Label forkserver child as `subint_forkserver`

Follow-up to 72d1b901 (was prev commit adding `debug_mode` for
`subint_forkserver`): that commit wired the runtime-side
`subint_forkserver` SpawnSpec-recv gate in `Actor._from_parent`, but the
`subint_forkserver_proc` child-target was still passing
`spawn_method='trio'` to `_trio_main` — so `Actor.pformat()` / log lines
would report the subactor as plain `'trio'` instead of the actual
parent-side spawn mechanism. Flip the label to `'subint_forkserver'`.

(this patch was generated in some part by [`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
Gud Boi 2026-04-23 11:43:06 -04:00
parent 72d1b9016a
commit 4425023500
1 changed files with 10 additions and 6 deletions

View File

@ -655,12 +655,16 @@ async def subint_forkserver_proc(
loglevel=loglevel, loglevel=loglevel,
parent_addr=parent_addr, parent_addr=parent_addr,
infect_asyncio=infect_asyncio, infect_asyncio=infect_asyncio,
# NOTE, from the child-side runtime's POV it's # The child's runtime is trio-native (uses
# a regular trio actor — it uses `_trio_main`, # `_trio_main` + receives `SpawnSpec` over IPC),
# receives `SpawnSpec` over IPC, etc. The # but label it with the actual parent-side spawn
# `subint_forkserver` name is a property of HOW # mechanism so `Actor.pformat()` / log lines
# the parent spawned, not of what the child is. # reflect reality. Downstream runtime gates that
spawn_method='trio', # key on `_spawn_method` group `subint_forkserver`
# alongside `trio`/`subint` where the SpawnSpec
# IPC handshake is concerned — see
# `runtime._runtime.Actor._from_parent()`.
spawn_method='subint_forkserver',
) )
return 0 return 0