My app gets stuck on splash screen in iOS 9 both on iPhone and simulator. I can run it on iOS 8 or lower on device and simulator with no problem. My colleague working on the same app has exactly the same problem.
There is no error or anything, just hangs on splash screen. If I stop it on xcode and try to run it from the phone or simulator directly, it would run without any problem.
By the way, I don't see didFinishLaunchingWithOptions or willFinishLaunchingWithOptions getting called!
In your "answer" you include the code:
+(void)initialize
{
titles = #[NSLocalizedString(#"CODE", nil), NSLocalizedString(#"ERROR", nil), NSLocalizedString(#"TROUBLESHOOTING", nil)];
}
This is indeed the source of your issue. It's wise to be very careful when implementing +load or +initialize. #bbum has a great article on exactly that topic.
+initialize is invoked the first time the class (or category) is touched - when the class is initialized +initialize is called by the class loading mechanism. There is no guarantee of when in the class loading process this may happen, which is part of your problem.
In your case you are using NSLocalizedString - which under the hood can be fairly heavy. It has dependancies on several other classes (NSString, etc) and can potentially access the file system. As #bbum points out in his article, that can lead to serious trouble. In your case, this may be a nasty deadlock.
Move your titles = #[NSLocalizedString... line to a more appropriate place in your object, like an initializer, awakeAfterUsingCoder:, etc. and your immediate problem should be solved. After doing so you should check your entire codebase for instances where +initialize and +load are implemented and audit them to make sure those uses are in line with #bbum 's recommendations.
OK I found the problem. It sounds ridiculous though!!
I am using UITabBarController and inside the first controller I have a UITableViewController with a customised datasource class which would initiate a hard code table header and these headers are localised!!
+ (void)initialize {
titles = #[NSLocalizedString(#"CODE", nil), NSLocalizedString(#"ERROR", nil), NSLocalizedString(#"TROUBLESHOOTING", nil)];
}
After I traced the stacks, I realised the process gets stuck right there with no trace and error! I still don't know why!
So I came up with a workaround:
+ (void)initialize {
titles = #[#"Code",#"Error",#"Troubleshooting"];
}
And only retrieve the localised value when returning the text:
- (NSString *)titleAt:(NSInteger)index {
return NSLocalizedString(titles[index],nil);
}
I have both debug and release set to NO
You sure "any SDK" also has arm64?
Ok, I think I found the answer.
You have to specify arm64 in all "Valid Architectures".
If you don't specify arm64 or forget one the app won't start and stays on the splashscreen.
Just verified this.
Is this an Xcode 7 bug?
I have an Obj-C Project I'm trying to migrate to Swift. I did succeed with various classes but recently ran into an issue I can't seem to make sense of. When I try to compile my current code base I get the following (SUPER UNHELPFUL ERROR MESSAGE)
Command /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/swiftc failed with exit code 1
My only assumption is its somehow related to my bridging-headers but Xcode isn't giving me enough information to figure out if this is actually true.
I'm using Cocoapods to add the CorePlot to my project. I'm trying to migrate the following class to Swift:
Obj-C Class (ScatterPlotContainer.h)
#import <Foundation/Foundation.h>
#class CPTScatterPlot;
#interface ScatterPlotContainer : NSObject
#property (nonatomic, strong) CPTScatterPlot *ahrsAlt;
#property (nonatomic, strong) CPTScatterPlot *calibration;
#property (nonatomic, strong) CPTScatterPlot *coreAlt;
#property (nonatomic, strong) CPTScatterPlot *pitch;
#property (nonatomic, strong) CPTScatterPlot *roll;
#property (nonatomic, strong) CPTScatterPlot *slip;
#end
Obj-c Class (ScatterPlotContainer.m)
#import <CorePlot/CPTScatterPlot.h>
#import "ScatterPlotContainer.h"
#implementation ScatterPlotContainer {
}
#end
Swift Conversion
import Foundation
class ScatterPlotContainer : NSObject {
public var ahrsAlt : CPTScatterPlot;
public var calibration : CPTScatterPlot;
public var coreAlt : CPTScatterPlot;
public var pitch : CPTScatterPlot;
public var roll : CPTScatterPlot;
public var slip : CPTScatterPlot;
}
My bridging headers file
#import <CorePlot/CPTScatterPlot.h>
What I've tried thus far
When I comment out the #import <CorePlot/CPTScatterPlot.h> from the Bridging headers file - I get an error in swift because it doesn't know what CPTScatterPlot is
I've also tried #import <CPTScatterPlot.h> which didn't work either.
Thoughts
So the only thing I can think of is perhaps because I'm using a cocoa pod there is some sort of module name I need to add. The error message really isn't that useful. Does anybody have a suggestion about some blaring error I've made or how to get a more descriptive error message to figure out what is going on?
I did the same all answer says but mine issue was not resolved. I did figured out that issue was related to broken function call.
A function syntax was not wrong but its calling mechanism was wrong.
To check the exact error for this issue check following:
Select issue navigator > Click on error will show logs for error > In that select All Messages tab.
This will show all detail logs for this error.
Scroll down and You got logs like, in my case
So, by reading this I figure out that something wrong with function calling. I browse my code and resolved it, Below was correct and wrong code.
Wrong Way:
var region = MKCoordinateRegionMake(self.mapView.userLocation.coordinate, span)
// It will not shown error here but when you build project compiler shows error.
Right Way:
let region = MKCoordinateRegion(center: self.mapView.userLocation.coordinate, span: span)
I run into this last night and nothing above was solving my problem.
I was about to do something very bad at my laptop when I saw, all by pure luck, that ONE (1) file was is text encoding set to UTF-16 ?!?! WTF??
This was the last file I was working on, and probably, one bad cut/paste "import" a strange character into the arena. I did a cut/paste of my code in this file to a bare bone text editor. I deleted the file, recreate it and paste back my code... and voilà! it work.
So do the above, but also check your file encoding! :-)
I had the same error message.
What helped, was to set the optimization level in the swift compiler settings to None.
This is not really a solution for me and I think that's one of the many bugs in the swift compiler.
Another solution for this issues is to check that you don't have 2 or more files with the same file names. It solved the problem for me.
Thank you #Kampai for the advice on going through the error log message. I read through, and some files were missing:
<unknown>:0: error: no such file or directory:
Somehow, some files were removed during a pull from GitHub. The files are in the directory, but not in the Xcode project.
Right click on a folder and click 'Add files to ...' to manually add missing files to Xcode. That fixed the problem for me.
This happened to me several times already, but now I know how to fix it \o/
I was getting the same error for including this code in a didSet block:
didSet {
// Test whether this view is currently visible to the user.
if super.isViewLoaded() && (super.view.window != nil) {
// (build fails even if this block is empty)
}
}
It took a lot of trial/error to hunt this down. Removing super. allowed the build to proceed.
had a horrible time with this bug for over 3 hours by meticulously going from file to file and reverting the changes and seeing if that file had the issue in it. I tried the first answer but didn't give me any answers. Found the issue and it was because I had a non computed property named the same as a computed property of a subclass. I really hope the debugger becomes more robust with handling these sorts of cases in future updates :(
Simply deleting derived data and cleaning helped me
1) Identify the file there the problem is. You can copy and paste the compilation instruction to the console and the last screen will contain the error description. Note the pid number there the problem was identified. Then scroll up and find the pid and related instruction - there will be one file per pid, so you will find the file you have problem it.
2) Look through the file and check all you last changes. If you have git initialized you can use
git diff <file name>
As for #Kampal, I'm still struggling to figure out how much to specify in a function call. For instance, creating a UIColor object sometimes requires that UIColor be specified, and sometimes doesn't.
These both work:
playButton.backgroundColor = .darkGrayColor()
playButton.setTitleColor(UIColor.whiteColor(), forState: UIControlState.Normal)
This yields the exit code 1 error on compilation, without any debugger warning. #time-sucking debug vortex
playButton.setTitleColor(.whiteColor(), forState: UIControlState.Normal)
So I have a new rule: when using a function that takes more than one parameter, be explicit.
Now back to playing swift: AVOID THE VORTEX
Since everyone else has been showing theirs, I'll show mine:
class Foo : UIView {
var pathPosition:Double = 0.0 { didSet {
pathPosition = min(max(0.0, pathPosition), 1.0) // crashes if this line is present
self.pathPosition = min(max(0.0, pathPosition), 1.0) // but not here
}}
}
Incredibly, this does not come up in Playground, but does fail when placed in code in a framework. Although it is legal syntax (used to work, still works in playground), the Swift compiler seems to want pathPosition to be qualified with self.. Note that is (relatively) old code and used to compile, maybe something broke in 6.1.
Edit:
I feel like I am going insane, but it feels like there is a greater complexity problem going on here, where surrounding code can impact this problem. I saw things compile last night, changed some code and settings again today, and it failed on me again. Today, I had to hack in a bunch of really stupid code to get it to work:
var pathPosition:Double = 0.0 { didSet {
// bug: insane!! - have to clobber the value before resetting!
let bugOldValue = pathPosition
self.pathPosition = 1.0 // fails without this nonsensical line!
self.pathPosition = min(max(0.0, bugOldValue), 1.0)
}}
For what it's worth, the actual error message I got, per the helpful instructions above, was:
PHI node has multiple entries for the same basic block with different incoming values!
%14 = phi double [ 1.000000e+00, %10 ], [ %11, %10 ], [ 1.000000e+00, %9 ], [ 0.000000e+00, %9 ], !dbg !4818
label %10
double 1.000000e+00
%11 = phi double [ %7, %entry ], !dbg !4815
LLVM ERROR: Broken function found, compilation aborted!
I'm scared for tomorrow.
I just had this same error, the problem was that I had overridden a method with a non-optional parameter and had made the parameter optional in the override. (the method parameter below)
func logNetworkCallDurationForMethod(method:String, path:String, milliseconds: UInt) {
}
override func logNetworkCallDurationForMethod(method:String?, path:String, milliseconds: UInt) {
}
Ran into this issue today actually. Was the result of a recent pull from git on a project where a file had been deleted, but it didn't update in my local project.
Clicking on the error brought up the location of the "missing" file, went and deleted it's reference in the Project Navigator. Fixed the error, did a clean, and compiled successfully.
This happened to me when trying to reference a method from an inmutable protocol argument(by mistake, I thought the member was a property):
Having an interface as follows:
public protocol NSValidatedUserInterfaceItem {
func tag() -> Int
}
Compilation crash
func validateUserInterfaceItem(anItem: NSValidatedUserInterfaceItem) -> Bool {
print(anItem.tag) // oopsie, tag is a function
return false
}
Compilation success
func validateUserInterfaceItem(anItem: NSValidatedUserInterfaceItem) -> Bool {
print(anItem.tag()) // this is cool for swift
return false
}
This happened to me and after reading the log in issue navigator I found out that I have two swift files with same name. This was creating the issue and I was getting build failed.
I got this error due to a missing file in my project. Added this file again and voila everything worked.
In my case it was wrong method overriding. Base class:
open func send(_ onSuccess: #escaping ((SomeType) -> Void)) -> SomeType { }
Subclass:
open override func send(_ onSuccess: ((SomeType) -> Void)) -> SomeType { }
As you see #escaping is missing. Swift3 converter in XCode8 doesn't consider inheritance relations, moreover, that type mistakes aren't marked as errors.
In My Case it was Simulator bug just uninstall app from simulator and clean project then run project.
I had accidentally dragged symlinks (aliases) to source files into the project instead of the actual files.
I had CoreData generated files twice (and added myself). Check the files are not duplicate.
Unfortunately this error is often caused by a glitch inside Swift's compiler. It is not always easy to find the reason. If cleaning doesn't work, my suggestion is to try to comment the last code you wrote (even the whole file if necessary). Usually commenting the last code you entered would restore the compilation and you'll get more meaningful errors. From there on you have to try to uncomment the code piece by piece until you get to the instruction which caused this error. The Swift compiler is still pretty young and from time to time it reports weird errors. These kind of errors are completely useless, because instead of helping the developers they only confuse them even more. I would suggest Apple to change the compiler to give more detailed information and avoid this annoying error from appearing anymore.
NSString const *kGreenColor = #"#00C34E";
I had above line in my Constant.h file. which was meant for preprocessors only.
Removing that line worked for me.
moving the Bridge file to project level resolve my problem.
In my case I had renamed a file. After committing I found that the file name still hasn't changed in the Xcode project (not sure why), the file was greyed out. Changing the name and committing again did the trick.
So we have to keep an eye out for this error, when making changes to files using source control.
In my case deleted a couple of files directly from SourceTree but their reference was still there in Xcode. Logs show their names. Removed them and error went away.
I'm having a really weird problem here with NSUInteger in iOS7,
everything is perfect before iOS7, I guess it's related to the 64-bit support in iOS7.
My code is like this, very simple:
if (blah blah blah) {
NSUInteger firstRow = 0;
firstRow = ([self.types containsObject:self.selectedMajorType] ?
[self.types indexOfObject:self.selectedMajorType] + 1 : 0);
...
}
According to my console,
[self.types containsObject:self.selectedMajorType] is true
[self.types indexOfObject:self.selectedMajorType]+1 is 1,
no doubt, and indexOfObject also returns an NSUInteger (according to Apple's document),
here's the screenshot:
but firstRow is always fking **0
This is so creepy I don't know what's going on with NSUInteger,
can someone help me? Thanks a lot!!
____new finding____
I guess this is the problem? It's weird..
I tried to recreate this scenario but I was always getting the expected result 1.
Here is the screen shot:
Here is the project, try running this and see if you still face the problem.
PS. I was using xcode 5.1 and iPhone 64bit Simulator.
=============UPDATE================
Here are some explanations on the lldb commands you used.
po : prints the objective C description of an object.
print / p : Evaluates a generalized expression in the current frame. Specify return type of the function if not used in your program.
Hope this screenshot will help you understand more.
After completing my app I run a test from Xcode, Product -> Test, it says 'Build Succeeded' and the app pops up for a split second, but then it prompt me with this message:
I have searched a lot but couldn't find any solution that works in this case. I want to mention that I have also tried changing the name of the application from, CSQTCConference, to CSQTC Conference, not sure how relevant this is.
I am planning to release my application to app store today, but this issue is holding it back. It will be helpful if you can suggest any pointers to resolve this issue.
I just comment the following line:
//XCTFail(#"No implementation for \"%s\"", __PRETTY_FUNCTION__);
in:
#import <XCTest/XCTest.h>
#interface KinderAppTests : XCTestCase
#end
#implementation KinderAppTests
- (void)setUp
{
[super setUp];
// Put setup code here. This method is called before the invocation of each test method in the class.
}
- (void)tearDown
{
// Put teardown code here. This method is called after the invocation of each test method in the class.
[super tearDown];
}
- (void)testExample
{
//XCTFail(#"No implementation for \"%s\"", __PRETTY_FUNCTION__);
}
#end
This solution worked for me, hope it helps :)
I'm not familiar with XCTest but I assume XCTFail will always fail, so no big surprise here.
You are running boilerplate example code, which created a test that always fails. If you haven't written your own tests what's the point of running them in the first place?
The real question is: Why is this happening?
Answer: You did a Command U, or selected test. This ran the default test, which always fails.
Correction: Comment out the fail line, and run it again.