Drop testnet3

Sjors commented at 4:22 pm on March 3, 2025: member

Testnet3 was deprecated in v28 with the introduction of testnet4. It will be dropped in v30 or later, please discuss in #31975.

There is no more [test] network section. The user is expected to set testnet4 explicitly, which should make future rotation of testnets easier.

If a [test] or [testnet3] section is present in bitcoin.conf a warning is issued:

0Warning: .../bitcoin.conf:18 Section [test] is not recognized.

Starting with -testnet3 will be greeted with an error message:

0Error: Testnet3 was dropped in version 30.0. Consider using '-testnet4' instead.

This also prevents a fallback to mainnet.

Starting with -test was already not possible with testnet3.

Review hints:

make a (Windows) Guix build and test the (un)installer

Guix hashes:

0(todo)

DrahtBot commented at 4:22 pm on March 3, 2025: contributor

The following sections might be updated with supplementary metadata relevant to reviewers and maintainers.

Code Coverage & Benchmarks

For details see: https://corecheck.dev/bitcoin/bitcoin/pulls/31974.

Reviews

See the guideline for information on the review process.

Type	Reviewers
Concept ACK	fjahr

If your review is incorrectly listed, please react with 👎 to this comment and the bot will ignore it on the next update.

Conflicts

Reviewers, this pull request conflicts with the following ones:

#32080 (OP_CHECKCONTRACTVERIFY by bigspider)
#29678 (Bugfix: Correct first-run free space checks by luke-jr)
#29039 (versionbits refactoring by ajtowns)
#28676 ([WIP] Cluster mempool implementation by sdaftuar)
#28201 (Silent Payments: sending by josibake)
#28122 (Silent Payments: Implement BIP352 by josibake)
#27260 (Enhanced error messages for invalid network prefix during address parsing. by russeree)
#26201 (Remove Taproot activation height by Sjors)

If you consider this pull request important, please also help to review the conflicting pull requests. Ideally, start with the one that should be merged first.

instagibbs commented at 4:27 pm on March 3, 2025: member

I’ve said it elsewhere, but I think it deserves a quite long deprecation cycle, along with some data gathering.

We don’t want services not upgrading for fixes because their test environment would break.

e.g., what exchanges with testnet environments have adopted testnet4 (or signet), much less would be ok with removing testnet3?

Does btcd support it? LND? Eclaire? CLN? etc.

Lots of work to be done here imo.

Sjors commented at 4:30 pm on March 3, 2025: member

I opened a separate RFC to discuss when to drop testnet3: #31975

I can rebase this for a while if we decide v30 is too early.

glozow added the label Tests on Mar 3, 2025

fjahr commented at 8:04 pm on March 3, 2025: contributor

Concept ACK

Sjors force-pushed on Mar 3, 2025

Sjors commented at 9:30 pm on March 3, 2025: member

I split 5d6aa13cf047e31005afca919ca9cf181e30905b into smaller commits. If we decide to punt this until after v30 then I’ll PR some of these separately. E.g. there’s some tests that can use testnet4 instead of testnet3 and perhaps we can drop GUI support earlier.

Sjors force-pushed on Mar 3, 2025

DrahtBot added the label Needs rebase on Mar 4, 2025

Sjors commented at 9:35 am on March 4, 2025: member

Will rebase after #31649 is rebased.

Sjors force-pushed on Mar 13, 2025

DrahtBot removed the label Needs rebase on Mar 13, 2025

Sjors force-pushed on Mar 14, 2025

Sjors commented at 8:25 am on March 14, 2025: member

Rebased after #31649.

in test/functional/wallet_crosschain.py:23 in 795876cfc2 outdated

19@@ -20,10 +20,7 @@ def skip_test_if_missing_module(self):
20     def setup_network(self):
21         self.add_nodes(self.num_nodes)
22 
23-        # Switch node 1 to testnet before starting it.
24-        self.nodes[1].chain = 'testnet3'
25-        self.nodes[1].extra_args = ['-maxconnections=0', '-prune=550'] # disable testnet sync
26-        self.nodes[1].replace_in_config([('regtest=', 'testnet='), ('[regtest]', '[test]')])
27+        # Node 1 is omitted, it was previously used with testnet3.

maflcko commented at 8:47 am on March 14, 2025:

This whole test seems a bit confusing, especially with this comment. The goal of the test is to check that a wallet doesn’t load cross-chain silently. It seems confusing to mention testnet3 and testnet4 here, but no other networks. Either, all test networks are tested, or only a single one. I’d argue that a single one should be sufficient.

Thus, a clean revert of git show 74a04f9e7ad6a16988149cc3438b9ce13c91cdb9 ./test/functional/wallet_crosschain.py seems more appropriate. Then, a follow-up to use the most stable test chain (signet?) and add a comment that only one test chain needs to be checked. Possibly, even better would be to just use the main chain, but I am not sure if the diff is trivial.

In any case, this change can probably be peeled off into a fresh/separate pull, because it makes sense on its own.

Sjors commented at 9:15 am on March 14, 2025:

I made a note to split this off, unless someone else gets to it first.

Sjors commented at 8:31 am on March 18, 2025:

#32088

in contrib/zmq/zmq_sub.py:37 in 795876cfc2 outdated

33@@ -34,7 +34,7 @@
34     print("This example only works with Python 3.5 and greater")
35     sys.exit(1)
36 
37-port = 28332
38+port = 48332

yancyribbens commented at 6:31 pm on March 14, 2025:

The wiki says the default port is 48333. Is this a typo?

Sjors commented at 6:36 pm on March 14, 2025:

48333 is the default p2p port

48332 is used for RPC, which is not defined in the BIP because it’s specific to Bitcoin Core (as is ZMQ).

Sjors force-pushed on Mar 19, 2025

Sjors commented at 9:08 am on March 19, 2025: member

Rebased after #32088 and #32057, which both simplify this PR.

I split the first five commits off into #32096.

in doc/bitcoin-conf.md:44 in 4e6e73a3b2 outdated

40@@ -41,7 +41,7 @@ Comments may appear in two ways:
41 ### Network specific options
42 
43 Network specific options can be:
44-- placed into sections with headers `[main]` (not `[mainnet]`), `[test]` (not `[testnet]`, for testnet3), `[testnet4]`, `[signet]` or `[regtest]`;
45+- placed into sections with headers `[main]` (not `[mainnet]`), `[testnet4]`, `[signet]` or `[regtest]`;

hodlinator commented at 9:10 pm on March 24, 2025:

Could this be part of #32096?

Another part is probably better to change at least as early: https://github.com/bitcoin/bitcoin/blob/b3162d10ea93cdec8e04150289ff61d6fed9a8e9/doc/bitcoin-conf.md?plain=1#L27

Probably better to use “regtest=1” or “signet=1”.

Sjors commented at 8:22 am on March 25, 2025:

This lists all the options available to the user, so I think it belongs here.

I’ll look into your suggestion on the next rebase.

Sjors commented at 8:17 am on April 2, 2025:

Changed to signet=1, guess that could have gone into #32096.

in contrib/linearize/example-linearize.cfg:12 in 4e6e73a3b2 outdated

 6@@ -7,9 +7,6 @@ host=127.0.0.1
 7 #mainnet default
 8 port=8332
 9 
10-#testnet default
11-#port=18332
12-

hodlinator commented at 9:12 pm on March 24, 2025:

Noticed while looking at #32096 that this file was untouched there.

Testnet3 has rpcport=18332 according to how chainparamsbase.cpp sets CBaseChainParams(..., rpc_port)… This config is setting the P2P ports for all networks so that they will collide with the default RPC ports and prevent successful startup (verified with the mainnet config line). Might as well just remove all ports IMO.

Sjors commented at 8:08 am on April 2, 2025:

I don’t really understand what this file is doing. Can you open a fresh PR to fix it?

in test/functional/feature_assumeutxo.py:106 in 4e6e73a3b2 outdated

102@@ -103,7 +103,6 @@ def expected_error(msg):
103         invalid_magics = [
104             # magic, name, real
105             [0xf9beb4d9, "main", True],
106-            [0x0b110907, "test", True],

hodlinator commented at 10:03 pm on March 24, 2025:

Why not move this to #32096? (23607f605c3156ca904f704666f7cbe4808cf76b there touches the entry from the vaguely related MAGIC_BYTES).

Why not have a “testnet4” entry?

Sjors commented at 8:23 am on March 25, 2025:

Indeed this should move.

(it’s handy if you mention the commit when making a comment)

We’ve only recently added a testnet4 snapshot, so maybe that’s why it wasn’t covered. I’ll add it in #32096.

in src/kernel/chainparams.cpp:225 in 4e6e73a3b2 outdated

220-        consensus.nMinerConfirmationWindow = 2016; // nPowTargetTimespan / nPowTargetSpacing
221-        consensus.vDeployments[Consensus::DEPLOYMENT_TESTDUMMY].bit = 28;
222-        consensus.vDeployments[Consensus::DEPLOYMENT_TESTDUMMY].nStartTime = Consensus::BIP9Deployment::NEVER_ACTIVE;
223-        consensus.vDeployments[Consensus::DEPLOYMENT_TESTDUMMY].nTimeout = Consensus::BIP9Deployment::NO_TIMEOUT;
224-        consensus.vDeployments[Consensus::DEPLOYMENT_TESTDUMMY].min_activation_height = 0; // No activation delay
225-

hodlinator commented at 11:07 pm on March 24, 2025:

(Inline thread in random position)

nit: src/external_signer.h references “test” and never “testnet4” which seems to be the argument the code would be sending in.

Sjors commented at 8:28 am on March 25, 2025:

I assume you’re talking about the chain argument? This is passed directly to HWI (or equivalent software). I’ll update the comment on the next rebase to swap “test” for “testnet4” which was missed when testnet4 was launched.

hodlinator commented at 9:31 am on March 25, 2025:

Exactly, from what I could see it would be set from ArgsManager::GetChainTypeString() in one case and CChainParams::GetChainTypeString() in another.

Sjors commented at 8:17 am on April 2, 2025:

I changed it to say “test network”

in src/kernel/chainparams.cpp:193 in 4e6e73a3b2 outdated

190@@ -191,101 +191,6 @@ class CMainParams : public CChainParams {
191     }
192 };
193

hodlinator commented at 10:11 am on March 25, 2025:

(Inline comment at random position):

Noticed doc/REST-interface.md still mentions testnet3 port: https://github.com/bitcoin/bitcoin/blob/4e6e73a3b2c5c3a4a07683bc3a6b59f8e27b78f3/doc/REST-interface.md?plain=1#L6-L7

Sjors commented at 10:15 am on March 25, 2025:

Speaking of things I’d like to delete entirely :-) (but won’t)

Sjors commented at 8:17 am on April 2, 2025:

Fixed

ryanofsky referenced this in commit 80e47b1920 on Apr 1, 2025

DrahtBot added the label Needs rebase on Apr 1, 2025

doc: use "test nework" in external_signer.h 834f8372dc

doc: use signet=1 in bitcoin-conf.md example 9abe0d4abb

gui: drop testnet3 7536be28ba

test: drop testnet3 coverage

Except for the deprecation warning in feature_config_args.py.
This test now writes its own config file, which was previously done by
test_acceptstalefeeestimates_arg_support.

The deprecation warning test will be removed in a later commit.

5cb1738951

Disallow -testnet3

Abort node startup and inform user they should use -testnet4.
Modify the deprecation message test to cover this.

Remove testnet3 references from bitcoin-cli documentation
and bash helper.

902f6af2be

Drop testnet3 seed nodes 6125f46e14

Drop testnet3 from docs and dev tools

Also use "test network" instead of "testnet" where appropriate (e.g.
because it also applies to signet).

78a53c79ed

consensus: drop testnet3

This drops the chain parameters and ChainType::TESTNET.

b038fa4493

validation: drop testnet3 comment about BIP34 0f6e77b713

doc: add release note e402cf241a

Sjors force-pushed on Apr 2, 2025

Sjors commented at 8:17 am on April 2, 2025: member

Rebased after #32096.

DrahtBot removed the label Needs rebase on Apr 2, 2025

Drop testnet3 #31974

Code Coverage & Benchmarks

Reviews

Conflicts