Age | Commit message (Collapse) | Author |
|
Move CMessageHeader from net.h to protocol.[ch]pp, with the
implementation in the .cpp compilation unit (compiling once is enough).
This commit does *not* and should not modify *any* code, it only moves
it from net.h and splits it across protocol.cpp and protocol.hpp.
Indentation changes aside the closest thing to a modification of code is
the addition of the 'TODO' comment (the execution of which requires code
modifications and thus doesn't belong in this commit).
Signed-off-by: Giel van Schijndel <me@mortis.eu>
|
|
util.h doesn't use REF, serialize.h does, creating a dependency of
serialize.h on util.h, but util.h already depends on serialize.h. To
resolve this circular dependency the function 'REF' has now been moved
closer to one of its two points of use.
Signed-off-by: Giel van Schijndel <me@mortis.eu>
|
|
Signed-off-by: Giel van Schijndel <me@mortis.eu>
|
|
Explicitly make these global variables less-global to reduce the maximum
scope of this global state.
In my experience global variables tend to be a major source of bugs. As
such the less accessible they are the less likely they are to be the
source of a bug.
Signed-off-by: Giel van Schijndel <me@mortis.eu>
|
|
cause deadlocks
|
|
|
|
Spotted by Venkatesh Srinivas <me@endeavour.zapto.org>
|
|
|
|
Unify copyright notices.
|
|
Use 'unsigned char' rather than 'char' for pchMessageStart.
|
|
Qualify make_tuple with boost:: namespace.
|
|
Regarding https://bitcointalk.org/index.php?topic=28022.0
main.cpp has: "char pchMessageStart[4] = { 0xf9, 0xbe, 0xb4, 0xd9 };"
Per discussion on the thread linked, leaving the signedness of
pchMessageStart is unsafe for values > 0x80. This patch specifies
'unsigned char' in main.cpp and net.h.
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
|
|
db.cpp has a number of uses of make_tuple and has 'using namespace
std' and 'using namespace boost'. Without qualifying make_tuple,
std::make_tuple is preferred, which is incorrect. This patch qualifies
make_tuple.
Signed-off-by: Jeff Garzik <jgarzik@exmulti.com>
|
|
Signed-off-by: Jeff Garzik <jgarzik@exmulti.com>
|
|
|
|
To a variation on:
// Copyright (c) 2009-2010 Satoshi Nakamoto
// Copyright (c) 2011 The Bitcoin developers
|
|
|
|
|
|
|
|
This is needed for the script unit tests.
|
|
There were some problems with the existing testing setup:
- Makefile rules for test-file compilation used CFLAGS instead of
CXXFLAGS in makefile.unix
|
|
|
|
|
|
|
|
|
|
Check for duplicate txins in CheckTransaction.
|
|
|
|
avoid strAddress + validity checks
|
|
Identify as "Bitcoin + version number" when mapping UPnP port
|
|
Don't listen if on TOR (resolves #441).
|
|
|
|
|
|
Avoid references to addresses using strings, and use CBitcoinAddress
as much as possible. Also added some validity checks on addresses
entered using RPC.
|
|
Translation from "Open Bitcoin" to "Verstuur Bitcoins"
|
|
Makes Bitcoin identify itself as "Bitcoin + version number" instead of
the nondescript "libminiupnpc" when forwarding a port via UPnP.
|
|
Comment "deprecated"
|
|
|
|
|
|
|
|
Fix EncryptKeys crash introduced by a9ba471, identified by TD.
|
|
|
|
Fixes issue #424
|
|
Bugfix: don't overuse limited ExtractAddress
|
|
ExtractAddress was called with the keystore as argument in RPC and
UI, limiting results to own keys. This caused empty "address"
fields.
|
|
make SetHash160 return a value (as specified in the function signature)
|
|
|
|
|
|
keys indexed by address + introduced CBitcoinaddress
|
|
Fix for small change outputs
|
|
With the separation of CENT and MIN_TX_FEE, it is now reasonable
to create change outputs between 0.01 and 0.0005, as these are
spendable according to the policy, even though they require a fee
to be paid.
Also, when enough fee was already present, everything can go into
a change output, without further increasing the fee.
|