Added WRNG to some sections (WRNG: Dash Test should be mentioned here) And added info about windows 8 "neubot" app

Added WRNG to "Implemented Tests" section (WRNG: Dash Test should be mentioned here)

In the quest to put under control again the Neubot tree, let's get rid of "the most annoying notification strategy ever" (cit.)

This notification strategy was annoying in 2012, I can't imagine how much it would drive people crazy in 2017.

Per the agreement with Nexa, I am creating the "final" and frozen release of Neubot, and we will not need this notifier in there.

This pull request removes the master server's source code. I have recently re-implemented the master server in Go, have deployed it and the deployment seems to be working quite well. Hence, there is no need to keep master server code in here.

This pull request improves Neubot's manpage. The main motivation for improving the manual page was that I wanted to document data saved by the dashtest test. Yet, on the go, I noticed some other places in which the manpage could be improved as well.

As promised, you woke up and it's in your inbox.

I hope it pleases you. The option you presented me as the "first possibility" seemed much more appropriate, so you will notice I have done all the work based on what you had already written and sent me (and also on another instance of this text I found within the website).

I have not been able to shrink it to anything shorter than this without great loss of quality. This was the casewith the paragraph I had written on friday (and did not show you).

Anyhow, I have tried not to be perfectionist, as you will revise it, and we may change it many times more, but I hope it pleases you a bit already.

Best,

Felipe.

this isn't so much an issue as it is curiosity. I have been using the speedtest offered by Ookla and there, with my server being specified to a local one, the speed shows approximately 25 mbps for download. However, on the Neubot speedtest, with the server being a "remote server", the speed is shown to be close to half an mb per second (download). Which one is more accurate of a result and why? Because I did try shifting servers in the aforementioned Ookla site and after connecting to a distant server, the download speed shown was 5 mbps. Although not as low as shown with Neubot, nonetheless I can see the decrement. How is the remote server picked?

How do we know if these connection failures are the result of an ISP throttling our service? As in, they are blocking the connection to mlabs?

I've tried other throttling check software, and invariably have the same results.

Neubot Log.txt

I run Neubot on a server that has a VPN endpoint as the default route.

That VPN is slower than my 'normal' default route (my ISP), and I'd rather Neubot doesn't use it.

So say my ISP's default route is 192.168.1.254 on my local network, and I have an interface (eth0 for example) on that same subnet.

Neubot uses tun0 (my VPN interface) and uses the VPNs next hop.

So to be able to specify interface=eth0, or nextHop=192.168.1.254 would be really useful to me. I've been parsing the log files (manually I might add, I should automate that if this can't be done) to gather a list of IPs that Neubot connects to and then adding a route for it via 192.168.1.254.

Hi, I see a huge difference between speedtest.net and Neubot. On Italy fastweb I've got 600k on NB and 15M on SP. That's repeated several times (speed slighlty changes each time but the difference is still remarkable). Any suggestions? thanks alessio valentini win10 64bit Intel on Opera&FF

I have been looking at my Neubot 0.4.16.9 Web interface's "Results" tab. The results look too good. They do not seem to reflect the many, many times I have been connected to my Internet Service Provider''s (ISP's) modem but could not get significantly farther into the internet. I looked at the "Log" tab. I see many, many "ERROR" lines. There are many, many lines like the following lines. Is Neubot simply ignoring such errors instead of using a "0" (zero) as the connection speed or bandwidth for these instances?

• connect(): cannot connect to master.neubot.org:9773
• gaierror: [Errno 11001] getaddrinfo failed
• runner_rendezvous: connection failed: None
• connect(): cannot connect to releases.neubot.org:80
• * speedtest: connection failed

It is not obvious to me whether Neubot's includes such "0" speed/bandwidth instances in its calculations of bandwidth. But if it is not, then it should. If it does not, then its "Results" may look better than they should.

Perhaps Neubot should routinely use traceroute (Windows tracert) and log what the hops are for good connections. Then, when it encounters a connection error, it could do a trace route and try to determine where the problem is occurring. If, based on those results, Neubot can determine that the computer appears to be connected but just can't reach the regular servers or third-party servers that are usually accessible, then Neubot should record a valid "0" for the connection speed/bandwidth.