René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"quic: Accumulation buffer is not filled, if small packets arrive","articleBody":"I have the following situation:\nOn http/3 (actually webtransport) the underlying data stream consists of 1024 byte packages (coming from a third party server), I see on the reader that the bytes arrive in packets of 100 bytes on js directly from the nghttp3 parser without being accumulated in the buffer.\nThe reason is that in `ReceiveData` in `streams.cc`:\n```\n  } else if (reader_ \u0026\u0026 was_empty) {\n    printf(\"RD 10\\n\");\n    // Notify the reader once when the accumulator transitions from empty\n    // to non-empty. This wakes the reader exactly once per accumulation\n    // cycle. When the reader wakes and calls pull(), BeforePull() flushes\n    // the accumulated data, and the subsequent EntryRead() extends the\n    // flow control window. We do NOT notify on every callback — that\n    // would defeat coalescing by flushing tiny amounts each time.\n    reader_-\u003eNotifyPull();\n    printf(\"RD 11\\n\");\n  }\n```\ndoes actually not cause a wake, but an immediate execution at least for:\n```\n        console.log('before wakeup')\n        const fin = await wakeup.promise;\n        console.log('after wakeup')\n```\nin `createBlobReaderIterable` .\nHere is a log trace:\n```\nReceive data wt 100\nRD 1\nRD 2\nRD 4\nRD 5\nRD 8\nRD 10\nafter wakeup\nFlush acc not pull \n  webtransport:http3wtstream(3804) commitReadBuffer 100 +1ms\nread mark2 4061 1781992229358 406100 100\nbefore wakeup\nafter wakeup\nbefore wakeup\nRD 11\nReceive data wt 100\n```\nto prove it. And yes, it is woken up twice per cycle via  `Stream::FlushAccumulation()` (that is the \"Flush acc not pull\" printout).\n\nIf I can trust on what is gemini saying, it is that the call from C++ will immediately schedule all microtasks before return, but I do not trust it....\n\nIt is unclear, if `createBlobReaderStream(reader)` will show the same issues.\n\nMy first reflex is to wrap the call to the wakeup call in a `nextTick` or `setImmediate` call , but maybe @jasnell  or @pimterry know a better way?\n\n","author":{"url":"https://github.com/martenrichter","@type":"Person","name":"martenrichter"},"datePublished":"2026-06-20T22:16:54.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":4},"url":"https://github.com/64035/node/issues/64035"}