have verify_account_authority check passed-in keys #1384

jmjatlanta · 2018-10-16T16:11:23Z

Potential fix for #1366

The original code always returned false, due to not using the passed in keys to do anything.

This PR uses a different approach. Compare the public keys of what was passed in versus what is known.

oxarbitrage · 2018-10-16T19:03:19Z

libraries/app/database_api.cpp

@@ -167,7 +167,7 @@ class database_api_impl : public std::enable_shared_from_this<database_api_impl>
      vector<withdraw_permission_object> get_withdraw_permissions_by_giver(const std::string account_id_or_name, withdraw_permission_id_type start, uint32_t limit)const;
      vector<withdraw_permission_object> get_withdraw_permissions_by_recipient(const std::string account_id_or_name, withdraw_permission_id_type start, uint32_t limit)const;

-   //private:
+   // private:


we maybe just remove this comment ?

Yes, I tried (hence the space I accidentally added). It did not compile due to _db being used outside, plus some others. I was thinking of fixing all of it, but resisted the urge because I was unsure why it was originally commented out.

abitmore · 2018-10-17T00:45:33Z

IMHO this is not the correct approach, since you ignored multi-auth accounts at all.

Correct approach would be to call verify_authority function somehow with available public keys as "approved".

abitmore

As commented.

jmjatlanta · 2018-10-17T10:50:13Z

Thank you @abitmore for partially answering a question I had.

Is the purpose of this function to prove that the public keys are valid? Or is this function saying "If you send a transaction signed by the private keys related to these public keys, it will go through"?

If you want to prove that those public keys are related to this account, well, I think I am done.

If you want to take in to account multi-auth accounts, then more will need to be added. And now with projects being worked on that make keys more granular, I believe it would depend on what type of transaction you are doing as to if such a signature would pass or fail.

If my thinking and assumptions on the last paragraph are correct, perhaps it would be best to pause working on this issue until those features are added.

Thoughts?

pmconrad · 2018-10-17T13:45:05Z

Doc comment says

@return true if the signers have enough authority to authorize an account

IMO it is clear that multi-level signatures must be taken into account.

jmjatlanta · 2018-10-19T21:22:36Z

In the latest approach, I am using graphene::chain::verify_authority with a transfer operation and the passed in account and keys. I have not tested with a multi-level signature account. I will take that on next, unless someone raises their hand and says this approach is also invalid.

cogutvalera · 2018-10-20T07:55:20Z

libraries/app/database_api.cpp

+   std::vector<operation> ops;
+   ops.emplace_back(op);
+
+   try


why not just to create signed_transaction and call it's verify_authority method ?
So the full implementation of verify_account_authority might be like next:

bool database_api_impl::verify_account_authority( const string& account_name_or_id, const flat_set<public_key_type>& keys )const { // create a dummy transfer transfer_operation op; op.from = get_account_from_string(account_name_or_id)->id; std::vector<operation> ops; ops.emplace_back(op); signed_transaction trx; trx.signees = keys; trx.operations.push_back(op); return verify_authority(trx); }

It's less code and more simple IMHO.

Nice idea, but IMO the signees field should be used as an internal cache only, not part of the tx struct's interface.

graphene::chain::verify_authority will use signees returned by signed_transaction::get_signature_keys method, so it should work

jmjatlanta · 2018-10-23T01:12:49Z

I have added a muti-signature test (2 of 3) to wallet_test. Please let me know what you think.

cogutvalera

described in comments

cogutvalera · 2018-10-23T08:13:07Z

tests/tests/wallet_tests.cpp

+      try {
+         account_update_operation op;
+         op.account = nathan.id;
+         op.active = authority(2, public_key_type(nathan_key1.get_public_key()), 1, public_key_type(nathan_key2.get_public_key()), 1, public_key_type(nathan_key3.get_public_key()), 1);


this line is too long, please reduce it something like next example of code:

op.active = authority( 2, public_key_type(nathan_key1.get_public_key()), 1, public_key_type(nathan_key2.get_public_key()), 1, public_key_type(nathan_key3.get_public_key()), 1 );

cogutvalera · 2018-10-23T09:05:25Z

All tests (verify_account_authority & any_two_of_three) have passed successfully with no errors.
Nice work ! Thanks !

cogutvalera · 2018-10-23T09:06:00Z

IMHO just required little improvements with that line as commented above ...

cogutvalera

Thanks !

pmconrad · 2018-10-23T12:03:51Z

tests/tests/wallet_tests.cpp

+      edump((e.to_detail_string()));
+      throw;
+   }
+}


Please add a test where nathan's account has a multisig from one key and two other accounts.

Added what I think you are asking. It is not much different than the last, so I may be misunderstanding your request.

IMHO we should check different combinations as in any_two_of_three test, perhaps the same combinations, but with 3 accounts, 1 key per 1 account, just my thoughts ...

pmconrad · 2018-10-23T17:57:11Z

tests/tests/wallet_tests.cpp

+         account_update_operation op;
+         op.account = nathan.id;
+         op.active = authority(3, public_key_type(nathan_key.get_public_key()), 1, 
+               public_key_type(alice_key.get_public_key()), 1, public_key_type(bob_key.get_public_key()), 1);


Here you should use alice_id and bob_id, not their keys

pmconrad · 2018-10-23T17:57:34Z

tests/tests/wallet_tests.cpp

+   try {
+      fc::ecc::private_key nathan_key = fc::ecc::private_key::regenerate(fc::digest("key1"));
+      fc::ecc::private_key alice_key = fc::ecc::private_key::regenerate(fc::digest("key2"));
+      fc::ecc::private_key bob_key = fc::ecc::private_key::regenerate(fc::digest("key3"));


Please use ACTORS macro

pmconrad · 2018-10-23T17:57:59Z

tests/tests/wallet_tests.cpp

+      	public_keys.emplace(bob_key.get_public_key());
+      	BOOST_CHECK(db_api.verify_account_authority("nathan", public_keys));
+      }
+


Also need a negative test, e. g. alice_key missing.

cogutvalera · 2018-10-23T18:30:42Z

tests/tests/wallet_tests.cpp

-         op.active = authority(3, public_key_type(nathan_key.get_public_key()), 1, 
-               public_key_type(alice_key.get_public_key()), 1, public_key_type(bob_key.get_public_key()), 1);
+         op.active = authority(3, nathan_public_key, 1, 
+               alice.id, 1, bob.id, 1);


IMHO we don't need to split this line, it isn't long now

cogutvalera

Thanks !

pmconrad · 2018-10-23T18:43:15Z

Good, thanks!

Stale

abitmore · 2018-10-23T19:29:59Z

libraries/app/include/graphene/app/database_api.hpp

@@ -663,9 +663,12 @@ class database_api
      bool           verify_authority( const signed_transaction& trx )const;

      /**
-       * @return true if the signers have enough authority to authorize an account
+       * @brief Verify that the public keys have enough authority to authorize a transaction


Why this change? There is no transaction in the API.

abitmore · 2018-10-23T19:30:34Z

libraries/app/include/graphene/app/database_api.hpp

+       * @brief Verify that the public keys have enough authority to authorize a transaction
+       * @param account_name_or_id the account to check
+       * @param signers the public keys
+       * @return true if the passed in keys have enough authority to authorize a transaction


Why this? There is no transaction in the API.

abitmore · 2018-10-23T19:32:56Z

tests/tests/wallet_tests.cpp

+      try {
+
+         ACTORS( (nathan) );
+         graphene::app::database_api db_api(db);


I think it's better to move these test cases to database_api_tests.cpp.

abitmore · 2018-10-23T19:36:42Z

Code looks fine to me.

abitmore · 2018-10-23T20:34:30Z

libraries/app/include/graphene/app/database_api.hpp

@@ -663,9 +663,12 @@ class database_api
      bool           verify_authority( const signed_transaction& trx )const;

      /**
-       * @return true if the signers have enough authority to authorize an account
+       * @brief Verify that the public keys have enough authority to approve an operation


I think we should mention the account in this sentence and the return

I think the latest commit will be more to your liking. Please take a look.

initial fix for verify_account_authority

7d44c3a

oxarbitrage reviewed Oct 16, 2018

View reviewed changes

abitmore previously requested changes Oct 17, 2018

View reviewed changes

jmjatlanta added 2 commits October 19, 2018 16:09

improved way to verify account and keys

824ed4f

Removed unneeded method definition

48fa400

cogutvalera reviewed Oct 20, 2018

View reviewed changes

Added multiple signatures to test

c2b63f5

cogutvalera suggested changes Oct 23, 2018

View reviewed changes

shorten line length

5db31f5

cogutvalera previously approved these changes Oct 23, 2018

View reviewed changes

pmconrad reviewed Oct 23, 2018

View reviewed changes

add test for multisig

7e4e033

jmjatlanta dismissed cogutvalera’s stale review via 7e4e033 October 23, 2018 17:47

pmconrad suggested changes Oct 23, 2018

View reviewed changes

improved multisig test

e03d329

cogutvalera reviewed Oct 23, 2018

View reviewed changes

cogutvalera previously approved these changes Oct 23, 2018

View reviewed changes

pmconrad previously approved these changes Oct 23, 2018

View reviewed changes

fix short line

673ac2b

jmjatlanta dismissed stale reviews from pmconrad and cogutvalera via 673ac2b October 23, 2018 18:50

cogutvalera previously approved these changes Oct 23, 2018

View reviewed changes

abitmore reviewed Oct 23, 2018

View reviewed changes

moved tests, fixed comments

f523ad6

jmjatlanta dismissed cogutvalera’s stale review via f523ad6 October 23, 2018 19:51

cogutvalera previously approved these changes Oct 23, 2018

View reviewed changes

abitmore reviewed Oct 23, 2018

View reviewed changes

abitmore mentioned this pull request Oct 23, 2018

Fix #584 authorize transactions with non-immediate owner authorities #1259

Merged

4 tasks

clarify comment

b078310

jmjatlanta dismissed cogutvalera’s stale review via b078310 October 24, 2018 12:12

abitmore approved these changes Oct 25, 2018

View reviewed changes

abitmore added this to the 201812 - Feature Release milestone Oct 25, 2018

pmconrad merged commit 2e15e77 into develop Oct 25, 2018

pmconrad deleted the jmj_1366 branch October 25, 2018 12:37

pmconrad mentioned this pull request Oct 25, 2018

database_api::verify_account_authority is broken #1366

Closed

17 tasks

pmconrad added a commit that referenced this pull request Oct 25, 2018

Adapted new tests from #1384

bbb8730

pmconrad mentioned this pull request Oct 25, 2018

Resolve 1371: sorting of public_keys is inefficient #1401

Closed

pmconrad mentioned this pull request Dec 21, 2018

Testnet X-Mas Sneak Preview Release #1489

Merged

MichelSantos mentioned this pull request Feb 7, 2019

Release Notes for the 201902 Feature Release of BitShares Core #1565

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

have verify_account_authority check passed-in keys #1384

have verify_account_authority check passed-in keys #1384

jmjatlanta commented Oct 16, 2018

oxarbitrage Oct 16, 2018

jmjatlanta Oct 16, 2018

abitmore commented Oct 17, 2018

abitmore left a comment

jmjatlanta commented Oct 17, 2018 •

edited

Loading

pmconrad commented Oct 17, 2018

jmjatlanta commented Oct 19, 2018

cogutvalera Oct 20, 2018

pmconrad Oct 20, 2018

cogutvalera Oct 20, 2018

jmjatlanta commented Oct 23, 2018

cogutvalera left a comment

cogutvalera Oct 23, 2018 •

edited

Loading

cogutvalera commented Oct 23, 2018

cogutvalera commented Oct 23, 2018

cogutvalera left a comment

pmconrad Oct 23, 2018

jmjatlanta Oct 23, 2018

cogutvalera Oct 23, 2018

pmconrad Oct 23, 2018

pmconrad Oct 23, 2018

pmconrad Oct 23, 2018

cogutvalera Oct 23, 2018

cogutvalera left a comment

pmconrad commented Oct 23, 2018

abitmore Oct 23, 2018

abitmore Oct 23, 2018

abitmore Oct 23, 2018

abitmore commented Oct 23, 2018

abitmore Oct 23, 2018

jmjatlanta Oct 24, 2018

have verify_account_authority check passed-in keys #1384

have verify_account_authority check passed-in keys #1384

Conversation

jmjatlanta commented Oct 16, 2018

Choose a reason for hiding this comment

Choose a reason for hiding this comment

abitmore commented Oct 17, 2018

abitmore left a comment

Choose a reason for hiding this comment

jmjatlanta commented Oct 17, 2018 • edited Loading

pmconrad commented Oct 17, 2018

jmjatlanta commented Oct 19, 2018

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jmjatlanta commented Oct 23, 2018

cogutvalera left a comment

Choose a reason for hiding this comment

cogutvalera Oct 23, 2018 • edited Loading

Choose a reason for hiding this comment

cogutvalera commented Oct 23, 2018

cogutvalera commented Oct 23, 2018

cogutvalera left a 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

cogutvalera left a comment

Choose a reason for hiding this comment

pmconrad commented Oct 23, 2018

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

abitmore commented Oct 23, 2018

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jmjatlanta commented Oct 17, 2018 •

edited

Loading

cogutvalera Oct 23, 2018 •

edited

Loading