Vitali Lovich

Ah. Ok. I gave it a few minutes to run now I see that we have `CAPNP_SKIP_FUZZ_TEST` set in Ekam but not under CMake by default.

Hey @kentonv would you be able to sketch out a rough guideline of how you think the API enhancement would look like? Might have cycles from new hires who could...

cc @harrishancock for review.

For posterity, I cross-compiled with this CMake toolchain file in `~/MingwToolchain.cmake`: ``` set(CMAKE_SYSTEM_NAME Windows) set(CMAKE_C_COMPILER i686-w64-mingw32-cc) set(CMAKE_CXX_COMPILER i686-w64-mingw32-c++) set(CMAKE_RC_COMPILER i686-w64-mingw32-windres) set(CMAKE_FIND_ROOT_PATH /usr/i686-w64-mingw32) set(CMAKE_FIND_ROOT_PATH_MODE_PROGRAM NEVER) set(CMAKE_FIND_ROOT_PATH_MODE_LIBRARY ONLY) set(CMAKE_FIND_ROOT_PATH_MODE_INCLUDE ONLY) ``` ```...

You're obviously right about this not being an issue for 32-bit operating systems. I assumed this was an issue for Windows because I saw problems removing the writev on MacOS...

Hmm.... the fix for write seems straightforward to me but the writev fix seems less so, but it's likely I'm missing something obvious. My reading is that `FdOutputStream` calls ::writev...

> I've had a lot of trouble with GCC's "maybe uninitialized" warning producing completely nonsensical complains, which is why I have disabled it in CI. Really? I've generally not seen...

@kentonv do you have any thoughts on how to fix the compiler error in https://github.com/capnproto/capnproto/pull/1130/checks?check_run_id=1658987563 ? All that's being done in that spot is a mild refactor but it's complaining...

I think I ran across this today trying to split up a 3-way handshake that exchanges the interfaces peers should use (or maybe I'm just structuring the capnproto incorrectly)? client.capnp:...

Oh ok. @harrishancock raised it as a more serious compat concern.