I'm very new to cmake and I'm trying to build os x application using it, the problem I've faced is that I need to specify NSPrincipalClass inside Info.plist to be able to launch it. (otherwise I'm getting "No Info.plist file in application bundle or no NSPrincipalClass in the Info.plist file" error). The problem is that there is no way to do it using cmake MACOSX_BUNDLE_xxx variables.
I probably can use some workarounds like script that will add it after building or something like this but is there a better way? I might also modify default plist template but than is there a way to include it inside project repo so there will be no need to modify it on every machine or after reinstalling cmake etc?
I've solved this issue. I've created my own template (file name is MacOSXBundleInfo.plist.in) put it into cmake/modules/ folder inside project root, from there the only think left is to add
set(CMAKE_MODULE_PATH ${CMAKE_MODULE_PATH} "${CMAKE_SOURCE_DIR}/cmake/Modules/")
into your CMakeLists.txt
I'm getting the following error:
ld: library not found for -lGoogleAnalytics
clang: error: linker command failed with exit code 1 (use -v to see invokation)
I've spent some time googling but can't find how to fix this problem. I'm new to xcode and this is an existing project that I need to work on.
In my case, the project uses CocoaPods. And some files are missing from my project.
So I install it from CocoaPods: https://cocoapods.org/.
And if the project uses CocoaPods we've to be aware to always open the .xcworkspace folder instead of the .xcodeproj folder in the Xcode.
You need to set the "linker search paths" of the project (for both Debug and Release builds). If this library was in, say, a sibling directory to the project then you can set it like this:
$(PROJECT_DIR)/../GoogleAnalytics/lib
(you want to avoid using an absolute path, instead keep the library directory relative to the project).
All in all, the Xcode cannot find the position of library/header/framework, then you tell Xcode where they are.
set the path that Xcode use to find library/header/framework in Build Settings --> Library/Header/Framework Search Paths.
Say, now it cannot find -lGoogleAnalytics, so you add the directory where -lGoogleAnalytics is to the Library Search Paths.
In my case I had a project with lots of entries in "Build Settings > Other Linker Flags"
I needed to reduce it down to just
$(inherited)
-ObjC
Old settings:
Updated settings:
For me, I open the projectname.xcworkspace file and it all works.
If you have pods installed, make sure to open the workspace folder (white Xcode icon) not the project folder. This resolved the library not found for ... error. Very simple issue but I was stuck on this for a long time.
This worked for me:
Go to build setting -> Linking -> Other Linker Flags -> Remove all other than $(inherited)
Cd ios && pod update
If you are using Pods to include the GoogleAnalytics iOS SDK into your project, it's worth noting that since the 3.0 release your Other Linker Flags needs to include -lGoogleAnalyticsServices not the old -lGoogleAnalytics
If your library file is called libGoogleAnalytics.a you need to put -lGoogleAnalytics so make sure the .a file is named as you'd expect
None of the above worked for me, what did was making sure the Pod file platform :ios, '11.0' matched with the minimum deployment target in the XCODE setting
You can also try to lint with the --use-library option, as cocoapods lint libraries as framework by default since v0.36
The problem might be the following: SVN ignores .a files because of its global config, which means someone didn't commit the libGoogleAnalytics.a to SVN, because it didn't show up in SVN. So now you try to check out the project from SVN which now misses the libGoogleAnalytics.a (since it was ignored and was not committed). Of course the build fails.
You might want to change the global ignore config from SVN to stop ignoring *.a files.
Or just add the one missing libGoogleAnalytics.a file manually to your SVN working copy instead of changing SVNs global ignore config.
Then re-add libGoogleAnalytics.a to your XCode project and commit it to SVN.
In XCode 10.1, I had to set "Library Search Paths" to something like $(PROJECT_DIR)/.../path/to/your/library
For me it was a silly thing: my mac uploaded the file into iCloud, and that is why Xcode did not find it.
If you turn off the automatic upload, it wont happen again.
I currently have a project that I'm building with a makefile. This project includes some additional software (jflex) which is not on one of the default system search paths (it's installed to (/opt/local/bin/). My .profile file adds this directory to my PATH, and so building the project from a terminal window succeeds.
However, if I try to run the makefile from within XCode (XCode project with an externally managed makefile), it fails since it's not looking in /opt/local/bin for jflex, and as such can't find it.
How can I change the settings of my XCode project to correctly build from within XCode? I assume there's some kind of path setting in XCode, or some kind of additional argument I can have XCode give to make so that it doesn't fail. Alternately, I could hardcode the path into my makefile if I could detect that it was being run from XCode (although this is the less preferable option for me, since my makefile will sometimes need to be called from the command line and I'd rather keep it simple).
I'm relatively new to using XCode, so apologies if I've missed something obvious.
Instead of calling make directly, you could call a shell script as the "Build Tool" in the "External Build Tool Configuration" pane. Then modify the path in the shell script and call make from there, i.e.
#!/bin/sh
PATH=/opt/local/bin:$PATH
make
Don't forget to set permissions such that Xcode can run the script, and provide the full path to the script as the "Build Tool".
May not be the best way, But can you launch XCode from command line. It will inherit the Path from it.
Or, in the XCode launcher change:
XCode
to
PATH=$PATH:/opt/local/bin/ XCode
may work (depending on launcher)
Or, did you relaunch the launcher/window manager (logout and back in again after setting .profile)?
I'm currently trying to use CMake to generate an Xcode project file, but I would like to undefine certain user-defined settings that CMake generates automatically when making an Xcode project. More specifically, I want to undefine WARNING_CFLAGS that CMake wants to set for my build targets in Xcode. I've tried the following in my CMakeLists.txt file:
set(CMAKE_XCODE_ATTRIBUTE_WARNING_CFLAGS "")
However, that has no effect at all. Any advice?
Try this:
set_target_properties(<target> PROPERTIES XCODE_ATTRIBUTE_WARNING_CFLAGS "")
I managed to get my C++ project running in Xcode using cmake -G Xcode. Now I would like to debug it, because of the nice gdb fronted it provides.
However I need to change the working directory where gdb is executed otherwise it wont work. (It needs to read some data from files relative to the path of the working directory) What I can do is after gdb has stopped typing cd working_dir and then run it manually typing r. This works, but Xcode will complain constantly and is also not an elegant solution.
So somewhere in the build settings I should be able to set the working directory..but where?!
Just updating this solution to XCode 4.x
You go on Product -> Scheme -> Edit Scheme
On the tab Options you choose your working directory.
In Xcode 4 the working directory settings is under Scheme Settings.
In Xcode 3.x do a "Get Info" on the executable and look at the first tab - there is a setting there for working directory.
The default debug working directory is somewhere under ~/Library/Developer/Xcode/DerivedData/project-/Build/Products/Debug. You're better off using the full path.