Circle CI recently failing due to pod import

We have recently encountered continuously failing builds with our integration, here is the job that fails defined in our config.yml file.

version: 2
jobs:
    build_and_test:
        macos:
            xcode: "10.1.0"
        environment:
            FASTLANE_LANE: test
            BUNDLER_VERSION: 2.0.1
        steps:
            - checkout
            # Restore Gem cache
            - restore_cache:
                key: gem-cache-{{ arch }}-{{ checksum "Gemfile.lock" }}
            - run:
                name: Setup Script
                command: chmod 755 ./setup.sh & ./setup.sh ci
            # Save Gem cache
            - save_cache:
                key: gem-cache-{{ arch }}-{{ checksum "Gemfile.lock" }}
                paths:
                    - vendor/bundle
            # Restore CocoaPods cache
            - restore_cache:
                key: pod-cache-{{ arch }}-{{ checksum "Podfile.lock" }}
            - run:
                name: Fastlane
                command: bundle exec fastlane $FASTLANE_LANE
            # Save CocoaPods cache
            - save_cache:
                key: pod-cache-{{ arch }}-{{ checksum "Podfile.lock" }}
                paths:
                    - Pods

The output file indicates that this line or pod is what is causing the build failure.

:x:/Users/distiller/project/[Company]/Classes/Application/Configuration/Configuration+Application.swift:67:8: no such module ‘ZendeskSDK’
import ZendeskSDK
^

However, ZendeskSDK (we’re using 2.3.0) was not updated on our Podfile and neither was our config file in recent builds. Odd thing is we import ZendeskSDK in other files too, but they compile fine according to the output. So the failure must be on the CI, I’m not sure what the issue is or where to begin investigating.

Anyone have any idea why this might be happening? Or can lead me to a support person that can help us navigate this issue?

I don’t use this technology, but as a piece of general advice, you should debug this issue exactly the same as if you had it on your development laptop. There is nothing particularly special about CircleCI - it is just a Linux server with a bit of software on it.

Assuming “ZendeskSDK” is a Pod, could you try rebuilding without cache to see if that fixes it?

Failing that, can you create a step to install “ZendeskSDK”?

There are a few reports of the same error on the web - could you find out what the general fix is?