hellozeronet / zeronet Goto Github PK

Cons:

• It could be irratating if you just restart it or you want to put in autorun

Possible solutions:

• Open by default but can be disabled via --dont-open-browser command line parameter
• Create an another file that starts zeronet.py and opens the browser too: start.cmd
• After starting zeronet.py display a message: -- Press enter to open http://localhost:43110/ in your browser --

Reported on redit:

Didn't work:
@util.Noparallel(blocking=False)
AttributeError: 'module' object has no attribute 'Noparallel' -- Error happened, press enter to close --

maybe collision with other python package

My first try:
I created a test site with and published it, and still get SEEDING UPDATE FAILED.

No real content yet there:
http://127.0.0.1:43110/19u3W1KB8nzzEW3EiQYC1eNfQ6Q53o9Gj7

What kind of files can I use for the site?
Is there a tutorial for that?

What will be ZeroMarket and ZeroBay??? or why is it gray?

My version of miniupnpc doesn't understand "-e ZeroNet" but the rest of the command is right. Maybe there could be a test or a try/except to do it without the -e option.

Zeronet is taking a lot of time to load other website's like the one's present in the Zeroboards messages. I recommend that you should include a website load progress bar with some data in it so that user's can know how much percent the website has been loaded. It is very annoying to look at the connecting page without any estimate of how much data is left to be downloaded. Along with that the ZeroSearch (The search engine for zeronet) is not working. The page says internal server error. And one last thing is that please give some more documentation so that the developers could help you build it better.

When navigating a ZeroNet site the top URL in the browser doesn't change. It stays as the address for the top level of the site. This makes creating internal links to sites in ZeroNet difficult.

You can right click on a link and copy/paste it. But this leaves the "/media" path in the URL and when you visit it it displays at the top level, not in the iframe with the Websocket magic around it.

Would it be possible to change the top level URL when navigating? Maybe using a fragment identifier that the top level JS parses to navigate the iframe?

I've mostly been using Zeronet in Firefox/Chrome for Mac, and they've been working fine. However, when trying to connect in Safari, there are a few problems. One, the flipping red square is white with a 1px red border (after it drops; as it's dropping it's red), two, dynamic content doesn't work (ZeroBlog is blank). It seems to be a problem with Ghostery (maybe?) but it worked fine with Ghostery in Chrome. Web Console messages:
[Error] Sandbox access violation: Blocked a frame at "http://themacmini09.agrius.feralhosting.com:43110" from accessing a frame at "http://themacmini09.agrius.feralhosting.com:43110". The frame requesting access is sandboxed and lacks the "allow-same-origin" flag. global code (ghostery-min.js, line 8)

[Error] Error: Syntax error, unrecognized expression: .messages .message:not(.template error (all.js, line 7) tokenize (all.js, line 7) select (all.js, line 7) gb (all.js, line 7) find (all.js, line 7) init (all.js, line 7) n (all.js, line 7) (anonymous function) (all.js, line 654) j (all.js, line 7) fireWith (all.js, line 7) x (all.js, line 9) (anonymous function) (all.js, line 9)

I have tried whitelisting the site in Ghostery, and even disabling it (which only removed the first error).

Ran for a while - I browsed some site someone made with images. Suddenly crashed after attempting to post a message. Continues to crash when started. Rolled back entire git repo and started from scratch - continued to crash on load of inital home.

screen: http://gyazo.com/f6e26573a4e8ffd72c738f9ac95a9633

I wish I knew more about the python environment or I'd dig in and trace the stack and debug. Beginning to think whatever I posted as a message is what crashed it.

Visiting these causes crash:
http://127.0.0.1:43110/1Gfey7wVXXg1rxk751TBTxLJwhddDNfcdp
http://127.0.0.1:43110/1EU1tbG9oC1A8jz2ouVwGZyQ5asrNsE4Vr
http://127.0.0.1:43110/16tAyA58u8AiZMMu6tcsCYaM8KXWXMhiaJ

This works:
http://127.0.0.1:43110/1Jr5bnqSnnp94CfC7xrqPh4yYYDRkpzozD/

I'll leave it up overnight to continue seeding.

Not every new message reaches all of the peers.

It's related to #43

Already working on new network code without ZeroMQ. It will be released in 1-2 weeks.

Torrent like file splitting and distributing

After git pull :

Traceback (most recent call last):
  File "/usr/lib/python2.7/dist-packages/gevent/pywsgi.py", line 508, in handle_one_response
    self.run_application()
  File "/home/arthur/local/ZeroNet/src/Ui/UiServer.py", line 21, in run_application
    self.ws_handler.run_application()
  File "/home/arthur/local/ZeroNet/src/lib/geventwebsocket/handler.py", line 81, in run_application
    self.run_websocket()
  File "/home/arthur/local/ZeroNet/src/lib/geventwebsocket/handler.py", line 57, in run_websocket
    self.application(self.environ, lambda s, h: [])
  File "/home/arthur/local/ZeroNet/src/Ui/UiServer.py", line 49, in handleRequest
    return self.ui_request.route(path)
  File "/home/arthur/local/ZeroNet/src/Ui/UiRequest.py", line 41, in route
    return self.actionWebsocket()
  File "/home/arthur/local/ZeroNet/src/Ui/UiRequest.py", line 239, in actionWebsocket
    self.log.error("Wrapper key not found: %s" % wraper_key)
NameError: global name 'wraper_key' is not defined

I think the README should recommend installing upnpc via packages (for example apt-get install miniupnc) and then ZeroNet could check if upnpc is in the path instead of having to explicitly use --upnpc option.

Creating an issue about an enhancement discussed in irc that I'm working on.

Currently there are some security issues in regards to downloading sites. A good implementation would be to prompt a user prior to downloading a site in which would output metadata such as the site description and directory size, along with an accept or decline. Among the initial download, there is also no prompt to accept or decline the update of a site being published that grows in size.

Suggestions or enhancements to such an idea are of course welcome.

• ZeroNet doesn't play nice with noscript :(
• Even if I allow 127.0.0.1, it still doesn't load anything properly
• Arch/Firefox Stable/NoScript, HTTPS Everywhere, Clean Links, Disconnect, RequestPolicy, Self-Destructing Cookies

Hi team!
First of all kudos for you project! It's a great tool!

Now I would like to ring a bell about the way we have to use it to browse and found existing site.

I deeply think that if we want decentralized services to be adopted by most of people and not just skilled ones (We work for freedom and liberties for every internet user after all), we should implement some sort of decentralized search/discovery service and human readable URLs.

Search/Discovery service are made to ease the spread and retrieval of informations online during a solo browsing.

Human readable URLs are made to ease sharing and references of a previous browsing session.

Two main features that our browsers commonly implements (Of course it perform such a thing using other system services) and that need to be implement to ease the adoption of your project.

I'm not a professional of bitcoin but I know that bittorrent protocol implement resources discovery and advertising.

• Ability to black/whitelist sites
• Max site size
• Disable autoupdate features
• Restrict site setting changes (pause, delete) to selected users

My site uses cyrillic characters at title and description http://127.0.0.1:43110/18C5HmyTt5Ua4WdyESp3mgmrCwVMPc4yw2 Cyrillic characters tend to Python's UnicodeEncodeError.
After I've published it once, ZeroNet allows me to publish latin version of this fields, but downloads old version of site from some peer immediately.

This would require adding "allow-same-domain" to iframe sandbox. I somewhat understand the security implications, and this would allow other sites access to each others' localresources, but I wanted to open this ticket to track this request. Ultimately, in an effort to provide a rich user experience, this same domain issue will need to be conquered.

Short term, I'd like to see localresources opened up, maybe with an explicit user permission per site, and an explanation of the risks?

Every page loading always reports: UiServer Websocket error, please reload the page

Javascript console shows: http://gyazo.com/f1cee5522fe1516d501b8cea7339deb4

When running in debug mode, accessing the console raises Gevent #445, in UiServer and UiRequest.

Hi, i believe address should begin with Z
https://en.bitcoin.it/wiki/Base58Check_encoding
chart says 32 is the Z character

To make sure we get the latest updates realtime we have make sure we keep 2-3 active connections per site.

This should be added quickly so people won't have to say if they have access to the private key or not.

Similar to torrent's DHT solution, required features:

• Find peer to specified site file (to allow sites to add optionally download files)
• Ipv6 and tor address support for peers
• It should work using on Tor (TCP only, maybe UDP on clearnet)
• No other library dependency if possible
• Store peer ips to allow work without bootstrap server

Maybe its requred to separate the DHT from Peer Exchange

Please comment if you have other ideas/suggestions.

https://registry.hub.docker.com/u/genecyber/zeronet/

Norton

File blocked: tools\upnpc\upnpc-static.exe 

Threat name: Trojan.Gen.2 ( http://www.symantec.com/security_response/writeup.jsp?docid=2011-082216-3542-99 ) 

Pure python upnp implementation would be nice, but i could not find any good.

Here is my idea to make it work:

• When new user visit zeronet proxy he got a welcome message: Hello, welcome to my ZeroNet proxy, here is your authorization key: KzJp5B7mDpZ7kMHv67GowQRys9W9Hbaa5Rzj4PCoiyXfTk1fGAvH. Its never will be displayed again, so please save it now! Ok I saved it, Create user! or Login
  • The hash of the key is got stored in cookie
  • When the user clicks on logout button the cookie is got cleared
  • If he/she chooses Login then he will be asked for previous auth key

Problems:

• No ssl on ZeroNet webui yet
• The signing is done on server-side so the user must trust the proxy operator

Namecoin should be used for publishing sites, which would allow namecoin lookups to the the public key address of a zeronet site. This completely decentralizes the technology that handles name lookups.
I propose zeronet public addresses should go in a value field "zn" or "0net" within a domain name registered with namecoin. More info here: https://wiki.namecoin.info/index.php?title=Domain_Name_Specification#Value_field

If there is a .git directory inside a site folder, crazy things start to happen with ZeroNet, making sites not update properly.

Ignoring any .git directory should fix this.

In the past I had no content by myself, so it was easy to upgrade => replace

Now I have my own blog, my own sites, ... ZeroBlog changed, do I need to upgrade that also ? Do I need to upgrade MY blog? what's with my data?

BTW, one thing needs improvement: As more sites are coming online as more annoying it is WHERE the sites will get the place in the main screen, since it changes with each restart. Maybe we can add numbers to it, or "bookmark" them to a certain spot. What will be if we get hundreds of sites? Will we get a search button/sort button?

 src.File.FileServer UpnpPunch run error: NameError: global name 'SocketError' is not defined in FileServer.py line 50 > UpnpPunch.py line 176 > UpnpPunch.py line 40

replacing SocketError by socket.error gets us to another error :

src.File.FileServer UpnpPunch run error: UnboundLocalError: local variable 'data' referenced before assignment in FileServer.py line 50 > UpnpPunch.py line 176 > UpnpPunch.py line 44

adding return to the except block gets us another error :

src.File.FileServer UpnpPunch run error: TypeError: expected string or buffer in FileServer.py line 50 > UpnpPunch.py line 176 > UpnpPunch.py line 52 > re.py line 177

I gave up there...

Plugin ideas:

• Donate to ZeroProxy owner (visible on every page or on zerohello)
• Maybe the multiuser support (#58 #59), namecoin integration (#31), bitmessage (#65) also can be a plugin

Currently only UDP trackers supported. It could allow store TOR peer address by using peerid field.

Where i can set allow ip list for zeronet webui?

content.json modified timestamp should be in UTC, but looks like its not:
1JF4oezBUFMCbbct1uUnXHf4y3W3grUsXP "modified": 1421296119.0,

http://www.unixtimestamp.com/
The Current Unix Timestamp
1421283266
Thu, 15 Jan 2015 00:54:26 +0000

1421296119
Is equivalent to:
Thu, 15 Jan 2015 04:28:39 +0000

I found a vulnerability in EllipticCurvePoint.SignECDSA. You use random.randint to generate k. random.randint is not a cryptographically secure random number generator.

From Wikipedia:
3. Select a cryptographically secure random integer k from [1, n-1].

Using a weak PRNG for k can result in the compromise of the private key d. For example: "Such failure of RNG caused users of Android Bitcoin Wallet to lose their funds in August 2013."
https://en.wikipedia.org/wiki/Elliptic_Curve_Digital_Signature_Algorithm

• It would be nice to have a per-site Allow clearnet connections setting: If disabled then standard ipv4 peers is connected using Tor exit-nodes. If its enabled then only Tor peers are connected through the Tor proxy.
• Display warning to user: You are in Tor-only mode, you should also configure your browser to use Tor network. (maybe restrict external resources using Content Security Policy)

Problem:

• The torrent trackers does not allow to store tor/i2p addresses.

Possible solutions:

• DHT: Its not easy to write a good DHT implementation and Tor doesnt support UDP, so we need custom DHT (it requires reliable servers for bootstrapping and not sure if its going to work fine with current small number of nodes) or proxies that converts DHT traffic to tcp.
• IRC: Not really decentralized solution, so i would not be happy to use it. (some altcoins using it, eg. Namecoin so it could work)
• Bittorrent http trackers could still work, but without tor address support, so you could browse zeronet sites, but could not host zeronet sites on tor. Maybe its possible with peer-exchange combination.

Any help/idea/suggestion greatly welcomed. Thanks! :)

No idea why, but sometimes the getFile command response is lost in space:

[2015-01-29 17:52:32,391] DEBUG    src.File.FileServer FileRequest: update 1Gfey7wVXXg1rxk751TBTxLJwhddDNfcdp content.json
[2015-01-29 17:52:32,391] DEBUG    src.File.FileServer Someone trying to push a file to own site 1Gfey7wVXXg1rxk751TBTxLJwhddDNfcdp, reload local content.json first
[2015-01-29 17:52:32,507] DEBUG    src.File.FileServer Update for content.json looks valid, saving...
[2015-01-29 17:52:32,512] INFO     src.File.FileServer 1Gfey7..fcdp changed files: [u'messages.json']
[2015-01-29 17:52:32,513] DEBUG    Peer:104.xxx:15441 sendCmd: ping
[2015-01-29 17:52:32,515] DEBUG    Peer:104.xxx:15441 Sent command: ping
[2015-01-29 17:52:32,716] DEBUG    Peer:104.xxx:15441 Got response to: ping
[2015-01-29 17:52:32,717] DEBUG    Peer:104.xxx:15441 Ping: 0.203
[2015-01-29 17:52:32,717] DEBUG    Peer:104.xxx:15441 sendCmd: getFile
[2015-01-29 17:52:32,717] DEBUG    Peer:104.xxx:15441 Sent command: getFile
And its stops here, 20s later got the file from other peers, but not from this one

The socket is alive, because the ping command get the response, but the getFile response never arrives.

From server side everythiing looks allright:

[2015-01-29 17:52:31,207] DEBUG    src.File.FileServer FileRequest: ping None None
[2015-01-29 17:52:31,408] DEBUG    src.File.FileServer FileRequest: getFile 1Gfey7wVXXg1rxk751TBTxLJwhddDNfcdp messages.json
[2015-01-29 17:52:31,409] DEBUG    src.File.FileServer Opening file: data/1Gfey7wVXXg1rxk751TBTxLJwhddDNfcdp/messages.json
[2015-01-29 17:52:31,409] DEBUG    src.File.FileServer Sending file data/1Gfey7wVXXg1rxk751TBTxLJwhddDNfcdp/messages.json from position 0 to 86470
[2015-01-29 17:52:31,409] DEBUG    src.File.FileServer File data/1Gfey7wVXXg1rxk751TBTxLJwhddDNfcdp/messages.json sent

I think its mostly happens after my computer wakes from sleep.
Maybe the zeromq server side in not compeltely thread safe maybe its other problem.

Sooner or later i think we have to get rid of the zeromq (too much limitations), but would be nice to know whats the reason of this problem.

I'm having a slight problem with trying to add a site, I feel like its due to user error on my part, but nevertheless I'd like to get it working

I've done this in a directory /home/dydx/Development/zeronet-sites/dydx which contains index.html and style.css

➜  dydx  python /home/dydx/Development/github/ZeroNet/zeronet.py siteCreate
- Generating new privatekey...
- -----------------------------------------------------------
- Site private key: [removed for sake of posting to github issues] (save it, required to modify the site)
- Site address: 1GXtHRoh7495zJYLthQ8xPt2dRzm8DwWhb
- -----------------------------------------------------------
- Creating directory structure...
- Creating content.json...
Site:1GXtHR..wWhb Site not exits yet, loading default content.json values...
Site:1GXtHR..wWhb Opening site data directory: data/1GXtHRoh7495zJYLthQ8xPt2dRzm8DwWhb...
Site:1GXtHR..wWhb - data/1GXtHRoh7495zJYLthQ8xPt2dRzm8DwWhb/index.html (SHA1: 4065b2b27a289cba476a7ffbce11681cbcc9f6f9)
Site:1GXtHR..wWhb Adding timestamp and sha1sums to new content.json...
Site:1GXtHR..wWhb Verifying private key...
Site:1GXtHR..wWhb Signing modified content.json...
Site:1GXtHR..wWhb Saving to data/1GXtHRoh7495zJYLthQ8xPt2dRzm8DwWhb/content.json...
Site:1GXtHR..wWhb Site signed!
- Site created!

Everything looks good. I try to visit http://127.0.0.1/1GXtHRoh7495zJYLthQ8xPt2dRzm8DwWhb and get a message saying there were no peers found, which I suppose is correct given that no one has visited/seeded the page besides me

I post on the message board about trying to see if it works, and a user replied with this:

> Gives me hash failed, try to use siteSign command again

I suppose thats sort of odd, but I try the siteSign command and get notified that it succeeded:

➜  dydx  python /home/dydx/Development/github/ZeroNet/zeronet.py siteSign 1GXtHRoh7495zJYLthQ8xPt2dRzm8DwWhb                    
- Signing site: 1GXtHRoh7495zJYLthQ8xPt2dRzm8DwWhb...
Private key (input hidden):
Site:1GXtHR..wWhb Opening site data directory: data/1GXtHRoh7495zJYLthQ8xPt2dRzm8DwWhb...
Site:1GXtHR..wWhb - [SKIPPED] data/1GXtHRoh7495zJYLthQ8xPt2dRzm8DwWhb/content.json
Site:1GXtHR..wWhb - data/1GXtHRoh7495zJYLthQ8xPt2dRzm8DwWhb/index.html (SHA1: 4065b2b27a289cba476a7ffbce11681cbcc9f6f9)
Site:1GXtHR..wWhb Adding timestamp and sha1sums to new content.json...
Site:1GXtHR..wWhb Verifying private key...
Site:1GXtHR..wWhb Signing modified content.json...
Site:1GXtHR..wWhb Saving to data/1GXtHRoh7495zJYLthQ8xPt2dRzm8DwWhb/content.json...
Site:1GXtHR..wWhb Site signed!

I go to check the URL once more and see that I have 3 peers! But, it never seems to load. I get a message saying that content.json download failed

Do I just need to wait for more seeds to propagate, or is there an error in my setup somewhere?

Current timeout is 1sec/peer, it not works well with larger sites or slower connections.
We need to do the publish commands in parallel so we can give it longer timeout. (atm there is 5 sec timeout on receive time, so thats the maximum we can get now)

I2P has bittorrent support, and it works very well (but slow). So suggestion is to use it. There should be an option to use only I2P, only clearnet, or both.

More info:
https://geti2p.net/en/
https://geti2p.net/en/docs/applications/bittorrent

When pause pressed empty current WorkerManager's tasks

http://zeronet.readthedocs.org/

• Make ZeroFrame API reference uptodate
• Add Disqus to allow comments/questions
• Put the example site to github
• FAQ
• Description of content.json format
• Create ZeroNet site for the documents
• Network protocol
• New donation page
• Done / Future features
• More screenshots
• Dbschema version 2
• Handshake
• Trusted authorization providers
• New command: actionSiteSign
• New command: actionSiteClone
• New command: actionFileRules
• New command: actionCertAdd
• New command: actionCertSelect
• New command: actionCertSet (admin)

Sites protected by password or public-key based auth.

future enhancement. When publishing becomes popular, more security should be allowed at the site key level in the form of multisig or scriptsig of some type.

Hi team!
First of all kudos for you project! It's a great tool!

Now I would like to ring a bell about the way we have to use it to browse and found existing site.

I deeply think that if we want decentralized services to be adopted by most of people and not just skilled ones (We work for freedom and liberties for every internet user after all), we should implement some sort of decentralized search/discovery service and human readable URLs.

Search/Discovery service are made to ease the spread and retrieval of information