[Bug] مقایسه سرعت و کیفیت اتصال بین ۱.۶ و ۱.۵

[0xSYJiK]

Problem Summary / خلاصه مشکل

من در ورژن ۱.۵ با ۵ اکانت هر کدوم ۱ ایدی سرعت خیلی بیشتری نسبت به ورژن ۱.۶ با ۵ اکانت میگرفتم و سرعت ۱.۶ تا ۴ برابر کمتر هستش برام نسبت به ۱.۵

Reproducible Steps / مراحل بازتولید

ورژن ۱.۵: ۵ عدد ایدی
ورژن ۱.۶: ۵ اکانت (3/1 , 2/1 , 2/1 , 1/1 , 1/1) تعداد ایدی نسبت به اکانت

Environment Details / مشخصات محیط

windows
Processor 11th Gen Intel(R) Core(TM) i7-11800H @ 2.30GHz (2.30 GHz)
System type 64-bit operating system, x64-based processor

Expected vs. Actual / انتظار در برابر نتیجه واقعی

انظار سرعت بالا تر با توجه به بیشتر بودن تعداد ورکر بیشتر

ورژن 1.6 ۵ اکانت (3/1 , 2/1 , 2/1 , 1/1 , 1/1) تعداد 20 ورکر

در مقایسه ورژن 1.5 و 5 اکانت 15 ورکر

Logs / Screenshots / لاگ‌ها یا تصاویر

22:31:43  STATS   INFO    accounts=[bonhorse today=195 script=63 | hades today=98 | kerem today=294 script=9130 | syjikg today=195 script=3831 | troy today=98 script=1916]
22:32:08  CARRIER INFO    RST from server for session ce990229; closing
22:32:08  CARRIER INFO    RST from server for session 9ce1deea; closing
22:32:12  CARRIER INFO    RST from server for session a0775edb; closing
22:32:13  CARRIER INFO    RST from server for session 01880b7c; closing
22:32:17  CARRIER INFO    RST from server for session 2c885689; closing
22:32:23  CARRIER INFO    RST from server for session 8f6ce6ed; closing
22:32:35  CARRIER INFO    endpoint AKfycb...47Rycb blacklisted for 5m0s (still rotating across 8 others)
22:32:35  CARRIER INFO    relay returned non-batch payload via AKfycb...47Rycb (attempt 1/2); retrying alternate script
22:32:36  CARRIER INFO    endpoint AKfycb...47Rycb recovered (back in rotation)
22:32:42  CARRIER INFO    RST from server for session 60467235; closing
22:32:42  CARRIER INFO    RST from server for session 5f26dee4; closing
22:32:43  STATS   INFO    active=167 sessions=335/168 frames=1019/898 bytes=1.3MB/6.3MB polls=1174/0 rst=26 endpoints=8/9
22:32:43  STATS   INFO    endpoints: AKfycb...47Rycb@kerem ok=135 fail=1 today=137 script=3616 | AKfycb...T3zkvM@kerem ok=95 fail=1 today=97 script=3619 bl=3m48s | AKfycb...dwTURA@kerem ok=134 fail=0 today=135 script=1895 | AKfycb...Sw69uw@syjikg ok=135 fail=0 today=136 script=1917 | AKfycb...hv0kOQ@syjikg ok=135 fail=0 today=136 script=1914 | AKfycb...5qp2JQ@hades ok=135 fail=0 today=136 | AKfycb...2Uzniw@troy ok=135 fail=0 today=136 script=1916 | AKfycb...ZVhWfg@bonhorse ok=135 fail=0 today=136 script=35 | AKfycb...56GYXQ@bonhorse ok=135 fail=0 today=136 script=28
22:32:43  STATS   INFO    accounts=[bonhorse today=272 script=63 | hades today=136 | kerem today=369 script=9130 | syjikg today=272 script=3831 | troy today=136 script=1916]
22:32:46  CARRIER INFO    RST from server for session ac355c95; closing
22:32:48  CARRIER INFO    endpoint AKfycb...dwTURA blacklisted for 5m0s (still rotating across 8 others)
22:32:48  CARRIER INFO    relay returned non-batch payload via AKfycb...dwTURA (attempt 1/2); retrying alternate script
22:32:49  CARRIER INFO    endpoint AKfycb...dwTURA recovered (back in rotation)
22:32:51  CARRIER INFO    RST from server for session 76e27e93; closing
22:32:52  CARRIER INFO    endpoint AKfycb...dwTURA blacklisted for 5m0s (still rotating across 8 others)
22:32:52  CARRIER INFO    relay returned non-batch payload via AKfycb...dwTURA (attempt 1/2); retrying alternate script
22:32:53  CARRIER INFO    endpoint AKfycb...dwTURA recovered (back in rotation)
22:32:57  CARRIER INFO    endpoint AKfycb...47Rycb blacklisted for 5m0s (still rotating across 8 others)
22:32:57  CARRIER INFO    relay returned non-batch payload via AKfycb...47Rycb (attempt 1/2); retrying alternate script
22:32:58  CARRIER INFO    endpoint AKfycb...47Rycb recovered (back in rotation)
22:33:06  CARRIER INFO    RST from server for session 485a4791; closing
22:33:07  CARRIER INFO    RST from server for session a040a751; closing
22:33:14  CARRIER INFO    RST from server for session 5ce0c5a1; closing
22:33:15  CARRIER INFO    RST from server for session 0693e837; closing
22:33:16  CARRIER INFO    RST from server for session 088b56fa; closing
22:33:19  CARRIER INFO    RST from server for session cca55291; closing
22:33:20  CARRIER INFO    RST from server for session 5428bab6; closing
22:33:25  CARRIER INFO    endpoint AKfycb...dwTURA blacklisted for 5m0s (still rotating across 8 others)
22:33:25  CARRIER INFO    relay returned non-batch payload via AKfycb...dwTURA (attempt 1/2); retrying alternate script
22:33:27  CARRIER INFO    endpoint AKfycb...dwTURA recovered (back in rotation)
22:33:31  CARRIER INFO    endpoint AKfycb...47Rycb blacklisted for 5m0s (still rotating across 8 others)
22:33:31  CARRIER INFO    relay returned non-batch payload via AKfycb...47Rycb (attempt 1/2); retrying alternate script
22:33:31  CARRIER INFO    RST from server for session de626388; closing
22:33:31  CARRIER INFO    endpoint AKfycb...hv0kOQ blacklisted for 3s (still rotating across 8 others)
22:33:32  CARRIER INFO    relay returned HTTP 404 via AKfycb...hv0kOQ (attempt 1/2); retrying alternate script
22:33:32  CARRIER INFO    endpoint AKfycb...hv0kOQ recovered (back in rotation)
22:33:34  CARRIER INFO    endpoint AKfycb...dwTURA blacklisted for 5m0s (still rotating across 8 others)
22:33:34  CARRIER INFO    relay returned non-batch payload via AKfycb...dwTURA (attempt 1/2); retrying alternate script
22:33:36  CARRIER INFO    endpoint AKfycb...dwTURA recovered (back in rotation)
22:33:39  CARRIER INFO    endpoint AKfycb...hv0kOQ blacklisted for 3s (still rotating across 8 others)
22:33:39  CARRIER INFO    relay returned HTTP 404 via AKfycb...hv0kOQ (attempt 1/2); retrying alternate script
22:33:39  CARRIER INFO    relay returned HTTP 404 via AKfycb...hv0kOQ (attempt 1/2); retrying alternate script
22:33:40  CARRIER INFO    relay returned non-batch payload via AKfycb...47Rycb (attempt 1/2); retrying alternate script
22:33:41  CARRIER INFO    endpoint AKfycb...hv0kOQ recovered (back in rotation)
22:33:41  CARRIER INFO    RST from server for session 83be2b8d; closing
22:33:42  CARRIER INFO    endpoint AKfycb...dwTURA blacklisted for 5m0s (still rotating across 8 others)
22:33:42  CARRIER INFO    relay returned non-batch payload via AKfycb...dwTURA (attempt 1/2); retrying alternate script
22:33:42  CARRIER INFO    endpoint AKfycb...dwTURA recovered (back in rotation)
22:33:43  STATS   INFO    active=260 sessions=511/251 frames=1402/1238 bytes=1.5MB/8.2MB polls=1492/0 rst=37 endpoints=7/9
22:33:43  STATS   INFO    endpoints: AKfycb...47Rycb@kerem ok=166 fail=4 today=171 script=3616 bl=29m57s | AKfycb...T3zkvM@kerem ok=95 fail=1 today=97 script=3619 bl=2m48s | AKfycb...dwTURA@kerem ok=172 fail=5 today=178 script=1895 | AKfycb...Sw69uw@syjikg ok=177 fail=0 today=178 script=1917 | AKfycb...hv0kOQ@syjikg ok=171 fail=3 today=175 script=1914 | AKfycb...5qp2JQ@hades ok=178 fail=0 today=179 | AKfycb...2Uzniw@troy ok=178 fail=0 today=179 script=1916 | AKfycb...ZVhWfg@bonhorse ok=178 fail=0 today=179 script=35 | AKfycb...56GYXQ@bonhorse ok=177 fail=0 today=178 script=28
22:33:43  STATS   INFO    accounts=[bonhorse today=357 script=63 | hades today=179 | kerem today=446 script=9130 | syjikg today=353 script=3831 | troy today=179 script=1916]
22:33:44  CARRIER INFO    endpoint AKfycb...hv0kOQ blacklisted for 3s (still rotating across 8 others)
22:33:44  CARRIER INFO    relay returned HTTP 404 via AKfycb...hv0kOQ (attempt 1/2); retrying alternate script
22:33:45  CARRIER INFO    endpoint AKfycb...hv0kOQ recovered (back in rotation)

[SaeedDev94]

Can be related: #85

[Kianmhz]

v1.7.0 (release notes) ships the fix for the v1.5→v1.6 regression you're seeing.

Two root causes, neither of which we had nailed when v1.6 went out:

1. EnqueueInitialData was prepending bytes instead of appending (#147). When a SOCKS5 app wrote in several small chunks before the SYN drained (TLS handshakes, HTTP request lines, anything with a length prefix), the bytes shipped to the upstream in reverse order. The upstream silently misparsed or errored, and the bench harness's sized sink misread the body — which is why this hid in plain sight since the optimization landed. This is the real cause of "uploads feel broken / sites take forever to load on v1.6."
2. The worker pool collapsed under v1.6's per-bucket scaling (#146). v1.5 ran workersPerEndpoint × len(endpoints) workers; v1.6 changed it to (workersPerEndpoint + idleSlotsPerBucket − 1) × bucketCount, which starved configs with multiple deployments per Google account. v1.7 restores the v1.5-style scaling while keeping v1.6's per-account safety semaphore (so relay returned non-batch payload #56 stays fixed).

Plus a default tweak: idle_slots_per_bucket is now 2 (was 1) for the recommended multi-deployment setup.

Upgrade path: pull v1.7.0 client + server binaries, redeploy apps_script/Code.gs (optional but recommended — picks up the relay-loop guard from #129). No config changes required.

Closing this — please reopen or comment on the v1.7.0 release if you still see the regression after upgrading.