New Rewrite for V2.0

[zpeters]

https://gist.github.com/sdstrowes/411fca9d900a846a704f68547941eb97

[zpeters]

COMING SOON

This version of speedtest is not supported. The underlying method of speed testing is not (and has not) been reliable. There is a new version of the speedtest protocol that i am actively working on a rewrite for. Expected release date is September 2018

[zpeters]

Initial testing release of v2.0 is out for the testing. See "releases" for downloads. The current "test" is hard coded and there are no options at the moment. I believe we'll end up needing a "tuning" to adjust the size and amount of data based on your internet connection. Please send me any feedback at [email protected] or through the issues

[polyzen]

?       github.com/zpeters/speedtest/cmd/speedtest      [no test files]
?       github.com/zpeters/speedtest/internal/app       [no test files]
?       github.com/zpeters/speedtest/internal/pkg/cmds  [no test files]
?       github.com/zpeters/speedtest/internal/pkg/comms [no test files]
?       github.com/zpeters/speedtest/internal/pkg/server        [no test files]

https://aur.archlinux.org/cgit/aur.git/tree/PKGBUILD?h=speedtest-zpeters#n28

[theinventor]

I want to test for you too, I need a build .exe so i don't have to compile.. (Or did you put it in the .zip?)

[theinventor]

I was using this a lot in some automation and just realized its been broken a while :)

[zpeters]

@polyzen i need a little more context here. Sorry i haven't been working on this for a while.

@theinventor Check this release link. It looks like i still had it as a draft...published it now with exes

https://github.com/zpeters/speedtest/releases/tag/2.0.0-a

[zpeters]

Putting this out there for comments/suggestions, etc. I think the root of the inaccuracy is that we don't (and will never know) the formula that is used on the backend of speedtest.net to extrapolate the download speed. I have reversed settings for my environment that are generally with-in 10-30 megs of the "true" download/upload speed. Basically what this comes down to is that depending on the speed of your connection and the size of the file you may get some overhead that makes the download "measure" slower (setup time for the download is static even though the transfer is quick) or the connection may have some sort of bursting. The best i can approximate is that we need to keep dynamically increasing the download or upload file size until the file takes a sustained X seconds to download or upload. WHere i am stuck right now is what is X and how do you dynamically tune the upload/download speeds accordingly.

Any thoughts, ideas, suggestions are welcomed

[polyzen]

@polyzen i need a little more context here.

When I run go test ./... as I used to, I get the "no test files" responses quoted above.