feat: use sync interfaces for controller to avoid UniFFI Android bug#264
feat: use sync interfaces for controller to avoid UniFFI Android bug#264thomas-waite wants to merge 11 commits intomainfrom
Conversation
thomas-waite
requested review from
andy-t-wang,
aurel-fr,
lukejmann,
murph,
paolodamico,
sideround and
tomislavhoman
as code owners
|
@codex review |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 703e819c5e
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
|
@codex review |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 9606844d81
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| use once_cell::sync::Lazy; | ||
| use std::sync::Arc; | ||
| use tokio::sync::Mutex; | ||
| use std::sync::Mutex; |
There was a problem hiding this comment.
Switching MIGRATION_LOCK to std::sync::Mutex introduces lock poisoning semantics: if a migration processor panics while run_migrations() holds the guard, later try_lock() calls return Poisoned and are currently mapped to MigrationError::InvalidOperation("already in progress"). In that scenario, all subsequent migration attempts in the process will be rejected even though no run is active, which is a regression from the previous tokio::sync::Mutex behavior. Please distinguish WouldBlock from Poisoned (or recover the poisoned guard) so panics don’t permanently brick migrations.
Useful? React with 👍 / 👎.
1 participant
Use sync interfaces for controller to avoid UniFFI Android bug
There is a bug in UniFFI where Kotlin bindings will sometimes throw an error when calling an async Rust function: mozilla/uniffi-rs#2624
This PR updates the migration processor interfaces to use
syncratherasyncas a work around. It has been documented that the calling mobile clients will need to take care of the call being blocking.