GitHub issue #15360 in cline/cline, opened 2026-10-26 by gregtakacs, records a report that in long tasks the VS Code (Remote-WSL) extension host dies from V8 heap exhaustion, that Cline's panel resets and reopening the same task brings the crash back within 1-2 minutes so the task becomes a crash loop, that there were 33 of these exits in 6 days (18 on 10-21, 6 on 10-23, 9 on 10-26), and that while a turn streams the entire state is re-serialized and pushed to the webview about every 0.4-0.6 s, with one run on 10-26 issuing 1,194 of these pushes in 41 minutes, about 20 GB of JSON serialized, before the host died; the task involved had 5,824 messages and its messages file was 17.3 MB, the same size as each push. The author notes he has not proven whether the serialized strings are retained or are churn the garbage collector cannot keep up with. System cited: Cline 4.1.25, VS Code 1.142.0, Remote-WSL, Windows + WSL2.
Verbatim Excerpts from Cited Records
“In long tasks the VS Code (Remote-WSL) extension host dies from V8 heap exhaustion.”
·
cline/cline (issue 15360 opening body, What happened section)
·
2026-10-26T16:50:16Z
·GitHub Issue ↗
“There were 33 of these exits in 6 days (18 on 10-21, 6 on 10-23, 9 on 10-26).”
·
cline/cline (issue 15360 opening body, What happened section)
·
2026-10-26T16:50:16Z
·GitHub Issue ↗
“While a turn streams, the **entire state** is re-serialized and pushed to the webview about every 0.4–0.6 s:”
·
cline/cline (issue 15360 opening body, What happened section)
·
2026-10-26T16:50:16Z
·GitHub Issue ↗
“One run on 10-26 issued 1,194 of these pushes in 41 minutes, **~20 GB of JSON serialized**, and then the host died.”
·
cline/cline (issue 15360 opening body, What happened section)
·
2026-10-26T16:50:16Z
·GitHub Issue ↗
Counter-record: Issue open as of the fetch; both comments read. linear-code[bot] (2026-10-26) linked the report to CLINE-3692. MindShow (2026-11-14) added an independent report on Cursor / VS Code-compatible extension over Remote SSH to Linux: three remote extension host runs ended with the same V8 heap out of memory fatal error and SIGABRT on 2026-11-14, near a roughly 2 GiB heap ceiling, with time-to-crash about 144-219 seconds after host startup, providers.read bursts peaking around 242-256 log entries per second, and the largest persisted session messages file about 3.8 MB. MindShow states the OOM itself is confirmed but providers.read frequency is NOT proven to be the allocation-retention root cause, that code inspection found sendStateUpdate serializing full state and firing subscriber callbacks without awaiting backpressure, and that the local mitigation is a reversible local patch, not an upstream fix, with reload and representative long-task OOM validation still pending; the report does not claim the production crash is resolved. No maintainer fix and no retraction of the report appear in the thread. Related published claims C086 (#15427) and C089 (#15031) cover the same state-serialization payload on the JetBrains and cline-core surface, distinct reports; this claim documents only #15360 on the VS Code extension-host surface.
Falsification probe: If GitHub issue #15360 is deleted without archive, if the author edits the opening body so the quoted sentences no longer match verbatim, or if the author posts a retraction of the report, this claim drops to Withdrawn.
Audit verification: Verified 100% continuous verbatim substring against official GitHub REST API by
scripts/verify_claim_verbatim.py.