Feedback for Test 17 (internal testing)


That might be coming soon just not from maidsafe


Yes there are at least several on the way


Browser:clearweb-urls on Linux.Mint18.1 spawn a Firefox process but that doesn’t see Firefox present - only if Firefox is already open do those links work. The Web Hosting manager:Help links do work.

One error:

Hosting Manager: Uploading on top an existing file gives - ~“Core error content already exists…” or some message like that which disappears too quick and without instead an option to overwrite[=> delete and upload that file].

and a few observations:

Authenticator: detailing all permissions being requested is too much information, and worrisome; by default the request should be simple - “do you trust/are you expecting, this app?”… with an advanced option to barf the specifics.

Hosting manager: could do with having its title in the body, not just relying on the container title. Rather than just “Home”, suggest Web Hosting Manager and a header for the list as Your PublicIDs or similar orientation hints.

Hosting manager:Help links could do with being on the page - menu help is a rare place people will explore. That said, the Browser has no [Help] where perhaps that should be there.

Unclear what affect Web Hosting manager:[Clear Access Data] has.

Hosting Manager: Navigation might be improved if the header was links… so _public/hello/www-root saw hello jump up to top with hello expanded; and _public/ to the top with all minimised.

Browser:File:[Toggle SAFE browsing] perhaps is redundant.

Browser:History suggests no history and I wonder that kind of base capability will be expected in future.

One good improvement seems to be that refresh following an updated file, does work without need for some http cache header forcing that.



and in reply to certain of Readme questions:

Is network access restricted when IP is not whitelisted


Is network access limited to only the IP that is whitelisted

Appears to be.

Is access allowed after switching IP

Not obviously.

If IP is swapped while apps are currently running, do they lose network connectivity within roughly 60 seconds

Yes - switch to a VPN, sees the Authenticator [You have been disconnected from the network.Reconnect now.] message, which is clear and makes sense of what that (A) logo is for.
BUT if Authenticator is not open, that change is not obvious. The Authenticator does not pop in front; the tab doesn’t flash; the favicon doesn’t change colour - and the Hosting manager continues for a while to suggest loading and then “No files found please upload” but then obscure error if that attempted. “Core error: Routing interface error”.

If the popup for authentication works, should the disconnect alert not act in the same way?

Is losing internet connectivity reflected in the browser accurately

Yes but the delayed reaction to disconnect seems unecessary.

Once connection is established, does the browser allow a manual reconnect to the network without needing to login again

Yes but can that not be also automatic?


Hosting Manager: zero size files not properly catered for.
Appears to stall with “Uploading files 1/1 - NaN%” and risk of that from a directory upload with zero file.
Also, the cancel for that doesn’t appear to work. So, this requires crashing out of Hosting Manager to resolve that.
I wonder, zero file placeholder should be possible.

for later - I wonder the packaging of the Browser/Hosting Manager/etc could be made simpler… just a single binary would be less intimidating to noobs and perhaps the ABC instructions need to be made explicit so everyone feels brave enough to jump in - knowing what binaries they need to click on to run among the clutter of files in the unzipped folders atm might be too much?.. perhaps Windows/OSx is simpler but Linux is a step into the unknown for some.


Any more invites for the test network? Will it be public soon?


Not yet, there is already quite a bit of feedback that needs resolved, hopefully when the binaries are updated(with bug fixes for main issues reported), things can scale up quicker.

I can reproduce this issue too. @fergish / @Nigel Just as a test and possible temp workaround, can you try the following please to see if it resolves it. Not yet sure of the source of the issue but it can at least confirm if the issue is the same across the machines.

So “if clicking the authenticator icon in the toolbar does nothing”:

  1. Close the browser from the dock. Ensure it is closed and not just hidden
  2. Open terminal
  3. Type cd followed by space.
  4. Open finder to the folder where you have the browser downloaded. Drag and drop the folder to terminal such as
  5. It will add that path to terminal. Hit Enter
  6. Type / copy paste ./SAFE\\ Browser and hit enter.
  7. It should launch the browser.
  8. Try clicking the same button now. Expect this to work now and open authenticator if it is the same issue
  9. Close the browser from dock again
  10. Open browser by just double clicking the app (not via terminal)
  11. Now click authenticator toolbar button and again we can expect this to work now too.

If that works, should at least let you try the app while we figure out what caused this issue.


Absolutely, will do this tonight and get back to you. Can’t wait to be a part of the test again!


k small update, that didn’t take long once it was reproducible. Apparently there was a missing locales file which caused the issue which when running the browser from terminal created the locales file in the package itself.

So even without running the browser from terminal, just manually creating the locales file had the issue resolved. So such as at step 6 from above, just need to run mkdir -p ./SAFE\ && echo "{}" > ./SAFE\

Not sure why it was missed from the build process if it is expected pre compiled, but that does resolve it. If you dont want to muck about with the manual patch, I’ve removed the old binary and uploaded the patched app with updated signatures in the same location, the zip there should now be named safe-browser-v0.2.1-patched-osx-x64

Hope that helps and cheers for submitting the issue. something that can get added to the build process now hopefully :slight_smile:


@Viv I enjoy mucking about and glad you posted solution but I am very eager to get at it :smile:


Hi guys, is the google drive link already public or not? I’ve been not so active last weeks… I only have few hours a week and doing to much reading on many things…


Why did it go from public - hundreds of testers - to a more controlled internal trickle style testing, then back to a larger tester pool? I figured it would go the other way around. Was there another recent refactoring that I missed in the dev updates or something?

Anyway, awesome work SN devs, designers and PMs et al., and looking forward to participating again when it’s less lottery style testing. :wink:


Nope its not public yet.

Because security features were not yet in place and the network came under attack there is no way to test in public until security features have been implemented and tested.


wow internal testing?? Interesting path you guys took. I understand why this path was chosen. The team wants to iron out the bugs before rolling out to the public for alpha 2 network.

Hope everything went well.


I can’t stress it enough. The interface should be designed assuming user has little to NO technical experience. NONE NADA ZILTCH.

All advanced terminology and options should be hidden until explicitly chosen. A clear warning should also accompany the list of advanced options.

Again the UI should be designed as if it is intended to be used by a 4 - 6 year old. Such a UI would have the broadest reach. Allowing almost anyone to enjoy this fantastic network! :blush:


Okay looking forward to play with the binaries… UI also starts to mature I saw, nice…


Better than that perhaps, is that there is as little UI necessarily attributed to the network as possible… everything else can spawn as many varieties as needed, to match the variety of user-base.

To which, perhaps need at an early stage to consider option on some route to skins for applications.


You, @JPL, @davidpbrown, myself and I’m sure others feel very strongly about this. I think this stuff needs to be addressed asap. I’m just uncertain if the team is perhaps holding off on that a little until they hire more front end for when allocation of hard drive space etc, that will be critical in adoption.


Internal testing is not something that is new tbh. Think anything and everything that is written in code goes through internal testing first, starts from the time someone writes some code before it gets committed to the point of CI, soak testing, getting different teams to try it and then normally concludes in starting with people from the forum keeping up with dev updates trying it and then more and more people ofc until some bug arises anywhere in that flow and it scales back and starts again :slight_smile:

This time the slight change has been the last but one leg where there is quite a lot of repetitive testing and exhaustive checks across the platforms required for quite considerable change set MD and other things, we just also got in few trusted members of the forum to help test and debug the same along with the internal team(only a few of us internally anyways so not like a large team). So far this has worked out great as you can see from this thread where some of the obvious hurdles are raised and hopefully addressed by the team and allows them to continue with their code too.

It is also by no means meant to be something exclusive as all this is expected to yield is when things scale up to say more forum members, people have a better user experience of the same with the opening obvious issues addressed.