Sort .zip order, and other determinism changes to make sure builds
are repeatable for windows deps.

65615a3a Gitian fixes for 0.9.0rc1 build (Wladimir J. van der Laan)

- Add 'g++' package (virtualbox images don't have this by default)
- Workaround for determinism in Qt5 resources
- Pass --disable-maintainer-mode --disable-dependency-tracking to
  configure for libqrencode to avoid random errors about missing m4
  directory
- Fix typo -with-pic -> --with-pic

It is not necessary to rebuild dependencies after this commit.
Fixes #3610 and #3612.

69baec96 doc: Add historical release notes (Wladimir J. van der Laan)

c117d9e9 Support for error messages and a few more rejection reasons (Luke Dashjr)
14e7ffcc Use standard BIP 22 rejection reasons where applicable (Luke Dashjr)

Add all known release notes for versions since 0.3.12,
thanks to Matt Corallo for collecting them on the wiki.

Don't define BUILD_DATE at all when no git version information
is available. `version.cpp` will then define it for us correctly
to the last commit date.

This has been fixed and broken many times over 0.9 history
(21cc8bdc, ef1e984e), please don't touch this code unless you plan
on testing all possible scenarios including gitian builds.

Fixes #3570.

c32a486f Add more data-driven tests. (Matt Corallo)

Just for completeness. Only a few messages changed or moved since last time.

Don't change any translatable messages until 0.9 final unless really necessary.

df966d1b log parameter interactions to debug.log (Philip Kaufmann)

Prepare block connection logic for headers-first

Commit a22eed6a got reverted in a RPC documentation update, redo it.

`bitcoin-qt --help` was missing a final newline.

.rc's should be linked into the executable directly,
not through a helper library.

4412c5a7 Fix `getaddednodeinfo` RPC call with dns=false (Wladimir J. van der Laan)

The getaddednodeinfo RPC call, when invoked with the dns flag set to
false, returns a malformed JSON object with duplicate keys.

Change this to return an array of objects with one key as
shown in the help message.

Fixes #3581.

69d03bc6 qt: Fix tab order in send and receive tab (Wladimir J. van der Laan)
79fb0557 [Qt] minor receive tab improvements (Cozz Lovan)

3cf1f436 Mention `*` value for account in documentation for `getbalance` RPC (Wladimir J. van der Laan)

- remove btc address length from address validator
- add an optional btc address check in validated line edit that defaults
  to off and is used in GUIUtil::setupAddressWidget()
- an isAcceptable() check is added to validated line edit on focus out
  which only kicks in, when a validator is used with that widget
- remove an isAcceptable() check from sendcoinsentry.cpp
- remove obsolete attributes from ui files, which are set by calling
  GUIUtil::setupAddressWidget()
- move some more things to GUIUtil::setupAddressWidget() and remove them
  from normal code e.g. placeholder text

4c0b2cde Win32: use a more modern API call in FileCommit() (Philip Kaufmann)

As reported by @Subo1978, error messages don't have a trailing newline
anymore after commit b77dfdc9. Add back this newline.

e5aadb25 Updated fallback seed nodes from sipa's DNS seeds (Gavin Andresen)

c7f93325 Add check for valid keys in `importprivkey` (Wladimir J. van der Laan)

edd735da [Qt] Fix coin control labels update (Cozz Lovan)

d7aa1ec8 [Qt] fix bitcoin: URI strings (spelling fixes) (Philip Kaufmann)

The base58 armoring was checked, but not the resulting private key,
which could be out of range. Fix this by adding a check.

6986861f build: use -mwindows for gui targets when linking with mingw (Cory Fields)

use english for all reason strings in IsStandardTx()

This changes the block processing logic from "try to atomically switch
to a new block" to a continuous "(dis)connect a block, aiming for the
assumed best chain".

This means the smallest atomic operations on the chainstate become
individual block connections or disconnections, instead of entire
reorganizations. It may mean that we try to reorganize to one block,
fail, and rereorganize again to the old block. This is slower, but
doesn't require unbounded RAM.

It also means that a ConnectBlock which fails may be no longer called
from the ProcessBlock which knows which node sent it. To deal with that,
a mapBlockSource is kept, and invalid blocks cause asynchronous "reject"
messages and banning (if necessary).