diff options
| author | fanquake <fanquake@gmail.com> | 2023-06-02 16:03:50 +0100 |
|---|---|---|
| committer | fanquake <fanquake@gmail.com> | 2023-06-02 16:18:11 +0100 |
| commit | 436c185b05200e28696ef6ff64e679c3148f72bb (patch) | |
| tree | 1e6bd6f37c1137583f4f70193a5f6927f8e547d2 /src/univalue | |
| parent | b22408df162a224d94ac54e8443b57ef3fd2ca72 (diff) | |
| parent | cfbc8a623b5133f1d0b0c0c9be73b2b107e0d687 (diff) | |
| download | bitcoin-436c185b05200e28696ef6ff64e679c3148f72bb.tar.xz | |
Merge bitcoin/bitcoin#27256: refactor: rpc: Remove unnecessary uses of ParseNonRFCJSONValue() and rename it
cfbc8a623b5133f1d0b0c0c9be73b2b107e0d687 refactor: rpc: hide and rename ParseNonRFCJSONValue() (stickies-v)
6c8bde6d54d03224709dce54b8ba32b8c3e37ac7 test: move coverage on ParseNonRFCJSONValue() to UniValue::read() (stickies-v)
Pull request description:
Follow-up to https://github.com/bitcoin/bitcoin/pull/26612#issuecomment-1453623741. As per https://github.com/bitcoin/bitcoin/pull/26506#pullrequestreview-1211984059, `ParseNonRFCJSONValue()` is no longer necessary and we can use `UniValue::read()` directly:
> IIRC before that PR UniValue::read could only parse JSON object and array values, but now it will parse string/number/boolean/null values as well. laanwj pointed this out in https://github.com/bitcoin/bitcoin/issues/9028#issuecomment-257885368
The implementation of `ParseNonRFCJSONValue()` was already [simplified in #26612](https://github.com/bitcoin/bitcoin/pull/26612/files#diff-84c7a7f36362b9724c31e5dec9879b2f81eae0d0addbc9c0933c3558c577de65R259-R263) and [test coverage updated](https://github.com/bitcoin/bitcoin/pull/26612/files#diff-fc0f86b6c3bb23b0e983bcf79d7546d1c9eaa15d6e4d8a7b03b5b85955f585abR292-R312) to ensure behaviour didn't change.
To avoid code duplication, we keep the function to throw on invalid input data but rename it to `Parse()` and remove it from the header.
The existing test coverage we had on `ParseNonRFCJSONValue()` is moved over to `UniValue::read()`.
ACKs for top commit:
ryanofsky:
Code review ACK cfbc8a623b5133f1d0b0c0c9be73b2b107e0d687. Only change since last review is adding a new test
Tree-SHA512: 89be959d2353af7ace0c1348ba1600b9ac1d3c7b3cf7f0b59f6e005b7fb9d695ce3e8720e1be3cf77fe7e318a4017c880df108928e7179ec50447583d13bc781
Diffstat (limited to 'src/univalue')
| -rw-r--r-- | src/univalue/test/object.cpp | 27 |
1 files changed, 27 insertions, 0 deletions
diff --git a/src/univalue/test/object.cpp b/src/univalue/test/object.cpp index 5ddf300393..5fb973c67b 100644 --- a/src/univalue/test/object.cpp +++ b/src/univalue/test/object.cpp @@ -412,6 +412,33 @@ void univalue_readwrite() BOOST_CHECK_EQUAL(strJson1, v.write()); + // Valid + BOOST_CHECK(v.read("1.0") && (v.get_real() == 1.0)); + BOOST_CHECK(v.read("true") && v.get_bool()); + BOOST_CHECK(v.read("[false]") && !v[0].get_bool()); + BOOST_CHECK(v.read("{\"a\": true}") && v["a"].get_bool()); + BOOST_CHECK(v.read("{\"1\": \"true\"}") && (v["1"].get_str() == "true")); + // Valid, with leading or trailing whitespace + BOOST_CHECK(v.read(" 1.0") && (v.get_real() == 1.0)); + BOOST_CHECK(v.read("1.0 ") && (v.get_real() == 1.0)); + BOOST_CHECK(v.read("0.00000000000000000000000000000000000001e+30 ") && v.get_real() == 1e-8); + + BOOST_CHECK(!v.read(".19e-6")); //should fail, missing leading 0, therefore invalid JSON + // Invalid, initial garbage + BOOST_CHECK(!v.read("[1.0")); + BOOST_CHECK(!v.read("a1.0")); + // Invalid, trailing garbage + BOOST_CHECK(!v.read("1.0sds")); + BOOST_CHECK(!v.read("1.0]")); + // Invalid, keys have to be names + BOOST_CHECK(!v.read("{1: \"true\"}")); + BOOST_CHECK(!v.read("{true: 1}")); + BOOST_CHECK(!v.read("{[1]: 1}")); + BOOST_CHECK(!v.read("{{\"a\": \"a\"}: 1}")); + // BTC addresses should fail parsing + BOOST_CHECK(!v.read("175tWpb8K1S7NmH4Zx6rewF9WQrcZv245W")); + BOOST_CHECK(!v.read("3J98t1WpEZ73CNmQviecrnyiWrnqRhWNL")); + /* Check for (correctly reporting) a parsing error if the initial JSON construct is followed by more stuff. Note that whitespace is, of course, exempt. */ |