Hi everyone,
This is the Zcash Foundation’s Engineering update for Sprint 7 (March 31st - April 11th) of 2025.
During this sprint the team continued to focus on improving Zebra’s RPC methods to match current zcashd usage in order to enable zcashd deprecation. To that effect we added support for negative heights as a parameter for a number of RPC methods, fixed the way we parse ranges for the getaddresstxids RPC method and we finally merged a number of fixes, one to ensure that the getpeerinfo RPC reports inbound connections as well as outbound ones and another to add additional fields to the getrawtransaction RPC.
We have been working on a number of fixes and cleanups, including an issue reported by @AloeareV whereby errors were being reported when syncing whenever a secondary rocksdb instance was behind the primary one. We have also fixed a panic when trying to call a number of commands on the zebra database while the zebra process is running and a fix for a warning that can happen during zebra start up.
Finally, we’ve been reviewing and merging a number of PRs by external contributors. Our thanks to @str4d @nuttycom @PabloK @aphelionz and others for helping to improve zebra!
On the FROST project, we have been working on the last few suggestions reported from the security audit of the frost client and server, frostd, including a suggestion to rename functions and variables in participants and coordinator for improved clarity and a suggestion to improve HTTP error handling.
On the devops side we’ve been working on improvements and fixes to the Docker file, including improvements to how we test changes to our Docker images. We worked on a fix to explicitly set the HOME environment variable to ensure cache directories are set in the correct location as well as adding a default command to the test stage in the Docker file to prevent a gosu error. We also cleaned up our release images to remove experimental features and updated our documentation to reflect our new strategy with regards to optional zebra features. Finally, we fixed a flaky lightwalletd acceptance test which was causing a lot of CI failures.
Thank you for reading!