GCC documentation now clarifies integer types as “integral”.
The GCC documentation now uses “integral” instead of the potentially confusing “fixed-point” when referring to integer types.
The GCC documentation is updated to use the term “integral” instead of “fixed-point” for integer types like int or uint32_t. This change avoids confusion with fractional fixed-point numbers. The terminology now accurately reflects that these are whole number types.
In Details
The GCC documentation, specifically doc/rtl.texi, is updated to replace instances of “fixed-point” with “integral” when describing fundamental integer types. This clarifies the distinction between plain integers and actual fixed-point fractional number representations, which is a common source of ambiguity for those working with low-level type systems. Even toolchain developers might mistakenly conflate the two without this precise clarification.
For Context
Compilers like GCC translate human-readable code into machine instructions. To do this, they need a clear way to describe the different kinds of data they encounter, such as numbers. Sometimes, the documentation might use terms that can be misinterpreted. In this case, the term “fixed-point” was used to describe whole numbers (integers), which are numbers without any fractional part. However, “fixed-point” also refers to a specific way of representing numbers with both whole and fractional parts, often used in embedded systems or financial calculations. To prevent any confusion between plain integers and these fractional fixed-point numbers, the GCC documentation now consistently uses the term “integral” to refer exclusively to whole number types.