YISH
YISH
@pymumu 大佬你这里的逻辑是怎样的?查询 IPv6 ping 不成功,但是对应后台查的 IPv4 ping 可以,直接给 IPv6 的查询返回 SOA?
好的,谢谢。@pymumu 这样200ms不导致响应慢吗? 另外,命中次数是指多个上游返回的ip重叠数最高的那个?
- 增加了 200 ms 切换下一个 ping 。每个 ping 的超时时间 2 秒。 https://github.com/mokeyish/smartdns-rs/blob/237042fffd24a3235329506d003612a2c8545120/src/dns_mw_ns.rs#L363-L364 - ping 不通,选命中 IP 重复次数最多的那个。 https://github.com/mokeyish/smartdns-rs/blob/237042fffd24a3235329506d003612a2c8545120/src/dns_mw_ns.rs#L321-L325 - 三种响应模式 https://github.com/mokeyish/smartdns-rs/blob/237042fffd24a3235329506d003612a2c8545120/src/dns_mw_ns.rs#L203 https://github.com/mokeyish/smartdns-rs/blob/237042fffd24a3235329506d003612a2c8545120/src/dns_mw_ns.rs#L288 https://github.com/mokeyish/smartdns-rs/blob/237042fffd24a3235329506d003612a2c8545120/src/dns_mw_ns.rs#L330
@dream10201 你下载这个版本试试,https://github.com/mokeyish/smartdns-rs/actions/runs/9240086699 双栈优选和测速部分都重写了。 @pymumu 我再确认下。双栈优选 - 当不开启 `dualstack_ip_allow_force_aaaa`。 - 查询 IPv4 时,是不是不作任何处理?(你文档说:`smartdns默认情况下总是会返回IPV4地址`) - 查询 IPv6 时,当对应 IPv4 快过阈值,IPv6 返回 SOA。使之只能查询到更快的 IPv4 地址。 - 当开启 `dualstack_ip_allow_force_aaaa`。 - 查询 IPv4 时,发现存在 IPv6...
那就是解决了? @dream10201
这个版本是参考 candle 官方改的 https://github.com/huggingface/candle/blob/main/candle-kernels/src/indexing.cu https://github.com/mokeyish/candle-ext/blob/main/src/kernels/indexing.cu candle 官方他们说是要减少算子,这样更方便适配到其他硬件平台,所以才写个扩展库写了这个。 可能需要参考 pytorch 的源码看看,它为什么那么快。
@djc Hi, I implement the trait `RuntimeProvider` to support for querying over socks5. But current code will never execute the code after `await`. This PR sovles the problem. https://github.com/hickory-dns/hickory-dns/blob/main/crates/resolver/src/name_server/connection_provider.rs#L92 ```rust...
> Is this comment what this is resolving?: > > ```rust > // TODO: this immediate poll might be wrong in some cases... > match (*self.closure)(bind_addr, self.name_server) > .as_mut() >...
I haven’t received the 1.6 update on my MacOS yet, I’ll look into it and find an alternative method when it’s available.
where did you download the version 1.6 from? from here, the latest version is 1.5.12 https://github.com/obsidianmd/obsidian-releases/releases