aboutsummaryrefslogtreecommitdiff
path: root/libjava/classpath
diff options
context:
space:
mode:
authorEric Botcazou <ebotcazou@adacore.com>2025-08-05 09:14:44 +0200
committerMarc Poulhiès <dkm@gcc.gnu.org>2025-09-11 11:10:50 +0200
commit88a389ac8664360aaf8b2a628df8da24b678daa8 (patch)
tree2c21b8a355be2fa31fc81fe574cbe10baa7e2f4c /libjava/classpath
parentc93fe0be50bcf622181b3036c89ebda0a71132d7 (diff)
downloadgcc-88a389ac8664360aaf8b2a628df8da24b678daa8.zip
gcc-88a389ac8664360aaf8b2a628df8da24b678daa8.tar.gz
gcc-88a389ac8664360aaf8b2a628df8da24b678daa8.tar.bz2
ada: Get rid of TYPE_ALIGN_OK flag in gcc-interface
The TYPE_ALIGN_OK flag had originally been a GCC flag tested in the RTL expander and was at some point kicked out of the middle-end to become a pure Gigi flag. But it's only set for tagged types and CW-equivalent types and can be replaced by a explicit predicate without too much work. gcc/ada/ChangeLog: * gcc-interface/ada-tree.h (TYPE_ALIGN_OK): Delete. * gcc-interface/decl.cc (gnat_to_gnu_entity): Do not set it. * gcc-interface/gigi.h (standard_datatypes): Add ADT_tag_name_id. (tag_name_id): New macro. (type_is_tagged_or_cw_equivalent): New inline predicate. * gcc-interface/trans.cc (gigi): Initialize tag_name_id. (gnat_to_gnu) <N_Unchecked_Type_Conversion>: Replace tests on TYPE_ALIGN_OK with calls to type_is_tagged_or_cw_equivalent. (addressable_p): Likewise. * gcc-interface/utils.cc (convert): Likewise. * gcc-interface/utils2.cc (build_binary_op): Likewise.
Diffstat (limited to 'libjava/classpath')
0 files changed, 0 insertions, 0 deletions