Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
    Emil Thelin
    @gummiboll
    Haven't really tried any other packages except the vanilla-go-way, ie net/http/httptest
    Emil Thelin
    @gummiboll
    I think httptest is probably the best way to go. However, the test-server gives a http://127.0.0.1:<random port number>-url and NewPing() doesn't support a url with a port specified in it so it will throw an error. Would however be an easy fix and I think it would be good if hping supports a custom port being specified, ie mylg hping hostname:port
    Emil Thelin
    @gummiboll
    Committed a PR that fixes the test and the host:port-thing. :)
    Mehrdad Arshad Rad
    @mehrdadrad
    very cool, now it supports port and we don't use gock :) Thanks Emil! i'll merge the PR tonight.
    Emil Thelin
    @gummiboll
    Np :)
    Emil Thelin
    @gummiboll
    mehrdadrad: Regarding mehrdadrad/mylg@a6aa5e1
    mehrdadrad: Could it be that you generated the statik-file with a old copy of statik? See rakyll/statik#17
    Mehrdad Arshad Rad
    @mehrdadrad
    @gummiboll looks gofmt is not happy w/ updated statik.go as there is binary inside. This file should be skipped for test gofmt
    Emil Thelin
    @gummiboll
    @mehrdadrad : Yeah, but if you run gofmt with -d you see what it complains about and that is the same thing that PR I mentioned above fixes. I think that you might be running a old version of statik

    When I run gofmt with -d I get: ```@@ -1,7 +1,7 @@
    package statik

    import (

    Meh, doesn't github-markdown work with gitter? :(
    So if you update/build statik and then generate a new statik.go-file from the new binary I think it will pass gofmt. :)
    Mehrdad Arshad Rad
    @mehrdadrad
    Ok thanks @gummiboll I'm already behind the wheel, I'll fix it today :smile:
    Emil Thelin
    @gummiboll
    Tried it and yes, with the latest version of statik, the statik.go file passes gofmt. :)
    Mehrdad Arshad Rad
    @mehrdadrad
    Great, thanks for your help @gummiboll
    Emil Thelin
    @gummiboll
    Np :)
    Emil Thelin
    @gummiboll
    @mehrdadrad: Was trying out the new webgui for mylg and noticed it was a bit slow, checked firebug and it kinda looks like the npm includes from unpkg.com are getting a bit out of hand. Rendering /#/trace generates about 500 requests and most of them are against unpkg.com. Some are 200 but quite a few are 302:s also. Not sure if you've seen it also but might be good to look into.
    Mehrdad Arshad Rad
    @mehrdadrad
    I'll check how can I fix that. Last night I noticed this issue: #17738 github.com/golang/go
    Mehrdad Arshad Rad
    @mehrdadrad
    @gummiboll forgot to mention u at last message
    Emil Thelin
    @gummiboll
    @mehrdadrad: Weird. Tried it and I get the exact same symptom as you; works on macos but not on a linux/amd64
    Mehrdad Arshad Rad
    @mehrdadrad
    @gummiboll I think I found how to fix it, I'll commit it tonight
    Emil Thelin
    @gummiboll
    Great :)
    Mehrdad Arshad Rad
    @mehrdadrad
    @gummiboll Emil when you get a chance try the web trace I think it should be much better
    Emil Thelin
    @gummiboll
    @mehrdadrad: Looks a lot better now! But the trace itself doesn't seem to work for me. Only the first hop (ip/host) is printed and the other hops get 100% packetloss. Not sure if its my FF that bugs or something.
    Can check more thoroughly later.
    Emil Thelin
    @gummiboll
    @mehrdadrad: Yeah, tried both in FF and Chrome. Only the first and last hop gets printed with host/ip/asn/last/avg/etc. All other hops are empty with 100% packetloss. Getting some js-exceptions also
    Mehrdad Arshad Rad
    @mehrdadrad
    @gummiboll that's weird I'll try w/ another laptop and works. I'll test w/ other to see how it works. did you try UDP? maybe NAT issue
    Emil Thelin
    @gummiboll
    @mehrdadrad: Tried both UDP and ICMP. When I run it with ICMP it prints the first and the last hop as I said above and if I run it with UDP it only prints the first hop and then never seem to never reach the last hop
    Hmm. I seem to get the same result if I run it from cli mode which I'm fairly certain has worked before
    @mehrdadrad: I'm running macos sierra/go version go1.7.1 darwin/amd64
    Emil Thelin
    @gummiboll
    @mehrdadrad: Works without a glitch on linux tho
    Emil Thelin
    @gummiboll
    @mehrdadrad: Tried it with a mobile connection and then it works like a charm. The weird thing is that it has worked before and there is nothing extraordinary in front of my laptop. Just a airport. "regular" traceroute works without a glitch also
    Emil Thelin
    @gummiboll
    Heheh, sorry for wall of text :D
    Can do some more digging later and see if I can find what's causing it but now its time for dinner. Been a long day, woke up this morning with snow to my knees and the tractor broke
    Mehrdad Arshad Rad
    @mehrdadrad
    @gummiboll that's very interesting, I need to deep dive to this issue. thank you so much for testing. have fun w/ snow unfortunately we don't have snow @ Los angeles anymore. we need to drive 1-2 hours to see snow.
    Emil Thelin
    @gummiboll
    @mehrdadrad: Found the commit that's causing it. See http://pastebin.com/1uC2MZDb
    So its this one: mehrdadrad/mylg@4431ad2
    Not sure why it causes it to fail tho
    Emil Thelin
    @gummiboll
    Hehe, snow is a blessing and a curse. :D
    Mehrdad Arshad Rad
    @mehrdadrad
    @gummiboll I think you're behind a stateful firewall, i just did minor changes pls check when you get a chance. Thanks
    Emil Thelin
    @gummiboll
    @mehrdadrad: Unfortunately it didn't fix it. :/ I get exactly the same symptom as I got before
    Mehrdad Arshad Rad
    @mehrdadrad
    @gummiboll ok Thanks Emil for all testing, I need to find a similar network to see how I can fix it. the main change that I did @4431ad21c37398f955dcf75a48e969fc3f9d5ca0 is put IP identification instead of kernel at each packet to have more control to get back the ICMP messages. maybe your access point change or manipulation ip packet header before translating if nat enabled
    Emil Thelin
    @gummiboll
    @mehrdadrad: the weird thing is that im just sitting behind a vanilla AirPort with no special configuration whatsoever. Nerver noticed any network-weirdness either. But something clarely is bugging
    @mehrdadrad: I can spin up a virtual Linux box tomorrow and see if i get the same issue, if so i can give you access to it so you can debug if you like?
    Might be harder to debug tho since it will essientially be double NAT:ed
    Mehrdad Arshad Rad
    @mehrdadrad
    @gummiboll the access that would be great. i think that's related to packet identification that I added recently but I would like to see what change happen to packet identification. maybe tcpdump capturing help.
    Emil Thelin
    @gummiboll
    @mehrdadrad: Haha, this is the weirdest thing ever.
    @mehrdadrad: Spun up a linux-box in virtualbox on my laptop and on it, mylg trace works