Menu

Automating Android Play Store Releases, Part 5: The Bug That Shipped Live
πŸ“°

Automating Android Play Store Releases, Part 5: The Bug That Shipped Live

DEV Community: cicdΒ·cynthia wahomeΒ·4 days ago
#KJ3A1juu
#devtocicd#devto#dev#comment#staff
Reading 0:00
15s threshold

Part 5 of a 5-part series on automating a multi-app Android release pipeline. Part 4 β†’ ended on the first real production release going out cleanly. This part covers what showed up right after β€” a bug that had nothing to do with signing, versioning, tracks, or storage, and everything to do with a comment that stopped being true nine minutes after it was written.

TL;DR

Right after the first real production release, both apps started showing the same login screen β€” staff's.

Root cause: this org has exactly one self-hosted runner. Both apps build sequentially, on the same disk, sharing the same Metro bundler cache.

A comment already in the codebase claimed this couldn't happen , because staff and student "each get their own runner." That was true β€” for eight days, on a different runner setup, before a later, unrelated commit quietly removed the isolation the comment depended on.…

Continue reading β€” create a free account

Join HashtagPLUS to read full articles, follow hashtags, vote, and join the conversation.