Andy Pan
Andy Pan
> I see @saghul gave his approval on this a long time ago but I guess that will not suffice to merge this PR. Since @bnoordhuis @vtjnash seemed to have...
Another two weeks have passed, still zero response... Helplessly and despairingly ping @bnoordhuis @vtjnash @saghul @libuv/collaborators
I have to admit that the latest commit has bloated this PR. The original version would be way more concise, maybe we should go back to using a placeholder file...
Hi @vtjnash, do you have a minute to review the latest commit?
这个问题主要是 gnet v1 里出现的，因为 v1 有编码器和解码器的功能，导致解码器在识别第一个特殊包的场景中不是很灵活，但是 gnet v2 移除了编码器和解码器，现在整个数据解析逻辑都要放在 `OnTraffic()` 里了，更加灵活，就没必要再设置一个所谓的 handler 了，用户自己判断一下第一个包然后处理就行了。
不需要每次都判断，你用 `sync.Once` 或者更简单地在 `c.Context` 里设置一个 bool，第一个包解析完之后就设个标志，下次判断一下这个标志，跳过就行了。
关闭之后有等待 2*MSL 时间之后再重新启动绑定地址吗？否则的话处于 `TIME_WAIT` 状态的连接无法重新绑定，除非使用 `SO_REUSEADDR`.
I think we may want to cancel those outdated CI jobs for this PR manually to make room for the latest one. @azat
[New failure](https://github.com/libevent/libevent/actions/runs/9145506779/job/25144530087) arose, caused by [Bug 106416](https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106416) - -Wint-conversion should be an error, not a pedwarn
> Still something wrong Yes, with -Wincompatible-pointer-types and -Wint-conversion being turned into error level on MinGW somehow, lots of code in `libevent` can't pass the GCC examination. Every time I...