Tezos integration #292

Adman · 2018-07-27T12:08:37Z

Updated fields of messages trezor/trezor-common#170

prusnak · 2018-07-27T12:23:30Z

Please fix the failure in https://travis-ci.org/trezor/trezor-core/jobs/408917280#L852-L854

Adman · 2018-07-27T15:17:05Z

@prusnak fixed

tsusanka · 2018-08-24T12:58:17Z

@Adman could you add python-trezor support and some device tests along the way please? We have two types of tests:

unit tests, directly included in trezor-core/trezor-mcu repositories, which you've added so that's great
device tests (aka integration tests), which are part of the python-trezor repository and are "trezor model independent". In other words those are tests that can be run against both Trezor One and Trezor T. They actually send protobuf messages to the device and expect some other protobuf messages with results

Have a look on the Cardano's or Ripple's PR for example for some inspiration.

Adman · 2018-08-27T10:43:35Z

@tsusanka ok, I'll add python-trezor support and make a PR.

Adman · 2018-08-28T13:28:45Z

@tsusanka Can you give me a hint how to run device_tests in python-trezor?
When running python -m pytest -v from python-trezor/trezorlib/tests/device_tests folder it fails
StopIteration ERROR: could not load python-trezor/trezorlib/tests/device_tests/conftest.py.

unit_tests work properly.

prusnak · 2018-08-28T13:33:28Z

Run emulator first via PYOPT=0 ./emu.sh Then in another window run "pytest" in device_tests directory.

Adman · 2018-08-31T12:50:19Z

@tsusanka @prusnak Here's the python-trezor implementation: trezor/python-trezor#302

tsusanka · 2018-09-03T07:45:50Z

@tsusanka @prusnak Here's the python-trezor implementation: trezor/python-trezor#302

Awesome! We are currently in the process of finalizing our next release. When that's done we'll review your PR completely. Thanks

tsusanka · 2018-09-05T13:02:13Z

@Adman What do you think about supporting just one curve instead of all three? Supporting all three brings unnecessary complexity. We can add the other later but I believe in the first version of Tezos support we should use just one. You can choose (maybe do a bit of research what is used the most?) but I suggest secp256k or ed25519. What do you think?

prusnak · 2018-09-05T14:49:14Z

As per agreement on Gitter, we'll support only Ed25519 curve = tz1. Relevant commit to trezor-common: trezor/trezor-common@b6f59d3

Please update this PR (and PR to python-trezor) to use only tz1/Ed25519 curve. Thank you!

Adman · 2018-09-07T12:12:34Z

@prusnak @tsusanka done.

tsusanka · 2018-09-07T12:33:08Z

@Adman please do a rebase (let me know if you need help), the tests should be passing then

Adman · 2018-09-07T12:44:24Z

@tsusanka rebased, but master is failing because of cardano's get_public_key and address tests https://travis-ci.org/trezor/trezor-core/jobs/425716279#L1550

prusnak · 2018-09-07T12:55:19Z

You did not rebase the commits - you did merge. Rebase is being done with git rebase master. https://git-scm.com/docs/git-rebase

Be careful, this can be quite dangerous, but if there are no conflicts, it should be pretty straightforward.

Adman · 2018-09-07T13:12:30Z

@prusnak There were some conflicts, but I hope everything's fine now.

tsusanka · 2018-09-10T15:11:14Z

Now that python-trezor is merged, please do a rebase and set run_xfail = tezos in setup.cfg. That way we'll see the tests are passing. I'm working on a review, I'll publish it tonight/tomorrow.

Signed-off-by: Adrian Matejov <adrian.matejov@simplestaking.com>

Adman · 2018-09-10T16:53:23Z

@tsusanka Ok, I rebased and set run_xfail = tezos

tsusanka

Thanks for you PR, consider this as a preliminary review, it's mostly nit-picks. I've requested some changes and have a general question:

Do you plan to use the GetPublicKey functions in particular on your front-end side? Or did you implement them just to fit the others? Since ed25519 does not support non-hardened derivation, therefore no xpub functionality, I am not sure what the use case is. If we don't have any in mind, let's remove it.

tsusanka · 2018-09-10T12:46:32Z

src/apps/tezos/get_address.py

+    node = await seed.derive_node(ctx, address_n, TEZOS_CURVE)
+
+    sk = node.private_key()
+    pk = ed25519.publickey(sk)


Let's replace these two lines with:

pk = seed.remove_ed25519_prefix(node.public_key())

so it's consistent with stellar and others.

you mean seed.remove_ed25519_prefix(node.private_key()), right?

Ah, no, you are right, sorry

No, private_key does not have a prefix. node.public_key() is computed so you can use it.

tsusanka · 2018-09-10T12:52:30Z

src/apps/tezos/get_address.py

+    sk = node.private_key()
+    pk = ed25519.publickey(sk)
+    pkh = hashlib.blake2b(pk, outlen=20).digest()
+    address = b58cencode(pkh, prefix=tezos_get_address_prefix(0))


The zero in tezos_get_address_prefix(0) looks kind of magic. Maybe let's simplify this to b58cencode(pkh, prefix="tz1") or maybe better b58cencode(pkh, prefix=TEZOS_ED25519_PREFIX) ?

Also why is the function called b58cencode and not b58encode?

it's called b58cencode because it is "check_encode" (not pure "encode")

Will call the prefix TEZOS_ED25519_ADDRESS_PREFIX

Let's not be afraid of long names and use base58_encode_check

tsusanka · 2018-09-10T13:01:26Z

src/apps/tezos/get_public_key.py

+    node = await seed.derive_node(ctx, address_n, TEZOS_CURVE)
+
+    sk = node.private_key()
+    pk = ed25519.publickey(sk)


tsusanka · 2018-09-10T13:05:57Z

src/apps/tezos/layout.py

+
+
+def split_address(address):
+    return chunks(address, 18)


Let's use common.layout.split_address. It has size=17, but that doesn't matter that much or will be refactored together later

If I use 17, the address in transaction won't fit on display.

Ok, keep it.

tsusanka · 2018-09-11T09:12:14Z

src/apps/tezos/sign_tx.py

+
+def _encode_common(operation, str_operation):
+    operation_tags = {"reveal": 7, "transaction": 8, "origination": 9, "delegation": 10}
+    result = ustruct.pack("<b", operation_tags[str_operation])


All those ustruct.pack are supposed to convert unsigned ints to bytes, aren't they? And the ints will be always from 0 to 255? I believe it should be <B, but one way or another let's refactor this to use app.common.writers. If you indeed want to serialize uint8 you can use write_uint8.

tsusanka · 2018-09-11T09:15:06Z

src/apps/tezos/sign_tx.py

+
+
+def _encode_contract_id(contract_id):
+    return ustruct.pack("<b", contract_id.tag) + contract_id.hash


For example this:

def _encode_contract_id(w: bytearray, contract_id): write_uint8(w, contract_id.tag) write_bytes(w, contract_id.hash)

tsusanka · 2018-09-11T09:24:56Z

src/apps/tezos/sign_tx.py

+
+    opbytes = _get_operation_bytes(msg)
+
+    watermark = bytes([3])


Could you maybe put a small comment why '3'?

https://groups.google.com/forum/#!msg/tezos-developer-community/bD7DFkLT8QE/X4dNDp0ABgAJ

I will add a small comment to code

tsusanka · 2018-09-11T09:26:33Z

src/apps/tezos/sign_tx.py

+    else:
+        raise wire.DataError("Invalid operation")
+
+    opbytes = _get_operation_bytes(msg)


Have a look on Stellar, we usually declare a bytearray, which we then pass as an argument. I'm providing an example in one of the following comments.

Adman · 2018-09-11T11:36:11Z

@tsusanka Requested changes applied.

It would be cool if the common split_address function had optional parameter size, because using size=17 the address won't fit on display. I kept the function there for now.

GetPublicKey - @jurajselep will post a comment about it here.

jurajselep · 2018-09-11T11:59:59Z

When we inject transaction, origination or delegation into Tezos node for first time, we need to "reveal" public_key.
http://tezos.gitlab.io/betanet/api/p2p.html?highlight=reveal

We are using GetPublicKey on front-end side for operation preapply and forge .

http://tezos.gitlab.io/betanet/api/rpc.html?highlight=reveal#post-block-id-helpers-forge-operations
http://tezos.gitlab.io/betanet/api/rpc.html?highlight=reveal#post-block-id-helpers-preapply-operations

{ 
 "kind": "reveal",
 "source": $contract_id,
 "fee": $mutez,
 "counter": $positive_bignum,
 "gas_limit": $positive_bignum,
 "storage_limit": $positive_bignum,
 "public_key": $Signature.Public_key 
}

tsusanka · 2018-09-11T12:15:57Z

src/apps/tezos/get_address.py

    pkh = hashlib.blake2b(pk, outlen=20).digest()
-    address = b58cencode(pkh, prefix=tezos_get_address_prefix(0))
+    address = b58cencode(pkh, prefix=TEZOS_ED25519_ADDRESS_PREFIX)


Nice. Now we can use this also in sign_tx._get_address_by_tag and remove tezos_get_address_prefix completely.

tsusanka · 2018-09-11T12:19:26Z

@jurajselep thanks for the explanation, that's fine by me.

@Adman please note, that the tests are failing currently becuase of style

Adman · 2018-09-11T12:43:36Z

@tsusanka Tests should be passing, but travis build failed with message The command "curl -LO "https://github.com/google/protobuf/releases/download/v${PROTOBUF_VERSION}/protoc-${PROTOBUF_VERSION}-linux-x86_64.zip"" failed and exited with 56.
I assume it's problem on travis side.

Signed-off-by: Adrian Matejov <adrian.matejov@simplestaking.com>

Adman · 2018-09-11T13:03:20Z

@tsusanka Ok, I restarted the build and it is passing. If you have further comments on this PR I will fix them.

tsusanka · 2018-09-11T13:24:31Z

Great. Looks good to me. I'd like @jpochyla to have a final look later, but this is on a good way to be merged. Please be patient, he's busy with other things.

@jpochyla the only thing where I'm not sure is whether _encode_zarith could be simplified or moved to writers with a better name.

tACK dc6e405

tsusanka · 2018-09-11T13:25:12Z

(grr, the closing was of course unintentional)

tsusanka · 2018-09-11T13:36:02Z

Maybe one more thing. Could you add a brief README? Have a look on Ripple and feel free to omit the maintainer/reviewer fields. It doesn't have to be verbose, but some links to documentation/explorer etc. would be nice.

Adman · 2018-09-11T14:37:49Z

@tsusanka Readme added.

Signed-off-by: Adrian Matejov <adrian.matejov@simplestaking.com>

tsusanka · 2018-09-13T13:06:16Z

Thank you!

Adman · 2018-09-13T20:19:09Z

Thanks all for the collaboration!

prusnak added this to the v2.0.8 milestone Jul 30, 2018

prusnak added altcoin firmware labels Jul 30, 2018

szymonlesisz mentioned this pull request Sep 7, 2018

Tezos integration trezor/connect#208

Merged

tsusanka self-requested a review September 10, 2018 12:44

Adrian Matejov added 5 commits September 10, 2018 18:19

tezos: added tezos cryptocurrency

84be7cd

Signed-off-by: Adrian Matejov <adrian.matejov@simplestaking.com>

tezos: display_address -> show

f2d5a36

Signed-off-by: Adrian Matejov <adrian.matejov@simplestaking.com>

tezos: removed duplicated code

2a15331

Signed-off-by: Adrian Matejov <adrian.matejov@simplestaking.com>

tezos: supporting only ed25519 curve

c7b7df7

Signed-off-by: Adrian Matejov <adrian.matejov@simplestaking.com>

tezos: run_xfail set

7c60790

Signed-off-by: Adrian Matejov <adrian.matejov@simplestaking.com>

tsusanka suggested changes Sep 11, 2018

View reviewed changes

tezos: fixed nitpicks

dc6e405

Signed-off-by: Adrian Matejov <adrian.matejov@simplestaking.com>

tsusanka closed this Sep 11, 2018

tsusanka reopened this Sep 11, 2018

tsusanka approved these changes Sep 11, 2018

View reviewed changes

tezos: added Readme

1709902

Signed-off-by: Adrian Matejov <adrian.matejov@simplestaking.com>

jpochyla approved these changes Sep 13, 2018

View reviewed changes

tsusanka merged commit 7918e1f into trezor:master Sep 13, 2018



		def _encode_contract_id(contract_id):
		return ustruct.pack("<b", contract_id.tag) + contract_id.hash

Tezos integration #292

Tezos integration #292

Conversation

Adman commented Jul 27, 2018

prusnak commented Jul 27, 2018

Adman commented Jul 27, 2018

tsusanka commented Aug 24, 2018

Adman commented Aug 27, 2018

Adman commented Aug 28, 2018 • edited Loading

prusnak commented Aug 28, 2018 via email

Adman commented Aug 31, 2018

tsusanka commented Sep 3, 2018

tsusanka commented Sep 5, 2018

prusnak commented Sep 5, 2018

Adman commented Sep 7, 2018

tsusanka commented Sep 7, 2018

Adman commented Sep 7, 2018 • edited Loading

prusnak commented Sep 7, 2018

Adman commented Sep 7, 2018

tsusanka commented Sep 10, 2018 • edited Loading

Adman commented Sep 10, 2018

tsusanka left a comment • edited Loading

Choose a reason for hiding this comment

tsusanka Sep 10, 2018 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

prusnak Sep 11, 2018 • edited Loading

Choose a reason for hiding this comment

tsusanka Sep 10, 2018 • edited Loading

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

Adman commented Sep 11, 2018

jurajselep commented Sep 11, 2018 • edited Loading

Choose a reason for hiding this comment

tsusanka commented Sep 11, 2018 • edited Loading

Adman commented Sep 11, 2018 • edited Loading

Adman commented Sep 11, 2018

tsusanka commented Sep 11, 2018 • edited Loading

tsusanka commented Sep 11, 2018 • edited Loading

tsusanka commented Sep 11, 2018

Adman commented Sep 11, 2018

tsusanka commented Sep 13, 2018

Adman commented Sep 13, 2018

Adman commented Aug 28, 2018 •

edited

Loading

Adman commented Sep 7, 2018 •

edited

Loading

tsusanka commented Sep 10, 2018 •

edited

Loading

tsusanka left a comment •

edited

Loading

tsusanka Sep 10, 2018 •

edited

Loading

prusnak Sep 11, 2018 •

edited

Loading

tsusanka Sep 10, 2018 •

edited

Loading

jurajselep commented Sep 11, 2018 •

edited

Loading

tsusanka commented Sep 11, 2018 •

edited

Loading

Adman commented Sep 11, 2018 •

edited

Loading

tsusanka commented Sep 11, 2018 •

edited

Loading

tsusanka commented Sep 11, 2018 •

edited

Loading