-dev

download and install these packages:

wget http://launchpadlibrarian.net/16469340/libogremain-1.4.9_1.4.9.dfsg1-1_i386.deb
wget http://launchpadlibrarian.net/16469345/libogre-dev_1.4.9.dfsg1-1_all.deb

I got the above packages from here:

https://launchpad.net/ubuntu/intrepid/i386/libogremain-1.4.9/1.4.9.dfsg1-1

https://launchpad.net/ubuntu/karmic/i386/libogre-dev/1.4.9.dfsg1-1

modify some source:

cd gazebo-0.8.0
vi server/gui/StatusBar.cc

change:

#include <stdio.h>   // change this
#include <cstring>    //  to this

run these commands (each line by itself), better yet append them to the end of ~/.bashrc to make it permanent:
For Bash: .bashrc (executed when you shart a shell) OR .bash_profile (executed when you log in)
For csh and tcsh: .cshrc
For sh and ksh: .profile

export PATH=/usr/local/bin:$PATH
export CPATH=/usr/local/include:$CPATH
export LIBRARY_PATH=/usr/local/lib:$LIBRARY_PATH
export PKG_CONFIG_PATH=/usr/local/lib/pkgconfig:$PKG_CONFIG_PATH
export LD_LIBRARY_PATH=/usr/local/lib:$LD_LIBRARY_PATH

build the source

sudo scons install

test that it worked:

gazebo worlds/simpleshapes.world

A Note on creating the ~/.gazeborc file: http://www.mail-archive.com/playerstage-gazebo@lists.sourceforge.net/msg00851.html

Disable .DS_Store file creation for network connections.

defaults write com.apple.desktopservices DSDontWriteNetworkStores true

-(void)copyToClipboard:(NSString*)str
{
NSPasteboard *pb = [NSPasteboard generalPasteboard];
NSArray *types = [NSArray     arrayWithObjects:NSStringPboardType, nil];
[pb declareTypes:types owner:self];
[pb setString: str forType:NSStringPboardType];
}

A quick note on Boost. If you want to use the Boost libraries, the good news is that you can just #include the boost classes that you want to use in your code (see bottom for Boost usage in code).
However if you want to use libs such as Boost::Filesystem or Boost::iostreams, they MUST be compiled into a library file. This is accomplished with the bjam tool which is kinda like an easy to use compiler for Boost instead of using GCC.

I can't believe that there are no tutorials on the net for compiling Boost on Mac OS X. I got it compiled for Visual Studio 2005 after reading numerous articles and then I adapted it for Mac OS X. Actually it's pretty easy to compile on Mac, a little easier than it was on Windows...OK, here we go!

Download Bjam and the Boost source from http://www.boost.org/

Compiling for Xcode 3.1:

1. You need to download and copy bjam.exe to

/usr/sbin

This will make the command available in the Terminal.
If it does not, check your paths in ~/.profile using vi or TextEdit. Find the line that starts with export PATH= and copy bjam to one of the paths listed (note: each path is separated by a colon ":")

2. CD to the boost directory (for example):

cd ~/Desktop/Boost/boost_1_37_0

3. To compile for Visual Studio 2005 as static release libs:

bjam toolset=darwin variant=release threading=multi link=static

4. The static debug version:

bjam toolset= darwin variant=debug threading=multi link=static

NOTE: If you want dynamic libraries (.dylib files instead of .a files) then use"link=shared instead of"link=static.

This will take a while (maybe 20 minutes) to run each of the above commands, and you"ll need both.
A Note about the naming of the output files:

Building with link=static gives you files with the following naming convention:

libboost_filesystem-xgcc40-mt-d-1_37.a

Building with link=shared gives you files with the following naming convention:

libboost_filesystem-xgcc40-mt-1_37.a

Including Boost in your C++ code

#include<iostream>
#include<boost/any.hpp> // any is an actual library

C++ Usage

boost::any a(5);
a = 7.67;
std::cout << boost::any_cast <double> (a) << std::endl;

Using the Compiled Libraries in Xcode

Using the libraries can be accomplished by dragging the static (.a) lib files to tyhe Xcode project. If you really want to use dynamic library files, then you will need to drag to the Xcode project, and also have a script run after the binary is built.
This is because dynamic libraries need to be copied inside the .app bundle because they are linked to at runtime, as opposed to compile time for static libs.

The actual process of scripting the dylib copy and modifying the darwin executable are already covered in another article I wrote here: How To use Dynamic Libraries in Xcode 3.1

Using the Header files in Xcode

Okay, it's getting to be all a big pain in the ass now, but we're almost done.

The headers can usually be dragged to the Xcode project for individual .h files. However because the Boost headers are in nested folders, it's a lot easier to just add the path to the Xcode project.

Go to: Project | Edit Project Settings | Build tab | User Header Search Paths
Add the path to where Boost is located on your hard disk. A good place is to put it somewhere inside the /Developer folder or /usr/local/include/. Also DO NOT check the "Recursive" option. It won't work if you do.

It's up to you where you want to put it....

Compiling for Visual Studio 2005/2008:

1. You need to download and copy bjam.exe to

C:Windows\system32

2. CD to the boost directory:

cd C:Program FilesBoostboost_1_37_0

3. To compile for Visual Studio 2005 as static release lib (use"msvc-8.0 for VS 2008):

bjam toolset=msvc-8.0 variant=release threading=multi link=static

The static debug version:

bjam toolset=msvc-8.0 variant=debug threading=multi link=static

Here is a dynamic dll, debug runtime version (for Visual Studio 2008). It will produce files with the following file name convention:" 'libboost_math_c99-vc90-mt-sgd-1_37.lib'

bjam toolset=msvc-9.0 variant=debug threading=multi runtime-link=static

This will take a while (maybe 20 minutes) to run each of the above commands, and you"ll need both.
A Note about the naming of the output files:
Building with link=static gives you files with the following naming convention:
boost_filesystem-vc80-mt-1_37.lib

Building with link=shared gives you files with the following naming convention:
libboost_filesystem-vc80-mt-1_37.dll

Update: Other options for compiling Boost with Bjam

Within the final build of the libraries, you will get a lib file like the following:
'libboost_math_c99-vc90-mt-sgd-1_37.lib'

The "lib" prefix is for static libraries. Use link=static
The 's' letter is to static linking to runtime. Use runtime-link=static
The 'd' is debug, use variant=debug
The 'g' is using debug runtime, to enable this you will need: runtime-debugging=on.

For more information, you can find the official documentation for building Boost here:

http://www.boost.org/doc/libs/1_38_0/more/getting_started/index.html

Well, thats it. Thanks for reading and if you have anything to add please leave a comment below (Sorry about the registration, I get too much spam otherwise).

Cheers,
Brock

Note: You can skip to the bottom to see the settings you need to set, that stop the debugger to jump around the place randomly.

The point of the debugger is that you can examine variables and program logic, "stepping through" the code, a single line at a time.

In the picture below, you can see a sample from Xcode's debugger.
The way the debugger works is when a breakpoint stops the program's execution, you can view the current values that all the variables hold.
The line that Xcode highlights blue hasn't actually been executed yet, it is just stopped on the line so you can process the logic yourself and then you press continue and check if it executes as you expect.

"

"

"

"

In the image above, you can see that the variable "number1" still holds the value of 2. This is because as I said above the debugger has stopped, the blue highlighted line has not been executed and will not be until the debugger moves to the line after it.

On a side note, DO NOT ENABLE ZEROLINK, or for that matter any other compiling optimisations such as SSE3 or Fast, etc. If you do, you will have weird problems with the debugger stepping through lines of code in the random orders.
What I mean by this is that the debugger will not step from the current line to the next line and the next. It will go to random lines in a seemingly random order. The reason this happens is because zerolink causes optimisations in the compiled application. The problem with this is the human understandable order of line by line gets thrown away and you therefore cannot debug using your logic.

You can check if Zerolink is enable by selecting:
"Project > Edit Active Target".
"

"

"

"

"

"

"

"

"

"

"

"

"

"

"

"

You can disable Zerolink by double clicking on the word "YES", or whatever it says then simply type NO.

"

"

"

"

"

"

"

"

"

"
"

"

"

"

"

"

"

"

"

"

"

Update:

Sorry about the long article on something so simple. I've decided not to delete the above as there are some parts that you may find useful (and pretty pictures!). After playing around with different settings I believe i've figured out the settings that actually matter."

Make sure the following have been set:

1) Instruction Scheduling = None
2) Optimization Level " " " = None [-O0]
3) ZERO_LINK " " " " " " " " "= None

The second thing you need to be careful of is that Xcode has 2 areas these settings can be set:

Project Settings: "Project > Edit Project Settings"
Target Settings: ""Project > Edit Active Target"

It is useful being able to have separate settings for the Project and the Target. However, the Target settings OVERRIDE the Project settings. So it is best to change the Target settings, not the Project settings.

If I discover any other settings that apply, I will update this post. If you see any mistakes or would like to add anything, leave a comment and i will update it