In Xcode, how can I call a 'shell script' which is a Perl script that copies the .app and .dsym files to a different directory?
I want to pass the name of the project and/or the project's root directory to the script. I want to have the script called every time I build in release and distribution modes but not in debug mode.
For anyone else who is wondering, you can also do a simple copy via
[Click on Scheme Name] -> Edit Scheme... -> [Select Scheme] -> Run "Target" -> Post-actions
and add a cp command. In my case, for quick testing and ease of use, I wanted to copy my binary back to the project directory so that I could process some data files. I used the following:
cp ${TARGET_BUILD_DIR}/${TARGET_NAME} ${PROJECT_DIR}/"binaryFileName"
Right-click on your target and choose Add->New Build Phase->New Run Script Build Phase.
Inside the "Script" text area, add:
if [ ${CONFIGURATION} != "Debug" ]
then
/usr/bin/perl "${PROJECT_DIR}/myperlscript.pl" "${PRODUCT_NAME}" "${PROJECT_DIR}"
fi
Change the location and name of myperlscript.pl to be the location and name of your script.
Make sure to move the Run Script step to be after Link Binary With Libraries, since you're copying the built .app.
(Might also want to add the "perl" tag to the question)
Instead of a script, Xcode 12.4 has an build phase for copying files:
Show the project navigator (blue folder button)
Click on your project (blueprint icon)
Click "Build Phases" header in the opened settings pane
Click the + in the top left of the settings pane
Click "New Copy Files Phase"
Change Destination to Absolute Path and set your path
Click the + at the bottom to add a file to be copied
Save. If you run without saving, your new phase is ignored.
If you don't want an absolute path or one of the destinations provided, you'll have to use a Run Script Phase instead.
Related
When I build my project with Xcode 8, it saves the final build in ~/Library/Developer/Xcode/DerivedData/MyProject-[add-lots-of-random-chars-here]/Build/Products/Release-iphoneos. Is there any way to make Xcode copy the app bundle to a user-specified path after building it? e.g. how can I make Xcode copy the built app bundle to /MyBuilds after building it?
I know that I can change the path for storing derived data in my project's settings in Xcode but doing so will of course make Xcode store all data (including intermediate stuff like object code etc) in this location which I don't want. I really only want Xcode to copy the final, ready-for-distribution app bundle to a user-specified location without any intermediate files used in the build process.
How can I do that?
The solution using a script in "Build Phases" does not work properly since Xcode is not finished building the app when running the script. Here is a solution with a script that runs after all build tasks are finished:
Go to "Edit Scheme"
Click on the triangle next to "Build"
Select "Post-action"
Press the + button and select "New Run Script Option"
Select your app name in "Provide build settings from"
Add the following shell script:
Script:
PRODUCT="${BUILT_PRODUCTS_DIR}/${TARGET_NAME}.app"
cp -R "${PRODUCT}" ~/Desktop
Add a shell script to your build phases to copy the product:
Script:
PRODUCT="${BUILT_PRODUCTS_DIR}/${TARGET_NAME}.app"
cp -R "${PRODUCT}" ~/Desktop
Certainly replace ~/Desktop with a target directory of your choice.
I have an Xcode project with a large number of targets where I would like to include a settings bundle for apps built under the Ad-hoc and Debug configurations, but not under the Release configuration.
Build Phases don't seem to allow for making themselves conditional on configuration (they can obviously be conditional on target, but doubling the number of targets in the project would make it completely unusable).
That leaves writing a custom Build Rule. My plan is to exclude the Settings.bundle from all targets, and create a build rule that conditionally copies it into the product package, but applicable examples are really hard to find.
The build rule I've started has the Process setting set to "Source files with names matching:" and Settings.bundle as the name. The Using setting is "Custom script:".
My custom script is as follows (with the caveat that my bash scripting is on a cargo cult level):
if [${CONFIGURATION} = 'Debug'] then
cp -r ${INPUT_FILE_PATH} ${DERIVED_FILES_DIR}/.
fi
Finally, I have ${DERIVED_FILES_DIR}/Settings.bundle listed as an output file.
Since I'm here, it should be obvious that it's not working. My first question is whether there is somewhere I can view the output of the build rules as the execute to make sure that 1) it's actually being executed and that 2) I don't have a stupid syntax error somewhere.
Also, what's the proper location (in the form of an environment variable) to copy the output to?
I finally figured it out.
For each target for which you want to conditionally include the settings bundle, choose its Project from the source list, choose the target, and switch to the "Build Phases" tab.
Click the "Add Build Phase" button and choose "Add Run Script".
Then enter the following for the script:
if [ "${CONFIGURATION}" == "Debug" ]; then
cp -r "${PROJECT_DIR}/Settings.bundle" "${BUILT_PRODUCTS_DIR}/${PRODUCT_NAME}.app"
fi
I know this question has been answered already, and the answer was very helpful to me, but I wanted to throw my own modified solution out there as well.
My requirement was to have different settings bundles for different build configurations, rather than just not including it at release. Assuming a simplistic approach of only Debug and Release configurations, here's how to do it:
Start by adding 2 settings bundles to the project, named Settings-debug.bundle and Settings-release.bundle and then remove these files from the Copy Bundle Resources build phase. Next add a user defined build setting called SETTINGS_BUNDLE, which has different values for each configuration:
Debug ${PROJECT_DIR}/relative/path/to/Settings-debug.bundle
Release ${PROJECT_DIR}/relative/path/to/Settings-release.bundle
Next add a run-script build phase (after Copy Bundle Resources) named Copy Settings Bundle with a modified version of the script in Frank's solution.
cp -r "${SETTINGS_BUNDLE}/" "${BUILT_PRODUCTS_DIR}/${PRODUCT_NAME}.app/Settings.bundle"
The difference here is that the copied bundle is always named Settings.bundle regardless of the source name.
You then need to add another build phase script to prevent code signing errors when the only changes are in the settings bundles. It forces the code signing step to occur on every build. This should run before the Compile Source Files build phase. I called mine Force Codesign.
touch "${PROJECT_DIR}/relative/path/to/main.m"
For complied sources, there is a poorly documented user defined build setting that can be added. Files can be both excluded and included from compilation
Go to your target's Build Settings > Tap the + button > Add User-Defined Setting
The key is either INCLUDED_SOURCE_FILE_NAMES or EXCLUDED_SOURCE_FILE_NAMES
The value is a space separated list of file paths
See reference:
http://lists.apple.com/archives/xcode-users/2009/Jun/msg00153.html
(Tested with Xcode 9.3)
I can't find when Xcode included this feature but EXCLUDED_SOURCE_FILE_NAMES is now directly available in Build Settings > Build Options > Excluded Source File Names.
So you no longer need to create a User-Defined Setting.
See below:
It will automatically add this line in your .pbxproj.
Settings.bundle is always copied into destination area no matter whether Release or Debug configuration. So, maybe you need the following code:
if [ ${CONFIGURATION} == "Release" ]; then
rm -rf ${BUILT_PRODUCTS_DIR}/${PRODUCT_NAME}.app/Settings.bundle
fi
I am no shell script expert but I think you need space between the square brackets and the condition. Also, quoting the variables may help:
if [ "${CONFIGURATION}" = "Debug" ] then
cp -r "${INPUT_FILE_PATH}" "${DERIVED_FILES_DIR}"/.
fi
As for the location, I use "$BUILT_PRODUCTS_DIR"/"$FULL_PRODUCT_NAME" for the root of my OS X app bundle.
An trying to get openframeworks to build me my application so that i can open it from anywhere and it will load the needed images from within the apps Resources folder.
I believe this is relative linking?
I have done it before, on an older xcode and oF 61.
To do this i dragged the needed images into the project file on the left and added it to the executable target, with in a 'build phase' -> 'copy files'.
Been trying all sorts of methods, ofSetDataPathRoot() which solved the problem last time isnt working for me this time.
Any ideas/help would be appreciated!
Thanks
First you need to tell xCode to copy your /bin/data directory into your application bundle by adding a build phase:
1. Click on your project in the Project Navigator
2. Select "Build Phases"
3. Toggle open the "Run Script" section and paste in the following:
cp -r bin/data "$TARGET_BUILD_DIR/$PRODUCT_NAME.app/Contents/Resources";
Then tell your app where to find the data folder relative to itself within the bundle.
Inside the setup method of your oF app:
ofSetDataPathRoot("../Resources/data/");
ofSetDataPathRoot() should solve this problem. Perhaps you are setting the replacement root path incorrectly?
Try calling ofToDataPath() yourself on a string path and print out the result, then use Terminal and cd inside the .app bundle to check if the path sounds correct. Paths are expressed relative to the location of the actual executable inside the .app bundle, so if the executable is at myApp.app/Contents/MacOS/myApp and the data files are at myApp.app/Contents/Resources then ofToDataPath( "texture.png" ) should return something like ../Resources/texture.png.
You can double-check the current working directory (myApp.app/Contents/MacOS in my example) by calling getcwd(), open up a terminal and type man getcwd for more info on that.
oF now sets data path root and does internal calls to ofToDataPath() by default. What version are you using?
Have you looked inside the product's package contents to make sure your resources are getting copies in the proper build phase?
I installed Mogenerator. Now what do I do? How do I use it?
The first problem I have is that I have no idea where it was installed to. During the install process, it only let me select the hard drive to install it on, not the directory. The most natural location would be the Applications folder, but it isn't there.
Next, the readme (which I found online) states:
Xmo'd works by noticing when your
*.xcdatamodel is saved. If the model file's Xcode project item comment
contains xmod, an AppleScript is fired
that creates a folder based on your
model's file name and populates it
with derived source code files from
your model. It then adds the new
folder to your project as a Group
Reference and adds all the source
files to your project.
There are several issues with the above statement that aren't clear:
What does "the model file's Xcode project item comment" refer to? How can I make it contain "xmod"?
Is adding this comment and having mogenerator monitor the .xcdatamodel file the only way to use mogenerator? Is there any way I can manually run mogenerator so that it recreates the generated files?
One more caveat to be aware of: You have to already set the Class properties of your entities to something different than NSManagedObject. Otherwise Xmo'd won't do anything.
Note: Xmo'd currently doesn't work with Xcode 4/5, afaik.
What I do is just add a "MOGenerator" target in Xcode:
Go to your project and click on "Add Target..." in the "Targets" section.
Choose "iOS -> Other -> Aggregate"
Go to "Build Phases"
Select from the Menu "Editor -> Add Build Phase -> Add Run Script Build Phase"
Paste your MOGenerator command into the Run Script section, for example:
PATH=${PATH}:/usr/local/bin
cd "${PROJECT_DIR}/MyApp"
mogenerator --human-dir Classes --machine-dir MOGenerated --model MyApp.xcdatamodeld/MyApp.xcdatamodel --template-var arc=true
Now you can update your MOGenerator-generated by simply running this target.
mogenerator is a script that is installed into your developer directory as I recall. However it might be installed into the Xcode scripts directory under your ~/Library.
What do you mean by manually triggering the application? You can trigger a build by "touching" the data model. Any save on the data model will trigger the build
In Xcode if you select the model file and hit ⌘I you will get its metadata. Click on the comments tab and add xmod there. mogenerator looks for that comment to know if it should generate files.
Update
You can run mogenerator from the command line as well as have it monitor your files. Type mogenerator --help in the Terminal to see the options.
I searched my hard drive and found the following files:
The application is installed to: /usr/bin/mogenerator.
The /Library/Application Support/mogenerator/ directory contains some .motemplate files.
⌘I doesn't work in Xcode 4 any more. please check out the command line tool. Here is the doc
Studying line 22 of make_installer.command, I found that /Developer/Library/Xcode/Plug-ins/Xmod.pbplugin is also installed.
And then, searching mogenerator GitHub Issues for "uninstall," I found official instructions on how to uninstall mogenerator from the creator himself.
using mogenerator:
download mogenerator
run and build the mogenerator project
locate the built file in the product group
copy the built file in to /usr/bin directory
in the terminal copy this code and hit enter:
mogenerator -m /Users/hashem/Desktop/Projects/myApp/myAppModel.xcdatamodel -O /Users/hashem/Desktop/Projects/myApp/managedObjects --template-var arc=true
NOTE: here first I have entered myApp.xcdatamodel file path, and next path is the location of generated files. if the file path contains space character be sure to add \ character before space in the file path. like /desktop/xcode\ projects/myApp/....
enjoy!
How do I use a relative path in Xcode project settings?
All paths in Build Settings are assumed relative to the directory that contains the .xcodeproj file. Use the standard Unix path tokens
. project directory
.. parent directory
So if your project file is trunk/Mac/proj.xcodeproj, and your headers are in trunk/Headers/foo.h, you would add ../Headers to your Header Search Paths.
Also there are two paths: $SRCROOT and $SDKROOT.
In the upper left corner next to the build/stop buttons, click on the name of your project and Edit Scheme...
In the left column, click on Run
Click on Options
Put a check next to Working Directory: Use custom working directory.
You can then change the relative path to anywhere you want.
EDIT: This is for Xcode 4.1
For Xcode 5:
Click on Product -> Scheme -> Edit Scheme.
Then follow ulu5's instructions: Click "Run", Click on "Options", and check the box "Use custom working directory."
The various answers currently here which recommend setting the working directory when executing a project by editing the scheme and then choosing whatever directory you want are missing what seems like a key part of the question: Relative Path. If you just use the file navigator in the UI you'll get an absolute path, likely with your own home directory in it, which isn't so good if the project you're working on is shared with other people.
To specify a working directory relative to the project folder in there, find the "Working Directory" field in the scheme (In XCode 10.1, that's Product | Scheme | Edit Scheme, then Options, then check "Use Custom Working Directory"), and use $PROJECT_DIR to get the path relative to the project.
Using Xcode 9:
It may be intended for Xcode to always use relative file paths based on the directory that contains the xcodeproj, but sometimes this does not seem to be true, and in my case this may have been due to the fact that the project (directory and all) was copied from an earlier version. I had to do:
Target(top left)->Edit Scheme->Use Custom Working Directory
and then specify to use the directory containing my project file.