Details
-
Bug
-
Status: Resolved
-
Major
-
Resolution: Fixed
-
None
Description
This is with gcc 7.3.0 and -Wconversion.
../src/arrow/compute/kernels/cast.cc: In instantiation of ‘void arrow::compute::CastFunctor<O, I, typename std::enable_if<arrow::compute::is_float_truncate<O, I>::value>::type>::operator()(arrow::compute::FunctionContext*, const arrow::compute::CastOptions&, const arrow::ArrayData&, arrow::ArrayData*) [with O = arrow::Int64Type; I = arrow::DoubleType; typename std::enable_if<arrow::compute::is_float_truncate<O, I>::value>::type = void]’: ../src/arrow/compute/kernels/cast.cc:1105:1: required from here ../src/arrow/compute/kernels/cast.cc:291:45: warning: conversion to ‘in_type {aka double}’ from ‘long int’ may alter its value [-Wconversion] if (ARROW_PREDICT_FALSE(out_value != *in_data)) { ~~~~~~~~~~^~~~~ ../src/arrow/util/macros.h:37:50: note: in definition of macro ‘ARROW_PREDICT_FALSE’ #define ARROW_PREDICT_FALSE(x) (__builtin_expect(x, 0)) ^ ../src/arrow/compute/kernels/cast.cc:301:45: warning: conversion to ‘in_type {aka double}’ from ‘long int’ may alter its value [-Wconversion] if (ARROW_PREDICT_FALSE(out_value != *in_data)) { ~~~~~~~~~~^~~~~ ../src/arrow/util/macros.h:37:50: note: in definition of macro ‘ARROW_PREDICT_FALSE’ #define ARROW_PREDICT_FALSE(x) (__builtin_expect(x, 0))
Attachments
Issue Links
- links to