Without this, old senders will throw a messy 404 traceback when talking
to a modern server.
Unfortunately 0.4.0 receivers don't make API calls in the right order,
so they throw a 404 before seeing our "you need to upgrade" message.
I was really confused about the Server-Sent Events syntax. This new one
is compatible with actual web browsers and the spec:
http://www.w3.org/TR/eventsource/
This requires a DB delete/recreate when upgrading. It changes the server
protocol, and app IDs, so clients cannot interoperate with each other
across this change, nor with the server. Flag day for everyone!
Now apps do not share channel IDs, so a lot of usage of app1 will not
cause the wormhole codes for app2 to get longer.
There are now three ways to invoke send:
* "wormhole send": ask for a text message
* "wormhole send FILENAME": send a file
* "wormhole send --text TEXT": send text message
This removes "side" and "msgnum" from the URLs, and puts them in a JSON
request body instead. The server now maintains a simple set of messages
for each channel-id, and isn't responsible for removing duplicates.
The client now fetches all messages, and just ignores everything it sent
itself. This removes the "reflection attack".
Deallocate now returns JSON, for consistency. DB and API use "phase" and
"body" instead of msgnum/message.
This changes the DB schema, so delete the DB before upgrading the server.
The "bytes % bytes" syntax only appeared on py3.5, so don't use it.
Updated travis to expect py3.4 works.
The twisted side is probably even more broken for py3.4 than it is for
py3.5.
This roughly parallels the way that blocking/eventsource.py and the pypi
"requests" modules work: the server can set the encoding (with
"Content-Type: text/event-stream; charset=utf-8"), and the EventSource
parser will decode accordingly. However eventsource_twisted.py *always*
returns unicode (on both py2/py3), even when the server hasn't set an
encoding. blocking/eventsource.py returns bytes (on py3, and str on py2)
when the server doesn't set an encoding.
In the future, eventsource_twisted.py should return bytes when the
server doesn't set an encoding.
eventsource_twisted.py includes an alternate approach that might be
necessary (a to_unicode() function instead of always using .decode), but
I won't be sure until enough of Twisted has been ported to allow the
EventSourceParser to be tested.
Also fix demo.py for python3.
* declare transit records and handshake keys are bytes, not str
* declare transit connection hints to be str
* use six.moves.socketserver, six.moves.input for Verifier query
* argparse "--version" writes to stderr on py2, stdout on py3
* avoid xrange(), use subprocess.Popen(universal_newlines=True)
* use modern/portable "next(iter)" instead of "iter.next()"
* use six.moves.input() instead of raw_input()
* tell requests' Response.iter_lines that we want str, not bytes
The main wormhole code is str (unicode in py3, bytes in py2). Most
everything else must be passed as bytes in both py2/py3.
Keep the internal "side" string as a str, to make it easier to merge
with other URL pieces.
The twisted.python.logfile in Twisted-15.4.0 is not yet compatible with
py3, but can be bypassed by not daemonizing the server (so it doesn't
write to a logfile). This has been fixed in twisted trunk, so when
15.4.1 or 15.5.0 comes out, this will no longer be needed. But I think
we'll leave it in place, since sometimes it's handy to run a server
without daemonization.
Just make up a code like NUMBER-STUFF, and add --code= to the
send-text/send-file command. Also don't use tab-completion on the
codewords part of the receiving side, unless you stuck to the even/odd
PGP wordlist. (tab still works for the channel-id).
Now the server allocates a channel randomly from set of available ids
with the shortest possible length. So concurrency=1 will always yield a
channel-id between 1 and 9 (inclusive). If we have 9 simultaneous
sessions, we'll start allocating channels from 10 to 99. 100
simultaneous connections kicks us into the 100-999 bucket, etc.
This is a proxy for the other client's version, and encourages both
sides to upgrade to the current version each time the server is
upgraded (which will be once per release).
To be useful, both sides must add -v. If the sender uses -v but the
receiver doesn't, the receiver won't show the verification string, so
the sender can't compare it to anything (and must either abort the
transfer or accept it blindly). Maybe the receiver should show the
verification string unconditionally. Maybe the sender should
indicate (in unprotected plaintext, along with the PAKE message) whether
the receiver should show it or not.
We used to use twisted.python.usage.Options, hence we depended upon
Twisted. Now we depend upon "argparse" instead, which is in the py2.7
stdlib (and on pypi for 2.6). This package will still (eventually)
provide Twisted support, but applications which need it will already
express a dependency on twisted themselves, so by removing the
dependency here, we make life easier for applications that don't use it.
Applications should feel free to pass wormhole.const.RENDEZVOUS_RELAY
here, but I figure it should be clear that you're using a public service
that's hosted *somewhere* external.
This fixes the situation where you start the receiver first, then start
the sender, then you hit TAB on the receiver.
This somewhat improves the situation where you start the receiver first,
hit TAB (getting nothing), then start the sender, then hit TAB on the
receiver again. The second TAB will list the channel-ids, but won't
insert the only one as it's supposed to. You must type something (which
you can erase) and then hit TAB again to get a unique channel-id
inserted. But at least you can tell which one to type.
The first TAB runs the completer with readline.get_completion_type()
equal to 9=TAB=try-to-insert. The second (and subsequent) TABs use
63=?=list-matches, and it won't go back to 9 until you type something.
If all the direct hints resulted in timeouts (e.g. they were to bad IP
addresses where connections just hang), the relay connection would fail.
The establish_connection() function had the same TIMEOUT as the
direct-hint connector, so it would give up just before the relay
connection was initiated.
