When I start build using eas android app-bundle I'm receiving this message:
Compressing project files and uploading to EAS Build. Learn more
ValidationError: "gitCommitMessage" is not allowed to be empty
eas.json
{
"build": {
"preview": {
"android": {
"buildType": "app-bundle",
"gradleCommand": ":app:assembleRelease",
"developmentClient": false,
"withoutCredentials": true
}
},
"production": {}
}
}
I don't know what happens
Such error message could be caused by a few things that I can think of.
The last commit on one or more of the files you made to the branch you are checked out to when running eas build has an empty message
If this is the case the solution would be to add a new commit (it could be an empty one, but with a message). Here is an example of how to do that:
git commit --allow-empty -m "Place commit message here"
Alternatively, but less likely, you may be running the eas build command with an empty string after a message flag "-m". Example:
eas build -p all -m
If this is the case, either remove the "-m" flag, or add a message to your build similarly to the above eas build -p all -m YOUR_COMMENT
Related
In my project I am using expo-sdk:43 along with expo-updates.
I have setup the channel property in eas.json file.
{
"cli": {
"version": ">= 0.52.0"
},
"build": {
"development": {
"developmentClient": true,
"distribution": "internal",
"channel":"dev"
},
"preview": {
"android": {
"buildType": "apk"
},
"channel":"preview"
},
"production": {}
},
"submit": {
"production": {}
}
}
I have tried to publish my updates through EAS update, but it is not working and the app is not receiving any update. I have tried to do an apk build for the application, and tried to get the output of Updates.Channel and it is outputting as undefined.
eas build -p android --profile --preview
eas update --branch preview --message "update"
But, when I build the app, the build is successful, in the eas logs I can see that it shows that channel property has also been configured.
Can someone please help what might be the possible issue in this case?
I have also tried to do the same steps with an expo-init command (But with expo sdk:47) and minimal app setup and it is working as expected.
Description:
During the deploy of firebase hosting, I received an error stating that 0 files were found. I have included my firebase.json file for reference.
Steps to reproduce:
Run the command firebase deploy --only hosting
Observe the error message stating that 0 files were found
Expected result:
The firebase hosting should be successfully deployed with the specified files.
Actual result:
An error is thrown stating that 0 files were found.
+ hosting: Finished running predeploy script.
i hosting[hosting-project]: beginning deploy...
i hosting[hosting-project]: found 0 files in hosting
+ hosting[hosting-project]: file upload complete
i hosting[hosting-project]: finalizing version...
+ hosting[hosting-project]: version finalized
i hosting[hosting-project]: releasing new version...
+ hosting[hosting-project]: release complete
+ Deploy complete!
Notes:
I have double-checked the file path in the firebase.json file and it
appears to be correct.
I have tried rerunning the deploy command multiple times with the same result
I have also tried deploying a
different project with the same firebase.json file, but the issue
persists.
In my firebase.json file, I am not targeting the dist folder directly because I am using the predeploy script to run npm run lint and npm run build before the deployment. However, I am not ignoring the dist folder with !dist% and !dist/*, which means I am not excluding it from the deployment.
Attached files:
firebase.json
{
"firestore": {
"port": "8080"
},
"functions": [
{
"source": "functions",
"codebase": "default",
"ignore": [
"node_modules",
".git",
"firebase-debug.log",
"firebase-debug.*.log"
],
"predeploy": [
"npm --prefix \"$RESOURCE_DIR\" run lint"
]
}
],
"hosting":{
"public":"hosting",
"ignore": [
"*",
"!dist/",
"!dist/*",
],
"rewrites":[
{
"source":"**",
"destination":"dist/index.html"
}
],
"predeploy": [
"npm --prefix \"$RESOURCE_DIR\" run lint",
"npm --prefix \"$RESOURCE_DIR\" run build"
]
}
}
I found a solution to resolve the issue of "0 files found" during the deploy of firebase hosting. It may not be the most elegant solution, but it does work. In the "predeploy" section of the hosting configuration, I added ".." at the end of the "npm run lint" and "npm run build" commands to go back to the root folder:
"npm --prefix \"$RESOURCE_DIR\\..\" run lint",
"npm --prefix \"$RESOURCE_DIR\\..\" run build"
This allowed me to target the dist folder directly in the hosting configuration, as specified in the official documentation. Here is the modified hosting configuration:
"hosting":{
"public":"hosting/dist",
"ignore": [
"**/.*",
"**/node_modules/**"
],
"rewrites":[
{
"source":"**",
"destination":"/index.html"
}
],
"predeploy": [
"npm --prefix \"$RESOURCE_DIR\\..\" run lint",
"npm --prefix \"$RESOURCE_DIR\\..\" run build"
]
}
I hope this helps anyone else who may be experiencing the same issue.
I'm using the new expo-dev-client and EAS workflow to attempt development with expo, while also including native dependencies (react-native-firebase in my case).
I've managed to build and install a preview of my app with EAS:
eas build --platform ios --profile staging
Where my "staging" profile is setup in eas.json like so:
{
"cli": {
"version": ">= 0.35.0"
},
"build": {
"production": {
"releaseChannel": "prod-v0",
"node": "14.17.4",
"ios": {
"cocoapods": "1.11.2",
"autoIncrement": "buildNumber"
}
},
"staging": {
"extends": "production",
"releaseChannel": "staging",
"distribution": "internal"
},
...
},
...
}
Note the releaseChannel is set to "staging" too.
So far so good. The app is building and I can install it on my device and confirm the releaseChannel using the expo-updates module with Updates.releaseChannel
Now I'd like to publish a pure javascript change, so I run...
expo publish --release-channel staging
But this gives the following warning (which I assume is why I don't see the update on my device):
ios: Expo.plist key: "EXUpdatesReleaseChannel": The value passed to the
--release-channel flag is to "staging", but it is set to "default".
So what/where is "Expo.plist", why isn't it in-sync, and how can I update it?
Update:
So I deleted the iOS folder from a previous expo run:ios build. The warning is now gone when running expo publish --release-channel staging. But still nothing seems to happen on my app - it still doesn't update.
I'm hoping to confirm my thought process when trying to test an Android application that's built with Expo's Turtle-CLI tool. I've got iOS running well, just need help getting Android working.
I have a build process that will spit out an .apk to build/android.apk. Here's the command that does that
turtle build:android \
--output $BUILD_DIR/android.apk \
--username $EXPO_USERNAME \
--password $EXPO_PASSWORD \
--config app.config.ts \
--release-channel $RELEASE_CHANNEL \
--type apk \
--mode debug
This successfully outputs the .apk in build.
Here is my detoxrc.json configuration
{
"testRunner": "jest",
"runnerConfig": "e2e/config.json",
"configurations": {
"ios": {
"type": "ios.simulator",
"binaryPath": "build/app-native.app",
"build": "./scripts/build_test_app ios",
"device": {
"type": "iPhone 11"
}
},
"android": {
"type": "android.attached",
"binaryPath": "build/android.apk",
"testBinaryPath": "build/android.apk",
"build": "./scripts/build_test_app android",
"device": {
"adbName": "059aaa47"
}
}
}
}
I then attempt to run my tests with
detox test --configuration android --loglevel trace
Which then yields (more logs at the bottom of this post)
No instrumentation runner found on device 059aaa47 for package com.foobar.mobilern
at ADB.getInstrumentationRunner (../node_modules/detox/src/devices/drivers/android/exec/ADB.js:250:13)
at Instrumentation.launch (../node_modules/detox/src/devices/drivers/android/tools/Instrumentation.js:19:24)
at MonitoredInstrumentation.launch (../node_modules/detox/src/devices/drivers/android/tools/MonitoredInstrumentation.js:18:5)
at AttachedAndroidDriver._launchInstrumentationProcess (../node_modules/detox/src/devices/drivers/android/AndroidDriver.js:284:5)
at AttachedAndroidDriver._launchApp (../node_modules/detox/src/devices/drivers/android/AndroidDriver.js:267:7)
at AttachedAndroidDriver._handleLaunchApp (../node_modules/detox/src/devices/drivers/android/AndroidDriver.js:120:7)
at AttachedAndroidDriver.launchApp (../node_modules/detox/src/devices/drivers/android/AndroidDriver.js:91:12)
at Device._doLaunchApp (../node_modules/detox/src/devices/Device.js:85:19)
at traceCall (../node_modules/detox/src/utils/trace.js:41:20)
If I omit the testBinaryPath key from .detoxrc.json, then I get this result.
Running adb -s 059aaa47 shell pm list instrumentation on the test device yields an empty response, so I'm inclined to believe that some testing setup isn't being completed thoroughly and/or perhaps I'm thinking about this wrong.
I guess the gist of my question is, is it possible to test one .apk without having to make use of some sort of testing harness / binary?
Full Log Output
I've created a repo I want to be cloned
after cloning, you run npm install
How do I echo a log message to the terminal that will show up at the end of the installation?
In your case, you can add a post-install script, that outputs, for example, a string to the console if you mark a version as alpha.
package.json
{
"version": "1.2.3-alpha.2",
"scripts": {
"postinstall": "node postinstall.js"
}
}
postinstall.js
const package = require('./package.json')
console.log('End of installation');
//Example using properties from package.json
if (package.version.includes('alpha')) {
console.log('Warning: Alpha version!')
}
BONUS
Append this to the command --loglevel verbose and all logs will be saved to npm-debug.log in current working directory.
It will show the logs in realtime + saves the log to directory its running.
You can just edit npm config npm config edit and add loglevel=verbose