I have written a couple Core Data iPhone apps and I've run into a consistent problem that I would expect to be a common pattern for iPhone development. Namely, I have one or more command line data loaders to create the Core Data database and then I have my iPhone app. Since the model files are shared and the loaders are a integral part of the project, I want it to be a single project.
What is the right approach to create a project with an iPhone OS target and a command line target? All my attempts have not worked well -- I have code signing, debugging, and configuration troubles.
Here's my best attempt so far:
At this point I have no executable. (By contrast, if you create a new Command Line Utility project then an executable is automatically associated with your target.)
If I create a Custom Executable then I have a few troubles: (1) changing the target doesn't change the executable because you can't link a custom executable to a target, (2) debugging the command line program is problematic because breakpoints set in XCode don't get registered (the chevrons at the line numbers turn orange when debugging the executable), so I have to break into the running code and then handle debugging manually from the console, (3) most frustratingly, if I create a properly provisioned distribution version of my iPhone app, then transferring the iPhone app to the device fails complaining that the provisioning profile doesn't match the Code Signing Entitlement. This is definitely not the case because I set the provisioning profile and entitlement file for only the distribution version of the iPhone target, but no matter how much fiddling I can't get the target installed on devices. (If I create a new project with identical code and code signing withOUT the command line tool then I can install an AdHoc provisioned app.)
So... is there some other pattern I should be following? Xcode configurations are serious voodoo.
Go to Build Settings and search for the app icon, and set the app icon in Asset Catalog App Icon Set Name, and here you can set resources for each configuration. This completes the process to create different environments and manage Xcode environment variables configuration.
Deployment Target refers to the oldest version of iOS that is capable of running your project. To change your deployment target, open up your project file in Xcode and check the setting under Build Settings -> Deployment(...) Check this answer to add earlier devices support.
Overview. Multiplatform apps broaden the experience of your app to each additional platform you support. In Xcode 14 or later, you can share your app's project settings and code across platforms using a single, multiplatform target.
OK, I think your problem need to be redefined.
As you mentioned:
Namely, I have one or more command line data loaders to create the Core Data database
Your requirement is to pre-populate the database for your core data iPhone app, and also, you want to let the tool to keep consistency with your latest Core Data models. Am I right?
If my understanding is correct. I suppose, instead of creating a dedicated command line tool to pre-populate the database, do it in the Unit Test target will be a better option.
Here is what I'm doing with my core data iOS app.
2 and 3 can be integrated into one testcase if you don't have much data to import.
⌘ + k
to remove everything previously created.⌘ + u
to run the testcase again and repopulate the database with data.[NSBundle bundleForClass:{Your testcase class name}]
to get the correct bundle path and thus you can load the Core Data Model and create the MOC instance. You can see how I do this in my open source project on github
Here is a brief code from my testcase.
- (void)setUp
{
[super setUp];
// Set-up CoreData
yourMocObj = ...;
}
- (void)tearDown
{
// Tear-down code here.
[yourMocObj save];
[super tearDown];
}
- (void)test_PopulateData{
// Get the path for the data file.
NSString *plistPath = [[NSBundle bundleForClass:[algorithms_iOSTests class]] pathForResource:@"AlgorithmsMasterViewData" ofType:@"plist"];
NSArray *dataArray = [NSArray arrayWithContentsOfFile:plistPath];
NSEntityDescription *yourEntity = [NSEntityDescription entityForName:.. inManagedObjectContext:yourMocObj];
[dataArray enumerateObjectsUsingBlock:^(NSDictionary *aCategoryDict, NSUInteger idx, BOOL *stop) {
// Category is a custom subclass of NSManagedObject.
Category *aCategory = [[Category alloc] initWithEntity:yourEntity insertIntoManagedObjectContext:yourMocObj];
[aCategory safelySetValuesForKeysWithDictionary:aCategoryDict];
}];
}
From my own experience, if your project starts off as an iOS project you are in for a world of pain, as it sounds like you have discovered. Sure, you can add OS X targets to your iPhone project, but the built executable isn't automatically added in Xcode. You can then add a custom executable, but the configuration lives in the user-specific project files which can be a dealbreaker if you're using source control and collaborating with other people. Also, as you have discovered, it doesn't integrate as well as it should with the debugger, etc.
If you want a more manageable project, I'd recommend starting out with an OS X project and then adding your iOS targets. This has worked much better for me.
This might be improved in Xcode 4, but in 3.2 these issues have been a real drag. At least you know you're not alone.
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With