same content, just moved to the subfolder.

Update by moving 0.14.3 to release-notes folder, deconflict with D337

Changes look good to me - no objections to making it version-independent at least in the beginning.

This file is referenced in several build / packaging files.
We cannot simply remove it without making sure whatever depends on it still works without it.

I tested both wallet and non-wallet build according to these instructions on an 11.0-RELEASE-p10 , with gmake check passing .

Manually tested, looks ok to me.

I confirm this passed extended regression tests on my machine.

I noticed on BU that clang-format from 5.x formats differently .

Ran full test suite

This also passed tests on my end.

In D312#5449, @deadalnix wrote:

There is a test that sign tx and check them, and other that use a specific set of flags. I think this is not ideal, but probably good enough for now.

Superceded by D326

Fix trace

Please have a look if you think the log message is ok.

Add a log message when nodestate is null, as requested

Remove the asserts as requested by deadalnix

Planning to do this cleanup AFTER D324 now ...

Add changes from review feedback

No disagreement there (about the asserts not doing much for reliability of the software).

Gladly moving to nullptr, I just didn't want to mix refactoring with semantic change here.

I think there will be an overlap with D310 (Sign using replay protection by default in raw transaction RPC).

Looks good to me.

My understanding is that REQ-DISABLE is not implemented at this stage.

I asked for others to review (repeating this request on Slack on Jul 10).
It seems no-one is available.

Handing back to review for an answer to the question I asked.

Remove the forward class declaration which is no longer needed.

To remove the forward class declaration in util.h which is no longer needed.

@deadalnix : please confirm whether you really want the consensus.h code changed to turn DEFAULT_MAX_BLOCK_SIZE into a simple number instead of 8 * ONE_MEGABYTE .

Add config to CConnman via constructor

There is no test affected by this change in default mode of operation?
That would be a lack of a test case somewhere.

Looks ok, tested as well.

Looks ok, builds and tests ok too.

Remove cfg_ptr local variable

Ubuntu Yakkety (16.10) and Debian 9 come with g++ versions that are able to build with this C++14 requirement.
To properly support C++14, we need to upgrade the gitian build which currently uses Ubuntu 16.04 (Trusty).

Plan to split up into smaller separate diffs.

Didn't see any semantic change, but due to amount of changes it would be good if another person went over it carefully as well to see if I missed any.
The scope removals and transformations of if..else's look ok to me - a good improvement.

Updates from review, before splitting D282 up into smaller pieces as requested

With @sickpig 's assistance, we are as far as the Linux wallet job passing.
https://travis-ci.org/ftrader-bitcoinabc/bitcoin-abc/builds/250686535

OK, I will look at splitting this up into digestable separate diffs.

The whole get_canonical_bitcoin_binary is an exercise in divination. This is improper for tests.

Returning to review, maybe someone else can answer my question?

Travis tests OK (ignoring the well known Wine issue)

It doesn't seem to be getting any better by just asking for g++5 .

Well, following Travis' own advice about how to set up the sources and package name for gcc seems to get you this:

According to Wikipedia:
GCC finished support for C++14 in GCC 5, and made C++14 the default C++ standard in GCC 6.

This one needs to come with a travis.yml upgrade.