Conversation
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
| enable_if_t < std::is_arithmetic<ArithmeticType>::value&& | ||
| !std::is_same<ArithmeticType, typename BasicJsonType::boolean_t>::value, | ||
| int > = 0 > | ||
| template<typename BasicJsonType, typename ArithmeticType, enable_if_t<std::is_arithmetic<ArithmeticType>::value && !std::is_same<ArithmeticType, typename BasicJsonType::boolean_t>::value, int> = 0> |
There was a problem hiding this comment.
Ugh, not sure this is a good change. Is there a setting that will prevent it?
| typename StringType, | ||
| enable_if_t< | ||
| std::is_assignable<StringType&, const typename BasicJsonType::string_t>::value && is_detected_exact<typename BasicJsonType::string_t::value_type, value_type_t, StringType>::value && !std::is_same<typename BasicJsonType::string_t, StringType>::value && !is_json_ref<StringType>::value, | ||
| int> = 0> |
There was a problem hiding this comment.
Interesting that this one stayed multi-line. Perhaps there is a max line length setting?
There was a problem hiding this comment.
There is, but I have not tried it yet.
There was a problem hiding this comment.
I think it is a bit better with max width of ColumnLimit=160.
| !std::is_same<ConstructibleArrayType, typename BasicJsonType::binary_t>::value&& | ||
| !is_basic_json<ConstructibleArrayType>::value, | ||
| int > = 0 > | ||
| template<typename BasicJsonType, typename ConstructibleArrayType, enable_if_t<is_constructible_array_type<BasicJsonType, ConstructibleArrayType>::value && !is_constructible_object_type<BasicJsonType, ConstructibleArrayType>::value && !is_constructible_string_type<BasicJsonType, ConstructibleArrayType>::value && !std::is_same<ConstructibleArrayType, typename BasicJsonType::binary_t>::value && !is_basic_json<ConstructibleArrayType>::value, int> = 0> |
There was a problem hiding this comment.
This one is much longer, must be some other trigger for single line vs multi-line.
| } | ||
| }; | ||
| { | ||
| {0xAB70FE17C79AC6CA, -1060, -300}, |
There was a problem hiding this comment.
Another questionable change. I find the whitespace inside the braces is helpful. Interesting that it did multi-line column alignment of comments on the includes and removed column alignment of the parameters here.
There was a problem hiding this comment.
I've seen it, but I have not found the config key to fix this yet.
There was a problem hiding this comment.
Cpp11BracedListStyle affects the whitespace.
gregmarr
left a comment
gregmarr
left a comment
There was a problem hiding this comment.
There are some interesting changes. I wonder if the result would be different if you went back to the astyle code and formatted again with the current options rather than incremental changes with different options.
| JSON_PRIVATE_UNLESS_TESTED: | ||
| // convenience aliases for types residing in namespace detail; | ||
| using lexer = ::nlohmann::detail::lexer_base<basic_json>; | ||
| JSON_PRIVATE_UNLESS_TESTED : |
There was a problem hiding this comment.
You can tell clang-format what this macro means, so it knows it's an access specifier: https://clang.llvm.org/docs/ClangFormatStyleOptions.html#macros
There was a problem hiding this comment.
Interesting! I'll give it a try.
| NLOHMANN_JSON_ABI_TAGS, \ | ||
| NLOHMANN_JSON_NAMESPACE_VERSION) \ | ||
| { | ||
| #define NLOHMANN_JSON_NAMESPACE_BEGIN \ |
There was a problem hiding this comment.
Looks like AlignEscapedNewlines: Right aligns to ColumnLimit if it's non-zero. That's unfortunate.
🔴 Amalgamation check failed! 🔴The source code has not been amalgamated. @nlohmann |
🔴 Amalgamation check failed! 🔴The source code has not been amalgamated. @nlohmann |
|
I call it a day. There seems to be some work needed to move the Furthermore, there seems to be the same version nightmare as with astyle. I am using Clang-Format 17.0.6 locally and Clang-Format 18.0.0 in the CI, and already the results are not the same. Not sure how to make it easy for contributors to use the "right" version. Finally, Clang-Format is incredibly slow. Like 7 minutes to format I'll pick this up next week. |
This is to be expected unforunately. Due to the complexity the devs decided not to be backwards compatible for every edge case.
clang-format 17.03 takes here ~19minutes locally. |
4 participants
This PR replaces Artistic Style by Clang-Format as tool to indent the code. Astyle introduced a breaking change a while ago which made working with the tool very cumbersome.