你好！所以我曾经抱怨，以太网在从Harmony 1.06切换到1.08之后就停止工作了。我在V1.0602下切换并继续开发。但是几周前，我发现了ETH的假错误（我的错），这让我觉得这是我在《和谐1.08》中出现大量问题的原因。在最初的尝试中，来自Microchip的某个灵魂试图帮助我，但是时间不够了，我不得不放弃。所以，现在，我放弃了一次新的尝试！事实上，它工作得更好。在我第一次尝试时，我甚至遇到了UDP连接的问题（99%失败）。现在，它们工作可靠，我得到NTP时间。这涉及DNS（池.ntp.org）。平工作。我确实有一个静态地址，所以没有经过DHCP测试（并且不需要）。但是我不能打开到（我的本地测试）服务器的TCP/IP连接。这里有一个奇怪的观察：第一个TCPIP_TCP_ClientOpen()的确给了我一个句柄0，但是它超时（超时，5秒）。因此，我用TCPIPpTCPUABLT（句柄，TRUE）中止该套接字，并在几秒钟后尝试。从那时起，我总是得到1的句柄。没必要尝试传输一些东西。下面是有趣的部分：在1.06中，我观察到了类似的情况：我试着在断开以太网电缆的情况下连接到服务器。当然失败了！但是，如果和我只有32的电缆连接失败，我得到了一个连接。再次：如果我有32（重复！）未能连接，连接与所有正确设置的基础设施不再可能。重新启动后，所有的工作…直到32次连接失败。也许这个观察是完全不相关的，也许它触发了某个人脑中的想法。无论如何，我真的很感激任何想法。尼克 以上来自于百度翻译 以下为原文 Hi! So I once complained, that the Ethernet quit working after having switched from Harmony 1.06 to 1.08. I switched back and continued developing under V 1.06.02. But a few weeks ago, I found a spurious bug (my fault) with the ETH that made me think that it was the reason for my massive problems in Harmony 1.08.Some kind soul from Microchip tried to help me during the initial attempt, but I ran out of time and had to give up.So now, I gave it a new try!In fact, it does work better. With my first attempt, I even had problems to get UDP-connections (99% fails). Now, they work reliably.I do get NTP time. That does involve DNS (pool.ntp.org). Ping works to. I do have a static address, so no DHCP tested (and not required).But I simply can't open a TCP/IP connection to (my local test) server.Here's a strange observation:The first TCPIP_TCP_ClientOpen() does give me a handle 0, but it times out (my time out, 5 seconds). So I abort that socket with TCPIP_TCP_Abort(handle, true) and try a few seconds later. From then on, I always do get a handle of -1. No need to try to transmit something.Now comes the interesting part: With 1.06, I observed something similar:I tried to connect to my server with the ethernet cable disconnected. Sure that fails! But if and only if I connected the cable within 32 fails, I got a connection. Again: If I had 32 (repeatedly!) fails to connect, a connect with all the infrastructure correctly set up was no longer possible. After a reboot, all works ... until 32 failed connections.Maybe that observation is completely unrelated, maybe it triggers an idea in someone's head.Anyhow, I really appreciate any ideas. Nick
2019-06-27 14:09
