NOMERGE: Support signet as parent chain #414

jtimon · 2018-09-21T05:50:12Z

signet has simpler signed blocks but no CT or any other extra feature from elements.

Suggestion to test this: #433

Dependencies:

QA: feature_fedpeg.py: Remove re-implementation of sync_all QA: feature_fedpeg.py: Remove re-implementation of sync_all #421

jtimon · 2018-09-22T03:21:21Z

fwiw, this seems to be working in the explorer, see jtimon/elements-explorer@df65994
But it would still be good to reuse the feature_peg.py modifying it a little bit to test this, even if it's only locally like when the parent chain is regtest before merging this.

jtimon · 2018-09-22T03:28:47Z

src/pow.cpp

 {
+    if (g_solution_blocks) {
+        if (hash == params.hashGenesisBlock) return true;


I guess we don't need or want this here.

You would need to provide the signature for the genesis block otherwise, no?

No, this is used to validate the parent chain. The sidechain doesn't validate the parent chain's genesis block (in fact, not even the parent chain should validate the genesis block, the genesis block is a consensus rule in itself and thus doesn't need to be validated, not when it's pow nor when it's blocksigned).

jtimon · 2018-09-27T00:35:33Z

This can now be tested with:

python3 ./qa/pull-tester/rpc-tests.py feature_fedpeg --parent_type=signet --parent_binpath=/home/user/code/bitcoin/src/bitcoind

Assuming you're building https://github.com/kallewoof/bitcoin/commits/signet ( b4c9e78 ) on /home/user/code/bitcoin/.

The changes to feature_fedpeg shouldn't break other ways to call the tests:

python3 ./qa/pull-tester/rpc-tests.py feature_fedpeg
python3 ./qa/pull-tester/rpc-tests.py feature_fedpeg --parent_type=bitcoin --parent_binpath=/home/jt/data/programs/binaries/bitcoin-0.16.3/bin/bitcoind
python3 ./qa/pull-tester/rpc-tests.py feature_fedpeg --parent_type=bitcoin --parent_binpath=/home/jt/data/programs/binaries/bitcoin-0.17.0/bin/bitcoind

stevenroose · 2018-09-27T09:13:16Z

qa/rpc-tests/feature_fedpeg.py

            self.binary = self.options.parent_binpath if self.options.parent_binpath != "" else None
            self.nodes.append(start_node(n, self.options.tmpdir, self.extra_args[n], binary=self.binary, chain=self.parent_chain))
+            if self.options.parent_type == 'signet':
+                self.nodes[n].importprivkey('8Hh8jNjkx1aSCgEk3iq9Vo2APZUSDQJVt3rJ2BRpb5Tavqb68vW')


Is this magic value related to the one in -signet_blockscript? Perhaps define them as variables somewhere earlier then together so that it's clear they are related.

This will eventually go away, I assume, unless we agree on a global default signet (seems unlikely, at least for upstream to bitcoin/bitcoin).

yeah, we could generate the block signing script and the wif dynamically by starting a node, getnewaddress, dumpprivkey, etc beforehand and then starting the actual testing nodes with the resulting script.
This is just the wif for the public key corresponding to the script.

This is what I was using for generating the p2pk sript and the corresponding wif to import:

jtimon@3a597eb

stevenroose · 2018-09-27T10:31:51Z

qa/rpc-tests/feature_fedpeg.py

@@ -140,13 +153,21 @@ def test_pegout(self, parent_chain_addr, sidechain):
                break
        assert pegout_tested

+    # TODO Remove when signet supports generate


Is there an upstream issue to track for this?

No, all we have for now is bitcoin/bitcoin@master...kallewoof:signet where I'm commenting directly on the code.
Pinging @kallewoof since I can't remember if I suggested this to him.

I intend to make a PR against bitcoin core. For now there is a PR against my repo here: kallewoof/bitcoin#4

stevenroose · 2018-09-27T12:22:04Z

src/primitives/bitcoin/block.h

+ * Contains a mapping of hash to signature data for each block header
+ * in signet networks.
+ */
+extern std::map<uint256,std::vector<uint8_t>> g_blockheader_payload_map;


Is this supposed to be like a global cache? It doesn't seem to get emptied ever.

An in-memory cache that, most importantly, doesn't inflate memory usage when signet feature is off.

This does mean an ever-growing set when on, yes. Wonder what cane be done with this.

@kallewoof thoughts?

It's ever growing, yep. It could be cached since header sigs are stored along with headers. I haven't thought too much about it, as even a 1 million block chain (that's 2x the lifetime of bitcoin right now) would only have ~100 MB of RAM use.

Also when it's a 11-of-15 multisig? Or even larger if this script doesn't have the same script limits as normal tx scripts.

Right you would basically store siglen + alpha per block. For a 15 multisig over ECDSA it'd get big quicker, I suspect, but I assume we will eventually switch to Schnorr multisigs long term.

jtimon · 2018-10-06T02:40:59Z

Now we can review signet more easily with kallewoof/bitcoin#4

jtimon · 2018-10-06T08:41:53Z

Needed rebase. Also, removed one TODO and its workaround in the tests since now generate works directly with signet.

jtimon · 2018-10-06T14:55:11Z

Moved to kallewoof/bitcoin#5 and removed the special case for the p2sh test for signet.

signed has simpler signed blocks but no CT or any other extra feature

jtimon · 2018-10-09T08:50:33Z

Updated, now it doesn't have the wif and blockscript hardcoded but generated aumotatically

jtimon · 2018-10-09T15:14:36Z

Fixed travis.

stevenroose · 2018-10-10T08:32:38Z

tACK 23cb9ae with parent 2aca2a9 (kallewoof/signet-0.17)

instagibbs · 2018-10-12T15:12:54Z

Could we get travis testing against a signet parent binary?

stevenroose · 2018-10-16T10:41:09Z

That would be ideal of course. Could be as simple as the Bitcoin extra Travis thing. But then there need to be some kind of "stable" signet release..

instagibbs · 2018-10-16T13:12:57Z

@stevenroose I'm simply reticent to merge something we can't be confident of stability of the protocol itself.

@kallewoof any plans on "finalizing" this, whatever that means to you? :) We clearly can adapt implementation later to later upgrades.

stevenroose · 2018-10-16T14:43:46Z

Maybe when the elements-0.17 work is done, signet is "released" and we can download a stable binary. I wouldn't mind to keep this open. Or either way just have a custom-compiled signet binary in Travis.

jtimon · 2018-10-16T23:00:43Z

Added a commit (not squashed yet) supposedly adapting the test to #433 as parent chain binary.

~~But even though everything else should have stayed the same, travis seems to be failing.~~
~~...it's not the first time assetdir.py randomly fails for me in elements-0.14, let me ask travis to try again...~~

instagibbs · 2019-03-28T21:08:05Z

@kallewoof have you or do you have intentions of releasing binaries for signet? We could at least get this ported to master for support/iteration.

kallewoof · 2019-03-28T23:22:53Z

@instagibbs Sorry for lack of activity. I am intending on proposing a BIP for signet in the near future. Things will be switched around a bit, I suspect, so no binaries yet.

Edit: to actually answer your question, I think having binaries for signet would be useful and will see about making those available. No timeframe at this point in time.

instagibbs · 2019-03-28T23:38:59Z

Ok just curious. Still interested at least in supporting as parent chain.

jtimon changed the title ~~WIP: Support signet as parent chain~~ WIP Support signet as parent chain Sep 21, 2018

jtimon force-pushed the e14-parentchain-signet branch 2 times, most recently from 91059bf to 4380518 Compare September 21, 2018 07:21

jtimon changed the title ~~WIP Support signet as parent chain~~ TOTEST Support signet as parent chain Sep 22, 2018

jtimon changed the title ~~TOTEST Support signet as parent chain~~ TOTEST: Support signet as parent chain Sep 22, 2018

jtimon commented Sep 22, 2018

View reviewed changes

jtimon force-pushed the e14-parentchain-signet branch 3 times, most recently from 0fd6752 to 1e8648a Compare September 27, 2018 00:23

jtimon changed the title ~~TOTEST: Support signet as parent chain~~ Support signet as parent chain Sep 27, 2018

stevenroose reviewed Sep 27, 2018

View reviewed changes

jtimon force-pushed the e14-parentchain-signet branch from 1e8648a to 9dcf7dd Compare October 6, 2018 08:39

jtimon force-pushed the e14-parentchain-signet branch from bbc58b9 to 4d9ad55 Compare October 6, 2018 14:55

jtimon force-pushed the e14-parentchain-signet branch from 4d9ad55 to f95af12 Compare October 7, 2018 06:07

jtimon added 3 commits October 9, 2018 09:55

Backport signet block serialization

3123072

Reuse block signing code for signet and elements

1f5e0a5

Support signet as parent chain

c4ddcee

signed has simpler signed blocks but no CT or any other extra feature

jtimon force-pushed the e14-parentchain-signet branch from f95af12 to 1518f6e Compare October 9, 2018 08:49

jtimon added 2 commits October 9, 2018 17:12

QA: feature_fedpeg.py: Move from --parent_bitcoin to --parent_type

b3a28f2

QA: Test support for signet as parent chain

23cb9ae

jtimon force-pushed the e14-parentchain-signet branch from 1518f6e to 23cb9ae Compare October 9, 2018 15:14

jtimon changed the title ~~Support signet as parent chain~~ NOMERGE: Support signet as parent chain Oct 16, 2018

SQUASH: Adapt signet fedpeg test to ElementsProject#433

cc7d24f

jtimon force-pushed the e14-parentchain-signet branch from 6ba0d07 to cc7d24f Compare October 16, 2018 22:18

jtimon added 0.14.1 and removed 0.14.1 labels Oct 16, 2018

jtimon mentioned this pull request Nov 22, 2018

NOMERGE: [0.17] Signet implementation #433

Closed

3 tasks

instagibbs added the needs port Needs backport to a different branch label Mar 25, 2019

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

NOMERGE: Support signet as parent chain #414

NOMERGE: Support signet as parent chain #414

jtimon commented Sep 21, 2018 •

edited

Loading

jtimon commented Sep 22, 2018

jtimon Sep 22, 2018

kallewoof Oct 6, 2018

jtimon Oct 6, 2018

jtimon commented Sep 27, 2018

stevenroose Sep 27, 2018

kallewoof Sep 28, 2018

jtimon Oct 4, 2018

jtimon Oct 6, 2018

stevenroose Sep 27, 2018

jtimon Oct 4, 2018

kallewoof Oct 6, 2018

stevenroose Sep 27, 2018

instagibbs Sep 27, 2018

instagibbs Sep 27, 2018

kallewoof Sep 28, 2018

stevenroose Sep 28, 2018

kallewoof Oct 6, 2018

jtimon commented Oct 6, 2018

jtimon commented Oct 6, 2018

jtimon commented Oct 6, 2018

jtimon commented Oct 9, 2018

jtimon commented Oct 9, 2018

stevenroose commented Oct 10, 2018

instagibbs commented Oct 12, 2018

stevenroose commented Oct 16, 2018

instagibbs commented Oct 16, 2018

stevenroose commented Oct 16, 2018

jtimon commented Oct 16, 2018 •

edited

Loading

instagibbs commented Mar 28, 2019

kallewoof commented Mar 28, 2019 •

edited

Loading

instagibbs commented Mar 28, 2019

NOMERGE: Support signet as parent chain #414

Are you sure you want to change the base?

NOMERGE: Support signet as parent chain #414

Conversation

jtimon commented Sep 21, 2018 • edited Loading

jtimon commented Sep 22, 2018

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jtimon commented Sep 27, 2018

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jtimon commented Oct 6, 2018

jtimon commented Oct 6, 2018

jtimon commented Oct 6, 2018

jtimon commented Oct 9, 2018

jtimon commented Oct 9, 2018

stevenroose commented Oct 10, 2018

instagibbs commented Oct 12, 2018

stevenroose commented Oct 16, 2018

instagibbs commented Oct 16, 2018

stevenroose commented Oct 16, 2018

jtimon commented Oct 16, 2018 • edited Loading

instagibbs commented Mar 28, 2019

kallewoof commented Mar 28, 2019 • edited Loading

instagibbs commented Mar 28, 2019

jtimon commented Sep 21, 2018 •

edited

Loading

jtimon commented Oct 16, 2018 •

edited

Loading

kallewoof commented Mar 28, 2019 •

edited

Loading