protobag VS bloaty

In many cases you do indeed want or need to use Protobuf for telemetry or other time-series data. The Protobuf RecordIO / TFRecord formats are very unhelpful for this use case.

Enter Protobag: https://github.com/StandardCyborg/protobag

Protobag is a library that helps you write and read time-series and other collections of Protobuf data using the standard Tar and Zip archives as containers. Protobag also includes code for leveraging self-describing messages to embed schemas in archives so that you never have data on disk that you can't read. Hope somebody finds Protobag helpful.

ESP32s aren't really ‘lower level’ in the sense that anyone is likely to write assembly code for them (compared to, say, 8051 or PIC), other than maybe some driver author at Espressif. The big win from using RISC-V, other than name recognition, is mainstream compiler support (which is nothing to sneeze at, especially when it's largely funded by someone else).

When I worked on Matter¹, the Xtensa and RISC-V versions were basically fungible from the software point of view. (And really, so were other vendors' various ARMs.) We did find that Bloaty McBloatface² didn't support Xtensa, so I had to write an alternative.

¹ https://github.com/project-chip/connectedhomeip/

² https://github.com/google/bloaty

I’ve gotten good insight into what takes up space in binaries by profiling with Bloaty (https://github.com/google/bloaty). My last profiling session showed that clang’s ThinLTO was inlining too aggressively in some cases, causing functions that should be tiny to be 75 kB+.

I'm surprised they didn't go for the binary size analysis tools like

https://github.com/google/bloaty

Or goweight.

I was able to shave a few % off our app binary with Bloaty. It’s pretty hard to use but once you figure out how to make regular expressions to properly classify things from your codebase, you can really visually analyze what your binary is composed of.

Bloaty