[dragotin]

No description provided.

[Copilot]

🟡 Changes recommended

Socket disconnects can still terminate or abort the process, and shutdown can silently drop accepted hydration requests.

4 open findings

• Prevent SIGPIPE from terminating the FUSE process · New
• Guard socket input after descriptor invalidation · New
• Synchronize shutdown checks with message enqueueing · New
• Correct the misspelled user-visible log message · New

What changed in this PR

Bounds hydration waits and improves socket/job coordination for FUSE file hydration.

Changes:

• Adds configurable hydration timeouts and result states.
• Reports queue/send failures promptly and makes transfer IDs atomic.
• Adds socket framing and timeout integration tests.

File	Description
INTEGRATION.md	Documents configuration and socket behavior.
src/​openvfsfuse/​config.json	Adds the default timeout.
src/​openvfsfuse/​main.cpp	Parses timeout configuration.
src/​openvfsfuse/​openvfsfuse.cpp	Implements bounded hydration waiting.
src/​openvfsfuse/​openvfsfuse.h	Stores the timeout argument.
src/​openvfsfuse/​sharedmap.cpp	Adds condition-variable waiting.
src/​openvfsfuse/​sharedmap.h	Defines hydration states and results.
src/​openvfsfuse/​socketthread.cpp	Propagates send and shutdown failures.
src/​openvfsfuse/​socketthread.h	Makes PostMsg report failure.
src/​openvfsfuse/​tests/​CMakeLists.txt	Registers the socket integration test.
src/​openvfsfuse/​tests/​socketthreadtest.cpp	Tests framing, timeout, and shutdown behavior.

🧠 Review effort: Balanced

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[TheOneRing]

Is a timeout here ideal at all?
I mean we should be able to see when the last update was to the file was performed, to see whether the file is still downloading.
The download of large file can just take time...

[Copilot]

🔵 Needs a closer look

Configuration conversion errors can terminate startup, socket initialization failures can hang the test, and reply-size documentation contradicts the implementation.

1 open issue

• Synchronize shutdown checks with message enqueueing

3 previously missed issues

In code that hasn't changed since last review

Document the 100 KiB socket input buffer limit

INTEGRATION.md:62

The unlimited-reply claim contradicts processSocketInput(): it discards buffered, unterminated data exceeding 100 KiB. A larger reply split across polling cycles can therefore be lost and cause a hydration timeout. Document this limit rather than promising arbitrary reply sizes.

Catch configuration type errors during startup parsing

src/​openvfsfuse/​main.cpp:100

A valid JSON config containing "hydrationTimeoutSeconds": "300" or null throws json::type_error here, outside the parsing catch. main() does not catch it, so startup terminates instead of reporting a controlled configuration error. Move setting extraction, including the ignoreApps conversions, inside the try block.

Check socket initialization before blocking on accept

src/​openvfsfuse/​tests/​socketthreadtest.cpp:90

If socket initialization fails, CreateThread() returns false without connecting, but the test proceeds to a blocking accept() with no client. The test then hangs instead of reporting the startup failure. Check the return value before accepting.

3 resolved since last review

• Prevent SIGPIPE from terminating the FUSE process
• Guard socket input after descriptor invalidation
• Correct the misspelled user-visible log message

_{Review effort: Balanced}

[dragotin]

About the timeout: The current built in timeout is the overall timeout after that a running request is stopped. It is true that it is independent of any progress on file download, but that is not in scope of this PR.