Add missing format for built-in floating-point types
authorUlrich Weigand <ulrich.weigand@de.ibm.com>
Tue, 6 Sep 2016 15:31:03 +0000 (17:31 +0200)
committerUlrich Weigand <ulrich.weigand@de.ibm.com>
Tue, 6 Sep 2016 15:31:03 +0000 (17:31 +0200)
commit49f190bcb7f074ea2e27d4e967e4fae9ed7dafb6
treeee2cd587d9bfd4df9156a6c70cbd60aae704317d
parentc413c44801e449f1f0b9828b81770e752b8219af
Add missing format for built-in floating-point types

Many callers of init_float_type and arch_float_type still pass a NULL
floatformat.  This commit changes those callers where the floatformat
that is supposed to be use is obvious.  There are two categories where
this is the case:

- A number of built-in types are intended to match the platform ABI
  floating-point types (i.e. types that use gdbarch_float_bit etc.).
  Those places should use the platform ABI floating-point formats
  defined via gdbarch_float_format etc.

- A number of language built-in types should simply use IEEE floating-
  point formats, since the language actually defines that this is the
  format that must be used to implement floating-point types for this
  language.  (This affects Java, Go, and Rust.)  The same applies for
  to the predefined "RS/6000" stabs floating-point built-in types.

gdb/ChangeLog:

* ada-lang.c (ada_language_arch_info): Use gdbarch-provided
platform ABI floating-point formats for built-in types.
* d-lang.c (build_d_types): Likewise.
* f-lang.c (build_fortran_types): Likewise.
* m2-lang.c (build_m2_types): Likewise.
* mdebugread.c (basic_type): Likewise.

* go-lang.c (build_go_types): Use IEEE floating-point formats
for language built-in types as mandanted by the language.
* jv-lang.c (build_java_types): Likewise.
* rust-lang.c (rust_language_arch_info): Likewise.
* stabsread.c (rs6000_builtin_type): Likewise.

Signed-off-by: Ulrich Weigand <ulrich.weigand@de.ibm.com>
gdb/ChangeLog
gdb/ada-lang.c
gdb/d-lang.c
gdb/f-lang.c
gdb/go-lang.c
gdb/jv-lang.c
gdb/m2-lang.c
gdb/mdebugread.c
gdb/rust-lang.c
gdb/stabsread.c