GitHub Throttling Fixed via Mirror | Generated by AI
Fetch completed instantly. Let me now give you the full analysis.
Here’s what I found:
== ROOT CAUSE ==
The slowness is NOT ncat or the proxy. It’s China→GitHub ISP throttling.
Speed tests:
Direct to GitHub: 33 KB/s
Via Hysteria2 proxy: 59 KB/s
Via ghfast.top mirror: 2.4 MB/s (70x faster)
Both direct and proxied paths are throttled by Chinese ISPs. The Hysteria2 proxy (DMIT LA) doesn’t help because the bottleneck is the last-mile China exit, not the proxy itself.
== NCAT SOURCE CODE ==
Found at /Users/lzwjava/projects/nmap/ncat/
ncat_connect.c lines 621-921: do_proxy_socks5() — standard SOCKS5 handshake, no performance issues.
ncat.h line 140: #DEFAULT_TCP_BUF_LEN = 8192 (8KB relay buffer) — adequate for throughput.
ncat_posix.c: relay loop uses select() + 8KB reads — not the bottleneck.
ncat is clean. The problem is the network path.
== SOLUTION APPLIED ==
Changed Qwen2.5-Coder remote to use ghfast.top GitHub mirror: https://ghfast.top/https://github.com/QwenLM/Qwen2.5-Coder.git
Fetch now completes instantly.
NOTE: Repo has 296 local commits ahead of upstream (looks like a fork merge history). You may want to rebase or reset: git rebase origin/main # preserve local commits git reset –hard origin/main # discard local commits
== FOR ALL GITHUB REPOS ==
To apply globally (all GitHub repos use ghfast.top): git config –global url.https://ghfast.top/https://github.com/.insteadOf https://github.com/
Or per-repo only. Note: pushes won’t work through the mirror — you’d need to temporarily override for push: git push https://github.com/… main
Want me to set up the global config?