Bitrise 16 Jan. 2018
- (iOS) Infer integration takes too much time.
- There's a hard limit of 120 minutes for maximum build time, but based on last email with Bitrise team yesterday (30 January), they are working to enable 2+ hours max build time for us.
- I've sent raw build logs for iOS builds that timed out, and they said the slowness comes from Infer process itself, so they actually have no idea how to optimize this from Bitrise side.
- Bitrise team requested me to allow access for their support team to read, run, and configure our pipeline (via an option in the settings page). However, I've put their request on hold as I'm still considering security concerns if they have access to sensitive information (e.g passwords, API keys, tokens).
- (Android) Possibility to increase RAM?
- Currently it's not possible from their side to increase more than 16GB RAM, because for Android, Bitrise is using GCE and they claim it's already the highest tier.
- They have a brand new, unreleased Enterprise plan. Basically, this is a self-hosted Bitrise, so we have to manage the servers ourselves (similar to Phabricator).
- By default, unlike regular PCs, virtual machines do not have swap memory configured. That's why the build gets very slow when it approaches the memory limit. Bitrise team suggested to try setting the swap memory by utilizing the script here: https://gist.github.com/viktorbenei/1037fb31b403eecc3dab77000589abc6
- (Testing) Possibility to run UI test in Device Farm?
- For an integrated, ready-to-use solution for Android only using simulator, they provide this article: https://blog.bitrise.io/introducing-solid-and-snappy-virtual-device-testing-for-android
- To run UI test on proper device, they suggest integration with Amazon Device Farm.
- They only provide the pipeline to correctly upload the test build to Amazon, but we have to manage the account ourselves.
- Amazon Device Farm account is not free.
- They also suggest to make the workflow asynchronous. The Bitrise workflow will only consist of building the test build and upload it to Amazon Device Farm. Once the test finishes, we can set the configuration in Amazon to report the result (e.g. send report via email, send message to Slack, or trigger another pipeline in Bitrise).