The problem is one of time and skill. Who can / will modify the Go code to be compatible with static compilation? How long will that take? Would we rather delay the release of PeachPub (might be…
But it's not just returning general data structures, right? It seemed like your intention was to make it super convenient by grouping the results of various sbot calls into more complex structures.…
It's the exact same work being done, just split into two.
I think there may be cases where the low-level calls already return the strings we want, or we don't need to parse the result, in which…
posted on ssb too just to see if anyone has an opinion
another advantage of pushing forward on this approach is we would make it easier for people to also make these FFI bindings in other languages besides rust
another blog post in similar vein: https://radu-matei.com/blog/from-go-to-rust-static-linking-ffi/
ok here is a slight variant approach to consider:
https://legionlabs.com/blog/rust_ffi_finding/
found based on @soapdog…
one more potential downside of two libraries:
potentially will have to write some functions twice.
If peach-web instantiates an instance of Sbot from the high-level library, then maybe it…
I think mostly on the same page, but I still dont know that most libary users would be more likely to use go-sbotcli-rs than peach-sbot.
Which also makes me wonder if it should be called…
what would you call the two libraries?
Could also be one library, which includes both the low-level calls (one-to-one correspondence with sbotcli calls), as well as additional higher level calls, and the library user (such as us, in…
This feels to me like deciding whether we want one or two libraries. And would mostly be a small distinction, as the parsing still needs to live somewhere.
Many API libraries will do parsing and…
Questions about what sbotcli calls to use to get the right filters will effect this code, but this is a basic PR going in the direction of parsing json, if we decide thats what we want to do @glyph