дзеркало
https://github.com/l0ng-ai/openidea.git
синхронізовано 2026-11-09 11:15:40 +02:00
Jaw
daac8e80b0
Usage counts asked while a language server loads the project come back incomplete and were re-asked only after a busy spell of 1.5 s or more, seen from the editor's own mount. An editor restored at launch mounts while its server starts, so it missed the spell's start, and a short load (vtsls loads a small project in about a second) never counted. Server spells are now followed once from app start, and the spell right after a start counts as loading whatever its length; later short spells are still ignored so request progress (jdtls) doesn't loop. The implementations gutter shares the same settle signal. Co-authored-by: jaw52 <[email protected]>