fix(api): the same game added to a run twice is one launch (#426)

Follow-up to #422, from the nesktop #3 review. POST /session/{id}/launch
for a game that already has an open launch (requested or running, no
stop requested) in the run now answers that launch with 200 instead of
inserting a second one. The check runs under the run's row lock that
addLaunch already takes, so two concurrent adds resolve to one launch.
OpenAPI documents the 200.

Tested: apps/api 208 pass against the local test Postgres, including a
new test adding the same game twice concurrently.
This commit is contained in:
Wanjohi
2026-10-17 16:51:25 +03:00
committed by GitHub
parent 0c348d3973
commit 3ce8aee427
3 changed files with 29 additions and 1 deletions
+7
View File
@@ -476,6 +476,11 @@ export namespace SessionApi {
description:
'Adds a launch to a run that its host has taken on and that is not being stopped; the host starts it in the box without stopping what already runs there. Answers with the launch: its `id` is what the launched game’s window carries, and what to stop when that window closes. A run that is waiting for a host, stopping or over is 409: ask for a new run instead. So is a run on a host that does not handle launches, with code `host_too_old`: stop the run and ask for a new one, as before launches. The run is still what is billed and what the allowance counts, so adding a launch asks nothing of either.',
responses: {
200: {
content: { 'application/json': { schema: Result(Launch.Info) } },
description:
'That game already has a launch in this run that is not over: it is answered instead of a second one'
},
201: {
content: { 'application/json': { schema: Result(Launch.Info) } },
description: 'The launch has been requested'
@@ -537,6 +542,8 @@ export namespace SessionApi {
conflict('That run is not running, so it cannot take another launch');
case 'hostTooOld':
hostTooOld();
case 'existing':
return c.json({ data: result.launch }, 200);
default:
return c.json({ data: result.launch }, 201);
}