- UPnP handles errors better
- it tries to bind external interface (previously sometimes it searched
on 127.0.0.1 resulting in no routers being detected)
The quick navigation key disrupted keybindings with keyboard modifiers,
like Ctrl-C for copy. This restricts the quick navigation only where no
keyboard modifiers are active.
Fixes#184
- delete key now works when message body is focused as well
- N for next message (down)
- P for previous message (up)
- R for reply
- C for compose
- F for find
- Find is now dynamic if the search text is least 3 characters long
Fixes Bitmessage#655
Addresses #155
Rows are deleted from a message list in multiple places, and this is an
attempt to refactor it so that it is done in one function. It's not used
anywhere yet.
MessageView does not currently load external resources (QTextBrowser by
default interprets all external resources as local file names and tries
to load them like that. This can, in the future, be implemented. For
example, if SOCKS (Tor) is used, the resource could be loaded through
the SOCKS too.
This commit is a skeleton for it that does not actually do anything and
can be filled with an implementation that does the loading.
Email addresses and URIs are now clickable when viewing a message in
plain text mode. Clicking an email address moves to the Send tab, while
clicking an URI has the same result as clicking an URI in html mode, it
will ask for confirmation before opening it in external handler.
HTML parser wasn't correctly handling img tags.
Now it also by defaults disabled external schemas to prevent
deanonymisation (even though the renderer actually doesn't support
external schemas at the moment)
Addresses #178
Attackers injected node addresses with port 0 into the network. Port 0
is unusable on many OSes and can't be listened on. PyBitmessage won't
accept nodes that have port 0 anymore.
- postpone initial sleep until the first getdata is received
- also sleep when received a getdata request for an object that hasn't
been advertised to the other node yet
There was a report that by quickly asking a large number of nodes if
they have an ACK object (which the attacker knows but it is injected
into the network by the recipient of the message), it can estimate how
an object propagates through the network, and eventually pinpoint an
originating IP address of the injection, i.e. the IP address of the
message recipient.
This patch mitigates against it by stalling when asked for a nonexisting
object (so that the attacker can't spam requests), and also upon
connection before sending its own inventory list (so that reconnecting
won't help the attacker). It estimates how long a short message takes to
propagate through the network based on how many nodes are in a stream
and bases the stalling time on that. Currently that is about 15 seconds.
Initial connection delay takes into account the time that already passed
since the connection was established.
This basically gives the attacker one shot per a combination of his own
nodes and the nodes he can connect to, and thus makes the attack much
more difficult to succeed.
Editing of blacklist labels affected the rerendering, because it emits
the changed signal too, and it caused an exception because the address
field was missing at that time. This works around both.
In some situations, it's not necessary to send an ACK. For example, when
the sender is blacklisted, when the message has no content, or when the
address has ACK sending disabled.
Also it's not necessary to rebroadcast empty messages into a mailing
list.
When advertising nodes and when establishing connections, private IP
range checks were not done. This could cause private IPs to be
advertised across the network. Also, some of the checks weren't
IPv6-aware.
Fixes Bitmessage#768
If you change, add or remove an item in a treewidget or addressbook,
messagelists will now autoupdate labels, and sender comboboxes will also
update if applicable.
Fixes#69
Message compose sometimes misinterpreted quoted message as HTML, causing
rendering screwups. Since we don't support HTML composing, we will treat
all quoted messages as plain text.
Zooming in message body view / compose works in single steps
irrespective of wheel sensitivity, and info about zoom level is
displayed in percent rather than font pixel size.
if singleWorker crashed, the thread couldn't be joined. This both makes
it so that it doesn't crash, as well as reorders the shutdown sequence
so that it is less likely to be triggered.
Fixes Bitmessage#549
- it didn't shutdown correctly
- it didn't handle exception correctly (however, if I understand
correctly, this will never be triggered if using blocking get, so it
doesn't affect PyBitmessage)
- flushing size check changed from 1 to 0 (I don't know why it was 1)
Flood mitigation was done both in the ObjectProcessorQueue as well as
receiveData threads. This patch removes the mitigation in receiveData
threads and cleans up the one in the ObjectProcessorQueue
- subject on replies was not parsed correctly (or should I say was
parsed more than necessary)
- unicode can handle invalid characters without needing a special
function
Fixes#164
The previous commit broke places where UserData was read from the
mesagelist line, because it's now a string rather than a QVariant. This
actually simplifies things.
On OSX and linux, the tray icon now raises window.
Starting a new instance should raise the window too, but it may not work
on linux (on my Ubuntu it just flashes the icon)
Fixes Bitmessage#743
Items like dialog box buttons were not localised. This loads the default
qt localisation, and also unloads the old localisation on locale change.
Still needs to be adjusted for frozen.
Addresses Bitmessage#737
Message will render as user is scrolling down. This prevents interface
freezes on long messages (such as inline img in text mode). Fixes
Bitmessage##366
Also a minor fix in text mode rendering.
If attempting to send to an email address from an unregistered
account, auto-send registration to mailchuck.
Use label if possible, otherwise random 12 characater address
Fixes#131
- addressGenerator got stuck (introduced recently when cleaning up
shutdown
- do not put addresses into API return queue if API is inactive
- improve translation
Settings GUI now contains a checkbox for UPnP and auto starts/stops the
thread when changed.
Default UPnP socket timeout decreased for faster thread
starting/stopping
Addresses #114
- delete from sent works when pressing delete key
- delete from trash deletes rather than moves
- delete with shift deletes rather than moves
Fixes#118
- changed almost all "print" into logger
- threads have nicer names
- logger can have configuration in "logger.dat" in the same directory as
"keys.dat", and the logger will pick the one named "default" to replace
the "console" and "file" that are in PyBitmessage otherwise
Example file for logging to syslog:
[loggers]
keys = root,syslog
[logger_root]
level=NOTSET
handlers=syslog
[logger_syslog]
level=DEBUG
handlers=syslog
qualname=default
[handlers]
keys = syslog
[handler_syslog]
class = handlers.SysLogHandler
formatter = syslog
level = DEBUG
args=(('localhost', handlers.SYSLOG_UDP_PORT),
handlers.SysLogHandler.LOG_LOCAL7)
[formatters]
keys = syslog
[formatter_syslog]
format=%(asctime)s %(threadName)s %(filename)s@%(lineno)d %(message)s
datefmt=%b %d %H:%M:%S
Python < 2.7.9 does not support anonymous SSL server through
ssl.wrap_socket, so we have to disable it. Works fine as client.
Try to prefer secp256k1 curve (again, requires python >= 2.7.9)
- include bitmsghash so that it bundles the C library PoW
- include sslkeys so that SSL works in the bundle
- PyBitmessage version now propagated from the shell build script
These are dummy key and certificates. They are loaded but otherwise not
used. This is a workaround against python's ssl_wrap deficiency, it
won't allow to execute a server-side SSL handshake without a loaded key
+ cert. Since generating key/cert locally would require additional
libraries, I opted for including a dummy one in the source.
