can't build working code #5
Comments
|
So the .deb I provided works fine, but when you build for yourself it doesnt? Odd. I just use "go get -v" I assume you are using the latest .deb I have posted? |
|
Yes, that's the situation. Here are two sessions - first, building and running, then running the installed .deb. As you can see, the first fails and the second one works: addy@blap:de-ip-hdmi$ go build -v addy@blap:de-ip-hdmi$ sudo de-ip-hdmi -interface enx0050b64d263a |
|
Interesting, Can you checkout d32cbab, build that, and see if that works? |
|
Yes, that revision builds and works. |
|
Woops! Okay, That's irritating, I suspect there is some commit that went a little wrong there, Could you do me a favour and "git bisect" between that and HEAD and find the offending commit? CC: @Shugyousha |
|
Sure, no problem... |
|
On Tue, Jul 12, 2016 at 12:52 PM, Ben Cox [email protected] wrote:
Hm, I think I screwed up this line. packet = append(macbytes[:6], packet[:6]...) it should be (note the second colon): packet = append(macbytes[:6], packet[6:]...) Can you check if it works with this change? If it does I can send a Thanks for the bug report! |
|
@Shugyousha yes, that fixes it! Thanks! (BTW, I say 'fix', but I also have a problem that although the video sender now starts and I can see the packets with wireshark, the ffmpeg process hangs on the FIFO so you never get any output from de-ip-hdmi itself (this is also true of the .deb package). I have the same problem with my own python code, so I assume it's something to do with FIFOs as it's affecting both of us. Since I want to transcode when I'm done I just write to separate audio/video files, so I'm not that bothered but worth mentioning in case anyone else is also having that issue). |
|
Here is the patch... |
|
@AdamLaurie are you plugging the device directly into your system, or is it though going though a switch? |
|
On Tue, Jul 12, 2016 at 09:56:15AM -0700, Adam Laurie wrote:
I thought I got the subslicing right on the first try but I obviously
I don't have the hardware so I can't test it but to me it sounds like If we close videodis at the end of main() and close the WriteCloser in diff --git a/main.go b/main.go
func randString(n int) string {
func BroadcastWakeups(ifname string, sendermac string) {
|
|
@benjojo I've tried both and it makes no difference. The patch fixes the original issue in that the sender now sees the keepalives and starts streaming. |
|
If you do a tcpdump, do you see the packets arrive at your interface? |
|
Yes (well, wireshark, as previously mentioned). The problem is definitely with the FIFO. In python, I can stream either the video or the audio to ffmpeg via a pipe/stdin, or I can write both the video and audio to separate files, but if I try to write either via a FIFO it hangs on the the first write. I'm guessing it's the same problem with the go code. |
|
Are you playing sound though the HDMI input? If not then use the |
|
Yes, I have both audio & video. Again, it all works fine if I write to files instead of FIFO. |
|
BTW, with reference to audio, you mention in your blog: "I found that the audio sample rate had changed to 44100hz, down from the 48000hz that danman reported. (Though this could be dependent on the input?)" I can confirm after some experimentation that at least here in the UK I'm still getting 48KHz, so I suspect it's just whatever the input is set to as in my case that matches the sources. |
|
Okay, Can you raise different issues for these things? We have solved the build issue, I can see two issues here, not being able to send data into a FIFO, and the audio sampling issue |
|
On 14/07/16 12:35, Ben Cox wrote:
The audio sampling thing is not an issue with your code - it was just an I will re-raise the FIFO issue though, and anything else I come across. cheers, AdamAdam Laurie Tel: +44 (0) 20 7993 2690 |
ghost
mentioned this issue
Hi, I'm new to go so sorry if this is a really dumb question, but do I have to do something special to create a binary prog from main.go? I've tried what seems to be the obvious choice: "go build main.go" but the resulting binary executes but doesn't actually work (the sender never starts streaming). The ready built package for Ubuntu/Debian works fine so I know my kit is all wired up correctly...
The text was updated successfully, but these errors were encountered: