Hmm... also since the new txindex code was landed, maybe defer this to next release? If someone has to downgrade because their project relies on removed RPC for some bizarre reason, they'll be pissed if they have to reindex the txindex.

I think a release note is well-deserved here. Yeah, I know it's been deprecated for 3 years already but I didn't even realize that until I read the help text just now. =D

We should also backport this: https://github.com/bitcoin/bitcoin/pull/12198

rebased; de-charred

no release note?

Looks good; some segwit tests removed of course but otherwise verbatim.

note: Logv already has virtual removed

rebase

rebase

rebase

rebase

rebase

last rebase & set the version number

added diff-range for easy comparison: https://github.com/bitcoin/bitcoin/compare/8181db88f~1..9b2704777

note to self: remove unsigned chars!

(Obviously if we ever introduce new tx format with scriptsig data redacted from txid calc, then we may have txid != hash. But that's a long long way off)

note to self: needs to be updated appropriately if D3115 is wanted.

ah changed my mind, it's fine :)

amend for nit

LGTM, except just recomment that line in script_tests.cpp .

use .format

Cool! TIL memset 0 is not necessarily good enough.

(is backport of PR12193)

This is vague and really not needed any time soon.

some commits extracted in D1915 D2263 D2265 D2266 D2267 D2268 D2269

note comment in D2928 was updated incorrectly due to out-of-order with this

(I'm happy to have this deferred to 0.19.8, if there is a desire for another 'good and stable' release after 0.19.6)

Testing done :-) ... this is robust!
(& rebased)

description nit:

In D3073#73361, @IntegralTeam wrote:

rebase onto master, and edit the description

rebase ; address comments

rebase; release notes

@IntegralTeam OK, now my diff is landed -- please rebase onto master, and edit the description to say that this is a backport like so:

use pushKV

oops, obviously should use pushKV here!

Updated test plan; ran a couple times with --extended. Only observed (once) an apparently unrelated failure in mining_getblocktemplate_longpoll.py (some kind of threading error), that I could not even reproduce in individual testing.

I'm also thinking that maybe all the script tests that include the CHECKDATASIG script flag should stop referring to it, except for tests of sigops counting. However if other implementations are relying on those flags being in script_tests.json then maybe they should be kept.

I am inclined to rename the flag SCRIPT_ENABLE_CHECKDATASIG to something more accurate (since it doesn't enable the opcode, it just enables counting its sigops) but that's debatable...

Mainnet indexing in background results: 8.5 hours

2019-05-23T03:08:04Z Using obfuscation key for /home/bitcoin/.bitcoin/indexes/tx
index: 0000000000000000
2019-05-23T03:08:04Z txindex thread start
...
2019-05-23T03:24:36Z Syncing txindex with block chain from height 200889
...
2019-05-23T04:28:09Z Syncing txindex with block chain from height 300655
...
2019-05-23T07:13:36Z Syncing txindex with block chain from height 400084
...
2019-05-23T10:53:08Z Syncing txindex with block chain from height 501172
...
2019-05-23T11:42:01Z Syncing txindex with block chain from height 582571
2019-05-23T11:42:32Z Syncing txindex with block chain from height 583169
2019-05-23T11:43:03Z Syncing txindex with block chain from height 583374
2019-05-23T11:43:26Z txindex is enabled at height 583818
2019-05-23T11:43:26Z txindex thread exit

In D3094#73113, @Fabien wrote:

test_runner.py does not run all the tests, there is a cutoff based on their expected duration. To run them all you need test_runner.py --extended.
For example, abc-p2p-compactblocks.py uses make_conform_to_ctor, but will not be tested without the --extended flag.
Did you also check these tests ? Anyway you should update the test plan.

I'm seeing that Core did a big refactor of GetBalance in https://github.com/bitcoin/bitcoin/pull/15747 .

Mainnet migration test (on a VPS with OK specs -- about 57 minutes)

...
2019-05-23T02:08:01Z Upgrading txindex database... [0%]
2019-05-23T02:13:57Z Upgrading txindex database... [10%]
2019-05-23T02:19:52Z Upgrading txindex database... [20%]
2019-05-23T02:25:57Z Upgrading txindex database... [30%]
2019-05-23T02:31:26Z Upgrading txindex database... [40%]
2019-05-23T02:37:09Z Upgrading txindex database... [50%]
2019-05-23T02:42:31Z