I need to install with carthage the package https://github.com/socketio/socket.io-client-swift
with a Xcode 12 project.
i have a CartFile :
github "socketio/socket.io-client-swift" ~> 15.0.0
I have tried this command :
carthage update --platform iOS --use-xcframeworks
I try :
modify the "Buid archicture only" on "NO" in my main project
modify the "Buid archicture only" on "NO" in the dependency "Starscream" project
but always i have the same error :
CompileSwift normal i386 (in target 'SocketIO' from project 'Socket.IO-Client-Swift')
cd /Users/admin/Documents/test2/Carthage/Checkouts/socket.io-client-swift
/Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/swift -frontend -c /Users/admin/Documents/test2/Carthage/Checkouts/socket.io-client-swift/Source/SocketIO/Engine/SocketEngine.swi$
/Users/admin/Documents/test2/Carthage/Checkouts/socket.io-client-swift/Source/SocketIO/Engine/SocketEngine.swift:27:8: error: could not find module 'Starscream' for target 'i386-apple-ios-simulator'; found: arm64, armv7-$
import Starscream
Maybe i need to create a ${ARCHS_STANDARD} in some env or/and custom file ?
I found that Carthage doesn't set the FRAMEWORK_SEARCH_PATHS correctly in many cases, which for me caused Socket.IO to be unable to find Starscream when compiling.
Here's a workaround script you can use to set the FRAMEWORK_SEARCH_PATHS:
# carthage.sh
# Usage example: ./carthage.sh build --use-xcframeworks --platform iOS
set -euo pipefail
xcconfig=$(mktemp /tmp/static.xcconfig.XXXXXX)
trap 'rm -f "$xcconfig"' INT TERM HUP EXIT
echo "FRAMEWORK_SEARCH_PATHS=\$(inherited) $(pwd)/Carthage/Build" >> $xcconfig
# echo 'BUILD_LIBRARY_FOR_DISTRIBUTION=YES' >> $xcconfig # You may or may not need this depending on your project setup
export XCODE_XCCONFIG_FILE="$xcconfig"
carthage "$#"
Related
Someone in my company created a Swift package SDK and now I was tasked to publish it for the customer in a binary way so that the end customers that will use the SDK will not be able to see the source code of it. This is how the SDK is built:
the SDK in xcode (p.s. the build folder is empty)
From my reading on the subject I understand that I need to export the files into an XCFramework file. However, the vast majority of guides I've encountered explain how to make this progress from a framework, and not from a package like in my case..
The only guide I found that seems exactly like what I need is this one, however I get an error right on the first relevant terminal command of xcodebuild -scheme [my scheme name] -sdk iphoneos -configuration Release ARCHS="arm64" BUILD_DIR="./Build". This is the main error line I get: xcodebuild: error: Building a Swift package requires that a destination is provided using the "-destination" option. The "-showdestinations" option can be used to list the available destinations.. Why would I need to specify a destination? I want the SDK to work for all the devices (ios 13+). Non of the forums I searched in online helped me solve this.
I also read Apple's instructions here but got very confused about how the terminal command in step 2 is supposed to look like in my case. Are some of the fields mandatory and some are not?
Any help would be much appreciated!!
Assuming you are working with iOS only and you need an xcframework for both device and simulator architectures, in order to generate an XCFramework from a swift package you need to:
Mark your Package as .dynamic (i.e. .library(name: "Foo", type: .dynamic, targets: ["Foo"]))
Archive the project for both simulator and device. This will generate a .framework file for each architecture.
Copy Modules folders (if any) into the .xcarchive files
Copy the bundles (if any) into the .xcarchive files
Create the xcframework with the frameworks created in step 2
Here's a bash script, based on the one from this swift forums post, to create an XCFramework from a swift package (in my case, I have my package inside an xcworkspace. I didn't try it as a standalone package, not sure if that can be done):
Change the input parameters as needed :), but most importantly replace <Your project name> and <your workspace>.
#!/bin/bash
PROJECT_NAME="<Your project name>"
PROJECT_DIR="./Packages/${PROJECT_NAME}" # Relative path to the directory containing the `Package.swift` file
BUILD_FOLDER="./build"
OUTPUT_DIR="${PROJECT_DIR}/Output"
SIMULATOR_ARCHIVE="${OUTPUT_DIR}/${PROJECT_NAME}-iphonesimulator.xcarchive"
DEVICE_ARCHIVE="${OUTPUT_DIR}/${PROJECT_NAME}-iphoneos.xcarchive"
rm -rf "$OUTPUT_DIR"
mkdir -p "$OUTPUT_DIR"
# 2 iterations: 1 for device arch and another for simulator arch
for PLATFORM in "iOS" "iOS Simulator"; do
case $PLATFORM in
"iOS")
ARCHIVE=$DEVICE_ARCHIVE
SDK=iphoneos
RELEASE_FOLDER="Release-iphoneos"
;;
"iOS Simulator")
ARCHIVE=$SIMULATOR_ARCHIVE
SDK=iphonesimulator
RELEASE_FOLDER="Release-iphonesimulator"
;;
esac
# Step 2
xcodebuild archive \
-workspace <your workspace>.xcworkspace \
-scheme $PROJECT_NAME \
-destination="generic/platform=${PLATFORM}" \
-archivePath $ARCHIVE \
-sdk $SDK \
-derivedDataPath $BUILD_FOLDER \
SKIP_INSTALL=NO \
BUILD_LIBRARY_FOR_DISTRIBUTION=YES
FRAMEWORK_PATH="${ARCHIVE}/Products/Library/Frameworks/${PROJECT_NAME}.framework"
MODULES_PATH="$FRAMEWORK_PATH/Modules"
mkdir -p $MODULES_PATH
BUILD_PRODUCTS_PATH="${BUILD_FOLDER}/Build/Intermediates.noindex/ArchiveIntermediates/${PROJECT_NAME}/BuildProductsPath"
RELEASE_PATH="${BUILD_PRODUCTS_PATH}/${RELEASE_FOLDER}"
SWIFT_MODULE_PATH="${RELEASE_PATH}/${PROJECT_NAME}.swiftmodule"
RESOURCES_BUNDLE_PATH="${RELEASE_PATH}/${PROJECT_NAME}_${PROJECT_NAME}.bundle"
# Step 3
if [ -d $SWIFT_MODULE_PATH ]
then
cp -r $SWIFT_MODULE_PATH $MODULES_PATH
fi
# Step 4
if [ -e $RESOURCES_BUNDLE_PATH ]
then
cp -r $RESOURCES_BUNDLE_PATH $FRAMEWORK_PATH
fi
done
# Step 5
xcodebuild -create-xcframework \
-framework "${DEVICE_ARCHIVE}/Products/Library/Frameworks/${PROJECT_NAME}.framework" \
-framework "${SIMULATOR_ARCHIVE}/Products/Library/Frameworks/${PROJECT_NAME}.framework" \
-output "${OUTPUT_DIR}/${PROJECT_NAME}.xcframework"
When the script ends, you will see 3 files in the output folder:
Device xcarchive
Simulator xcarchive
XCFramework
You can remove the xcarchives, as you won't need them anymore (you can also update the script to do that for you).
i am new in react native i want to run react native app in IOS after react-native init, app not run in IOS show some error
Showing All Messages
bash: Native/social_login/socialLogin/node_modules/react-native/scripts/../Libraries: No such file or directory
Command PhaseScriptExecution failed with a nonzero exit code
and build failed please help me
versions
"react": "17.0.1",
"react-native": "0.64.0"
command line tools : Xcode 12.4 (12D4e)
This happened also to me upgrading from 0.63 to 0.64. After trying all solutions I followed a solution moving the folder to a directory where the path contain no spaces and it works and build the app successfully.
Solution to React Native 0.64 build fail
In order for this to work properly follow these steps:
If you previously installed a global react-native-cli package, please
remove it as it may cause unexpected issues (i.e. npm uninstall -g
react-native-cli)
Move the project folder in a path with no spaces (i.e. ~/sub folder
name/ReactNativeApp won't work till you have spaces in the path, so
move in a path like ~/folder/ReactNativeApp)
Then cd into the project folder and upgrade react native to the
latest version with npx react-native upgrade and resolve conflicts if
any
After upgrading remove the node_modules folder and the yarn.lock from
the root and the podfile.lock and Pods folder from ios subfolder
Then cd back to the root and run yarn install && npx pod-install
Now run again your app in Xcode or your IDE and it works
Crazy and absurd that a space in the path-name could cause this issue
This is nothing just an issue with the scheme name for me, in my case my scheme name contains whitespace e.g. "ABC staging", which is not allowed, it got fixed after deleting and creating a new scheme with the name "ABC-staging".
in the case of react-native 0.67^, you may need to apply these changes as well
change line no 7
set -e. ==> set +e
in
node_modules/react-native/scripts/find-node.sh
then use patch-package using the following command :
npx patch-package react-native
this will create a patch file inside the patch folder in root, then you can add in the post-install script in package.json:
"postinstall": "npx patch-package"
This command will run each time a new package is getting added to the project and auto-fix react-native find-node. sh.
Try to run pod install in ios folder
cd ios && pod install
Then when it's done go back to your main folder and run
yarn run ios
If that doesn't work, check out the solutions here
Xcode 10.2.1 Command PhaseScriptExecution failed with a nonzero exit code
There seems to be an issue with coreutils on macOS.
What fixed it for me is:
brew install coreutils
brew install findutils
brew install gnu-sed
Finally change the node_modules/react-native/scripts/generate-specs.sh to:
#!/bin/bash
# Copyright (c) Facebook, Inc. and its affiliates.
#
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
# This script collects the JavaScript spec definitions for core
# native modules and components, then uses react-native-codegen
# to generate native code.
#
# Optionally, set these envvars to override defaults:
# - SRCS_DIR: Path to JavaScript sources
# - CODEGEN_MODULES_LIBRARY_NAME: Defaults to FBReactNativeSpec
# - CODEGEN_MODULES_OUTPUT_DIR: Defaults to React/$CODEGEN_MODULES_LIBRARY_NAME/$CODEGEN_MODULES_LIBRARY_NAME
# - CODEGEN_COMPONENTS_LIBRARY_NAME: Defaults to rncore
# - CODEGEN_COMPONENTS_OUTPUT_DIR: Defaults to ReactCommon/react/renderer/components/$CODEGEN_COMPONENTS_LIBRARY_NAME
#
# Usage:
# ./scripts/generate-specs.sh
# SRCS_DIR=myapp/js CODEGEN_MODULES_LIBRARY_NAME=MySpecs CODEGEN_MODULES_OUTPUT_DIR=myapp/MySpecs ./scripts/generate-specs.sh
#
# shellcheck disable=SC2038
set -e
THIS_DIR=$(cd -P "$(gdirname "$(greadlink "${BASH_SOURCE[0]}" || echo "${BASH_SOURCE[0]}")")" && pwd)
TEMP_DIR=$(gmktemp -d /tmp/react-native-codegen-XXXXXXXX)
RN_DIR=$(cd "$THIS_DIR/.." && pwd)
# find node path
source "$RN_DIR/scripts/find-node.sh"
NODE_BINARY="${NODE_BINARY:-$(command -v node || true)}"
USE_FABRIC="${USE_FABRIC:-0}"
cleanup () {
set +e
grm -rf "$TEMP_DIR"
set -e
}
describe () {
printf "\\n\\n>>>>> %s\\n\\n\\n" "$1"
}
main() {
SRCS_DIR=${SRCS_DIR:-$(cd "$RN_DIR/Libraries" && pwd)}
CODEGEN_MODULES_LIBRARY_NAME=${CODEGEN_MODULES_LIBRARY_NAME:-FBReactNativeSpec}
CODEGEN_COMPONENTS_LIBRARY_NAME=${CODEGEN_COMPONENTS_LIBRARY_NAME:-rncore}
CODEGEN_MODULES_OUTPUT_DIR=${CODEGEN_MODULES_OUTPUT_DIR:-"$RN_DIR/React/$CODEGEN_MODULES_LIBRARY_NAME/$CODEGEN_MODULES_LIBRARY_NAME"}
# TODO: $CODEGEN_COMPONENTS_PATH should be programmatically specified, and may change with use_frameworks! support.
CODEGEN_COMPONENTS_PATH="ReactCommon/react/renderer/components"
CODEGEN_COMPONENTS_OUTPUT_DIR=${CODEGEN_COMPONENTS_OUTPUT_DIR:-"$RN_DIR/$CODEGEN_COMPONENTS_PATH/$CODEGEN_COMPONENTS_LIBRARY_NAME"}
TEMP_OUTPUT_DIR="$TEMP_DIR/out"
SCHEMA_FILE="$TEMP_DIR/schema.json"
if [ -z "$NODE_BINARY" ]; then
echo "Error: Could not find node. Make sure it is in bash PATH or set the NODE_BINARY environment variable." 1>&2
exit 1
fi
CODEGEN_PATH=$("$NODE_BINARY" -e "console.log(require('path').dirname(require.resolve('react-native-codegen/package.json')))")
# Special case for running CodeGen from source: build it
if [ ! -d "$CODEGEN_PATH/lib" ]; then
describe "Building react-native-codegen package"
bash "$CODEGEN_PATH/scripts/oss/build.sh"
fi
describe "Generating schema from flow types"
"$NODE_BINARY" "$CODEGEN_PATH/lib/cli/combine/combine-js-to-schema-cli.js" "$SCHEMA_FILE" "$SRCS_DIR"
describe "Generating native code from schema (iOS)"
pushd "$RN_DIR" >/dev/null || exit 1
"$NODE_BINARY" scripts/generate-specs-cli.js ios "$SCHEMA_FILE" "$TEMP_OUTPUT_DIR" "$CODEGEN_MODULES_LIBRARY_NAME"
popd >/dev/null || exit 1
describe "Copying output to final directory"
gmkdir -p "$CODEGEN_COMPONENTS_OUTPUT_DIR" "$CODEGEN_MODULES_OUTPUT_DIR"
gcp -R "$TEMP_OUTPUT_DIR/$CODEGEN_MODULES_LIBRARY_NAME.h" "$TEMP_OUTPUT_DIR/$CODEGEN_MODULES_LIBRARY_NAME-generated.mm" "$CODEGEN_MODULES_OUTPUT_DIR" || exit 1
gfind "$TEMP_OUTPUT_DIR" -type f | gxargs gsed -i.bak "s/$CODEGEN_MODULES_LIBRARY_NAME/$CODEGEN_COMPONENTS_LIBRARY_NAME/g" || exit 1
gfind "$TEMP_OUTPUT_DIR" -type f -not -iname "$CODEGEN_MODULES_LIBRARY_NAME*" -exec cp '{}' "$CODEGEN_COMPONENTS_OUTPUT_DIR/" ';' || exit 1
echo >&2 'Done.'
}
trap cleanup EXIT
main "$#"
notice that some commands are starting with g like greadlink etc.
if your get invalid identifier error in react-native-xcode.sh then under Build Phases -> Bundle React Native code and images it should be: (notice the double quotes):
set -e
export NODE_BINARY="node ../node_modules/react-native/scripts/react-native-xcode.sh"
use patch-package react-native to patch it (if project is being developed on different machines then they must install the brew packages above)
For me it was just having a space in folder name which was in the path of project folder from root.
I had to delete the contents of my ~/.bash_profile file.
I don't even use bash, but some other script populated the file and that broke my build. Deleting the contents fixed the build, immediately. Might not work for you, but thought I'd share.
I'm running Xcode 13.4.1
Installed cocoapods using brew install cocoapods (https://brew.sh/index_es)
M2 apple chip
React Native 0.70
Actually after spending a whole lot of time trying different solutions none of them worked except for Giuseppe's answer. I had white spaces in my path
"/React Native/exampleProject"
1. Renamed folders/files to remove any white space in the path (creating a fresh project within the fixed path worked with no problems at all)
I didn't even had to execute pod install anymore in the ios folder
In my case, i had to export the right path to node
in your terminal type which node, copy the path and export it in ios/xcode.env
export NODE_BINARY="copied node path"
Ok, worth trying this.
(This usually happens if you had multiple imports or any imports missing)
As soon as we get this error message:
Always scroll up & read the issue if written in that log file. Any issues like syntax error or issue related to your js code.
If yes, you can fix that first & re build it.
In my case I had multiple imports of one of the RN components.
Adding $(ARCHS_STANDARD) to Valid Architectures in Build Settings solved it for me
As showed in the screen above, I couldn't find any xcodebuild command line to add/remove the embed frameworks, is there any way we can do so via command line (to remove or add the line)? As I need to automate the build process.
For conditionally embedding frameworks to your app you should make a script under Build Phases that copies your dynamic library to Frameworks folder of your app and then sign it e.g.:
# Path to your framework to copy. You must have two compiled versions: Debug(arm64, x86_64) and Release(arm64).
FRAMEWORK=".../$CONFIGURATION/MyFramework.framework"
# Frameworks folder of your app
NAME=$(basename $FRAMEWORK)
DESTINATION=${BUILT_PRODUCTS_DIR}/${FRAMEWORKS_FOLDER_PATH}/$NAME
# Copy (if needed) and sign
if [ ! -d $DESTINATION ]; then
cp -r $FRAMEWORK $DESTINATION
codesign --force --sign ${EXPANDED_CODE_SIGN_IDENTITY} ${OTHER_CODE_SIGN_FLAGS:-} --preserve-metadata=identifier,entitlements $DESTINATION
fi
For xcframework you can specify subfolder path in additional for a target arch:
# Release device only
if [ $PLATFORM_PREFERRED_ARCH == "arm64" ]; then
FRAMEWORK=".../MyFramework.xcframework/ios-arm64/MyFramework.framework"
fi
We have integrated OCLint in one of our project with below script.
source ~/.bash_profile
export PATH=$PATH:/usr/local/bin/
if [ -z "${SCHEME+x}" ]
then
export SCHEME="${PROJECT_NAME}"
fi
if [ -z "${WORKSPACE+x}" ]
then
export WORKSPACE="${PROJECT_NAME}.xcworkspace"
fi
cd "${SOURCE_ROOT}"
# Check if xctool and oclint are installed
if ! which -s xctool
then
echo 'error: xctool not found, install e.g. with homebrew'
exit 1
fi
if ! which -s oclint-json-compilation-database
then
echo 'error: OCLint not installed, install e.g. with homebrew cask'
exit 2
fi
# Cleanup before building
rm -f compile_commands.json
xctool -workspace "${WORKSPACE}" -scheme "${SCHEME}" clean > /dev/null
# Build and analyze
# OCLint Rule Index: http://docs.oclint.org/en/dev/rules/index.html
xctool -workspace "${WORKSPACE}" -scheme "${SCHEME}" -reporter json-compilation-database:compile_commands.json build
oclint-json-compilation-database -e Pods -- -max-priority-1=100000 -max-priority-2=100000 -max-priority-3=100000 \
# Final cleanup
rm -f compile_commands.json
It gives 'zero' warnings and 'zero' error though code contains the lots of warnings and errors.
It seems that it was working fine with old XCodes but not working with XCode 9.2.
Can anyone tell us what kind of changes required to run the OCLint in xcode 9.2 perfectly ?
We are expecting that xcode will show the warnings when we build the OCLint schema. But we are not getting any result now.
Please guide us if we are doing something wrong.
We faced a similar challenge few days back when we upgraded from xcode 8.2.1 to xcode 9.2 with oclint 0.11
Can you try upgrading to oclint-0.13.1 and do a retry and see if it helps.
There are some other tweaks also present but not sure that will fit or not ?
https://github.com/oclint/oclint/issues/245
(search texts for -fmodules and -gmodules, and delete all matches.)
https://github.com/oclint/oclint/issues/302
(set CLANG_ENABLE_MODULE_DEBUGGING=NO)
I am building a static library. The build setting has the Architectures set to: $(ARCHS_STANDARD) which is shown as Standard Architectures (armv7, armv7s, arm64) I build the lib choosing iOS Device AND then using the simulator (for example iPhone Retina).
Now that I have two builds (one inside Debug-iphoneos and the other inside Debug-iphonesimulator, I use lipo -create to create the aggregated lib:
lipo -create path/to/first/lib /path/to/second/lib -o MyLib.a
If I used this library in another project to simulate on any iOS device with 64-bit architecture, it gives symbol(s) not found for architecture x86_64. What really makes me so angry that the lib project itself is inside a workspace with another project that use the lib. I can simulate on 64-bit iOS simulator! (on all simulators and devices for that matter). What am I doing wrong?
Notes:
This is not duplicate Q. Before accusing me of that (since this is my second day trying to fix this stupid issue), I did search on Stack and Google. All answers don't help.
I am using Xcode 5.1.1.
I had the same trouble with building static library.
Finally I have found the basic solution. (You need to build universal library for x86_64/armv7/armv7s/arm64)
1) Click on the project file
2) Click on the target
3) Open "Build Phases"
4) Open "Run Script"
5) Add "/bin/sh" and the script below
##########################################
#
# c.f. http://stackoverflow.com/questions/3520977/build-fat-static-library-device-simulator-using-xcode-and-sdk-4
#
# Version 2.7
#
# Latest Change:
# - Supports iPhone 5 / iPod Touch 5 (uses Apple's workaround to lipo bug)
#
# Purpose:
# Automatically create a Universal static library for iPhone + iPad + iPhone Simulator from within XCode
#
# Author: Adam Martin - http://twitter.com/redglassesapps
# Based on: original script from Eonil (main changes: Eonil's script WILL NOT WORK in Xcode GUI - it WILL CRASH YOUR COMPUTER)
#
set -e
set -o pipefail
#################[ Tests: helps workaround any future bugs in Xcode ]########
#
DEBUG_THIS_SCRIPT="false"
if [ $DEBUG_THIS_SCRIPT = "true" ]
then
echo "########### TESTS #############"
echo "Use the following variables when debugging this script; note that they may change on recursions"
echo "BUILD_DIR = $BUILD_DIR"
echo "BUILD_ROOT = $BUILD_ROOT"
echo "CONFIGURATION_BUILD_DIR = $CONFIGURATION_BUILD_DIR"
echo "BUILT_PRODUCTS_DIR = $BUILT_PRODUCTS_DIR"
echo "CONFIGURATION_TEMP_DIR = $CONFIGURATION_TEMP_DIR"
echo "TARGET_BUILD_DIR = $TARGET_BUILD_DIR"
fi
#####################[ part 1 ]##################
# First, work out the BASESDK version number (NB: Apple ought to report this, but they hide it)
# (incidental: searching for substrings in sh is a nightmare! Sob)
SDK_VERSION=$(echo ${SDK_NAME} | grep -o '.\{3\}$')
# Next, work out if we're in SIM or DEVICE
if [ ${PLATFORM_NAME} = "iphonesimulator" ]
then
OTHER_SDK_TO_BUILD=iphoneos${SDK_VERSION}
else
OTHER_SDK_TO_BUILD=iphonesimulator${SDK_VERSION}
fi
echo "XCode has selected SDK: ${PLATFORM_NAME} with version: ${SDK_VERSION} (although back-targetting: ${IPHONEOS_DEPLOYMENT_TARGET})"
echo "...therefore, OTHER_SDK_TO_BUILD = ${OTHER_SDK_TO_BUILD}"
#
#####################[ end of part 1 ]##################
#####################[ part 2 ]##################
#
# IF this is the original invocation, invoke WHATEVER other builds are required
#
# Xcode is already building ONE target...
#
# ...but this is a LIBRARY, so Apple is wrong to set it to build just one.
# ...we need to build ALL targets
# ...we MUST NOT re-build the target that is ALREADY being built: Xcode WILL CRASH YOUR COMPUTER if you try this (infinite recursion!)
#
#
# So: build ONLY the missing platforms/configurations.
if [ "true" == ${ALREADYINVOKED:-false} ]
then
echo "RECURSION: I am NOT the root invocation, so I'm NOT going to recurse"
else
# CRITICAL:
# Prevent infinite recursion (Xcode sucks)
export ALREADYINVOKED="true"
echo "RECURSION: I am the root ... recursing all missing build targets NOW..."
echo "RECURSION: ...about to invoke: xcodebuild -configuration \"${CONFIGURATION}\" -project \"${PROJECT_NAME}.xcodeproj\" -target \"${TARGET_NAME}\" -sdk \"${OTHER_SDK_TO_BUILD}\" ${ACTION} RUN_CLANG_STATIC_ANALYZER=NO" BUILD_DIR=\"${BUILD_DIR}\" BUILD_ROOT=\"${BUILD_ROOT}\" SYMROOT=\"${SYMROOT}\"
xcodebuild -configuration "${CONFIGURATION}" -project "${PROJECT_NAME}.xcodeproj" -target "${TARGET_NAME}" -sdk "${OTHER_SDK_TO_BUILD}" ${ACTION} RUN_CLANG_STATIC_ANALYZER=NO BUILD_DIR="${BUILD_DIR}" BUILD_ROOT="${BUILD_ROOT}" SYMROOT="${SYMROOT}"
ACTION="build"
#Merge all platform binaries as a fat binary for each configurations.
# Calculate where the (multiple) built files are coming from:
CURRENTCONFIG_DEVICE_DIR=${SYMROOT}/${CONFIGURATION}-iphoneos
CURRENTCONFIG_SIMULATOR_DIR=${SYMROOT}/${CONFIGURATION}-iphonesimulator
echo "Taking device build from: ${CURRENTCONFIG_DEVICE_DIR}"
echo "Taking simulator build from: ${CURRENTCONFIG_SIMULATOR_DIR}"
CREATING_UNIVERSAL_DIR=${SYMROOT}/${CONFIGURATION}-universal
echo "...I will output a universal build to: ${CREATING_UNIVERSAL_DIR}"
# ... remove the products of previous runs of this script
# NB: this directory is ONLY created by this script - it should be safe to delete!
rm -rf "${CREATING_UNIVERSAL_DIR}"
mkdir "${CREATING_UNIVERSAL_DIR}"
#
echo "lipo: for current configuration (${CONFIGURATION}) creating output file: ${CREATING_UNIVERSAL_DIR}/${EXECUTABLE_NAME}"
xcrun -sdk iphoneos lipo -create -output "${CREATING_UNIVERSAL_DIR}/${EXECUTABLE_NAME}" "${CURRENTCONFIG_DEVICE_DIR}/${EXECUTABLE_NAME}" "${CURRENTCONFIG_SIMULATOR_DIR}/${EXECUTABLE_NAME}"
#########
#
# Added: StackOverflow suggestion to also copy "include" files
# (untested, but should work OK)
#
echo "Fetching headers from ${PUBLIC_HEADERS_FOLDER_PATH}"
echo " (if you embed your library project in another project, you will need to add"
echo " a "User Search Headers" build setting of: (NB INCLUDE THE DOUBLE QUOTES BELOW!)"
echo ' "$(TARGET_BUILD_DIR)/usr/local/include/"'
if [ -d "${CURRENTCONFIG_DEVICE_DIR}${PUBLIC_HEADERS_FOLDER_PATH}" ]
then
mkdir -p "${CREATING_UNIVERSAL_DIR}${PUBLIC_HEADERS_FOLDER_PATH}"
# * needs to be outside the double quotes?
cp -r "${CURRENTCONFIG_DEVICE_DIR}${PUBLIC_HEADERS_FOLDER_PATH}"* "${CREATING_UNIVERSAL_DIR}${PUBLIC_HEADERS_FOLDER_PATH}"
fi
fi
6) Hit "cmd + B" (Build Project)
7) Open Product in Finder
8) Navigate 1 directory up ("cmd + ↑"), and you will see "Release-universal" directory.
There will be your "fat/universal" library, You are ready to go!
I encountered this with a framework lib I'm using in one of my apps, when I tried to test it in iPhone Retina 64bit simulator.
I simply added x86_64 as an architecture to build for and set it to always build for all architectures. Worked a charm.
The lipo tool can not only create fat mach-o binaries, but it can inspect them:
xcrun lipo -info /path/to/libThing.a
This will output what architectures are in the file. Before you join binaries using lipo, run this to make sure the architectures you expect are present. It's also a good idea to run this on the product of a fat binary join.
In your case you need:
iPhoneSDK Configuration:
armv7, armv7s, arm64
iPhoneSimulator Configuration:
i386, x86_64
It seems that the iPhoneSimulator build product is not producing an x86_64 binary based on your question. Check your build configuration - in particular, "Build Active Architectures Only" (ONLY_ACTIVE_ARCH ) should be set to NO. The default is for this to be NO for Release, but YES for debug. If it is YES, only one architecture will be in the build product.
Go to your application project Target and look in Library Search Path.
Now check that your library file path should be written in double quotes:
"$(SRCROOT)/MyAppTest/TestFlight"
If there is no double quote then just add them and compile the project.
Hope it will work for you.