diff options
Diffstat (limited to 'gcc/ada/gnat_rm.texi')
| -rw-r--r-- | gcc/ada/gnat_rm.texi | 54 |
1 files changed, 54 insertions, 0 deletions
diff --git a/gcc/ada/gnat_rm.texi b/gcc/ada/gnat_rm.texi index b0e9f32..759ae5a 100644 --- a/gcc/ada/gnat_rm.texi +++ b/gcc/ada/gnat_rm.texi @@ -111,6 +111,7 @@ Implementation Defined Pragmas * Pragma Ast_Entry:: * Pragma C_Pass_By_Copy:: * Pragma Check:: +* Pragma Check_Float_Overflow:: * Pragma Check_Name:: * Pragma Check_Policy:: * Pragma Comment:: @@ -850,6 +851,7 @@ consideration, the use of these pragmas should be minimized. * Pragma Ast_Entry:: * Pragma C_Pass_By_Copy:: * Pragma Check:: +* Pragma Check_Float_Overflow:: * Pragma Check_Name:: * Pragma Check_Policy:: * Pragma Comment:: @@ -1402,6 +1404,58 @@ Checks introduced by this pragma are normally deactivated by default. They can be activated either by the command line option @option{-gnata}, which turns on all checks, or individually controlled using pragma @code{Check_Policy}. +@node Pragma Check_Float_Overflow +@unnumberedsec Pragma Check_Float_Overflow +@cindex Floating-point overflow +@findex Check_Float_Overflow +@noindent +Syntax: +@smallexample @c ada +pragma Check_Float_Overflow; +@end smallexample + +@noindent +In Ada, the predefined floating-point types (@code{Short_Float}, +@code{Float}, @code{Long_Float}, @code{Long_Long_Float}) are +defined as being unconstrained. This means that even though they +have well defined base ranges, there is no requirement that an +overflow exception be raised when the result of an operation is +outside this base range. This definition accomodates the notion +of infinities in IEEE floating-point, and corresponds to the +efficient execution mode on most machines. GNAT will not raise +overflow exceptions on these machines, instead it will generate +infinities and NaN's as defined in the IEEE standard. + +Although the generation of infinities is efficient, it is not +always desirable, and it is often the case that it would be +preferable to check for overflows, even if this resulted in +substantially less efficient code. This can be accomplished +by defining your own float subtypes, and indeed such types +can have the same base range as in: + +@smallexample @c ada +subtype My_Float is Float range Float'Range; +@end smallexample + +@noindent +In this example, @code{My_Float} has the same range as +@code{Float} but it is constrained, so operations on +@code{My_Float} values will be checked for overflow +against this range. + +However, it is often convenient to avoid the need to +define your own floating-point types, and instead use +the standard predefined types. The @code{Check_Float_Overflow} +configuration pragma achieves that. If a unit is compiled +subject to this configuration pragma, then all operations +on predefined floating-point types will be treated as +though those types were constrained and overflow checks +will be generated, resulting in a @code{Constraint_Error} +exception if the result is out of range. + +This mode can also be set by use of the compiler +switch @option{-gnateF}. + @node Pragma Check_Name @unnumberedsec Pragma Check_Name @cindex Defining check names |