|
|
|
|
@ -26,25 +26,25 @@ License: LGPLv2+
|
|
|
|
|
|
|
|
|
|
# Fix paths reported by the libIDL-config-2 tool to conform with Fedora
|
|
|
|
|
# multilib installation paths:
|
|
|
|
|
Patch0: libIDL-0.8.6-multilib.patch
|
|
|
|
|
Patch: libIDL-0.8.6-multilib.patch
|
|
|
|
|
# Remove an unused parent-node variable in the primary_expr part of the parser,
|
|
|
|
|
# which caused a compiler warning.
|
|
|
|
|
Patch1: libIDL-0.8.14-parser-primary_expr-unused-parent-node.patch
|
|
|
|
|
Patch: libIDL-0.8.14-parser-primary_expr-unused-parent-node.patch
|
|
|
|
|
# On platforms (such as 64-bit Linux), where long long int and long int are
|
|
|
|
|
# both 64-bit, we can have IDL_LL defined to ll (format with %%lld) while
|
|
|
|
|
# IDL_longlong_t, which is just gint64, may be ultimately defined to long int.
|
|
|
|
|
# This results in compiler warnings about the mismatch between the long long
|
|
|
|
|
# format and long parameter, even though the types are compatible. We can fix
|
|
|
|
|
# this with a cast to (long long) before formatting.
|
|
|
|
|
Patch2: libIDL-0.8.14-long-long-format-warnings.patch
|
|
|
|
|
Patch: libIDL-0.8.14-long-long-format-warnings.patch
|
|
|
|
|
# Instead of type-punning with sscanf, parse into a temporary with a type
|
|
|
|
|
# matching the format code and then memmove into the “integer” storage. This is
|
|
|
|
|
# no less platform-dependent, but does not invoke undefined behavior or produce
|
|
|
|
|
# a compiler warning.
|
|
|
|
|
Patch3: libIDL-0.8.14-lexer-sscanf-type-punning.patch
|
|
|
|
|
Patch: libIDL-0.8.14-lexer-sscanf-type-punning.patch
|
|
|
|
|
# Fix references to the old libIDL-config script by changing them to
|
|
|
|
|
# libIDL-config-2.
|
|
|
|
|
Patch4: libIDL-0.8.14-old-libIDL-config-script.patch
|
|
|
|
|
Patch: libIDL-0.8.14-old-libIDL-config-script.patch
|
|
|
|
|
|
|
|
|
|
BuildRequires: gcc
|
|
|
|
|
BuildRequires: make
|
|
|
|
|
|
Benjamin A. Beasley
3 years ago