The Gram Code Editor - 2026-10-03


  1. mort

    https://upload.yax.im/upload/SZwfxV0iMgQ_QkK-5Zb4py9E/Screenshot%202026-10-03%20at%2012.35.13.jpg

  2. mort

    I can see why you wanna get rid of protobuf hehe

  3. krig

    yeah protobuf and aws-lc-sys.. if I can get rid of those two that would make a massive difference

  4. mort

    I'm taking a stab at replacing protobuf with postcard by the way. No idea if it'll work but I have all the message types defined in terms of rust now at least...

  5. mort

    It'll probably be a massive PR which touches a lot because my structs will necessarily be different from the ones protobuf generates. Might end up being unmergable due to the amount of churn, or I might give up. But it might be valuable as an experiment still

  6. krig

    woah, nice! I started looking at it too but got distracted pretty quick. But then I’ll let you have a go at it 👍

  7. mort

    https://codeberg.org/GramEditor/gram/pulls/608 I created a draft MR so that the progress is visible

  8. ashleytg

    > https://upload.yax.im/upload/SZwfxV0iMgQ_QkK-5Zb4py9E/Screenshot%202026-10-03%20at%2012.35.13.jpg WOAH

  9. ashleytg

    WHAT IS THAT :0

  10. mort

    Protobuf 😛

  11. krig

    Google tech.. XD

  12. mort

    Yeah seriously

  13. mort

    I work on a C++ project which uses GRPC and it's such a pain, it's the only library in the world which changes its ABI based on whether you compile with asan enabled or not

  14. mort

    So when you get GRPC from your package manager (i.e compiled with address sanitizer disabled), you can't then compile your project with ASan enabled

  15. mort

    To work around this I once tried building my own GRPC but it's *massive* and depends on relatively specific versions of roughly the entire world so I gave up lol

  16. krig

    pre-AI slop…

  17. mort

    Huh both the remote crate and the rpc crate parses byte buffers into messages, I expected this to happen in just one place

  18. krig

    @mort ooh I see that your PR doesn't have the latest changes.. yesterday I cleaned up a bunch of rpc code that was left over from the old proprietary remote collaboration, and I think at least some of that duplication is gone now.. there was a whole unused websocket-based collaboration thing there

  19. krig

    I'm sure there is a lot left to remove still

  20. ashleytg

    > Protobuf 😛 i meant what are you using to see compile time for each thing-

  21. mort

    @krig Ah I didn't realise all that RPC websocket stuff was unused! Rebased now

  22. ashleytg

    has anyone run a linting tool on the source code, or check for _name variables?

  23. mort

    What about _name variables? As in unused variables?

  24. ashleytg

    yea :3 i've never compiled gram from source, so i'm not sure if there are any!