Hi! The r17-1009 change (in particular gfc_match_omp_variable_list hunk of it) fixed spelling of one diagnostic message, but has not adjusted corresponding testsuite, so these tests now fail.
Fixed thusly, tested on x86_64-linux, committed to trunk as obvious. 2026-06-02 Jakub Jelinek <[email protected]> * gfortran.dg/gomp/allocate-6.f90: Adjust dg-error wording after recent spelling fixes. * gfortran.dg/gomp/allocate-7.f90: Likewise. * gfortran.dg/gomp/allocate-14.f90: Likewise. * gcc.target/sparc/struct-ret-check-1.c: Fix comment spelling, specificed -> specified. --- gcc/testsuite/gfortran.dg/gomp/allocate-6.f90.jj 2026-03-27 10:17:20.579225568 +0100 +++ gcc/testsuite/gfortran.dg/gomp/allocate-6.f90 2026-06-02 08:24:36.891391166 +0200 @@ -68,7 +68,7 @@ common /com2/ y,z integer, allocatable :: alloc integer, pointer :: ptr -!$omp allocate(q) ! { dg-error "'q' at .1. is part of the common block '/com1/' and may only be specificed implicitly via the named common block" } +!$omp allocate(q) ! { dg-error "'q' at .1. is part of the common block '/com1/' and may only be specified implicitly via the named common block" } !$omp allocate(d(:)) ! { dg-error "Unexpected expression as list item at .1. in ALLOCATE directive" } !$omp allocate(a) align(4), align(4) ! { dg-error "Duplicated 'align' clause" } --- gcc/testsuite/gfortran.dg/gomp/allocate-7.f90.jj 2026-03-27 10:17:20.579225568 +0100 +++ gcc/testsuite/gfortran.dg/gomp/allocate-7.f90 2026-06-02 08:24:45.957268434 +0200 @@ -52,7 +52,7 @@ integer, parameter :: prm=5 !$omp allocate(used) allocator(omp_pteam_mem_alloc) ! { dg-error "Argument 'used' at .1. to declarative !.OMP ALLOCATE shall be in the same scope as the variable declaration" } !$omp allocate(n) allocator(omp_pteam_mem_alloc) ! { dg-error "Unexpected dummy argument 'n' as argument at .1. to declarative !.OMP ALLOCATE" } -!$omp allocate (x) align(128) ! { dg-error "'x' at .1. is part of the common block '/com1/' and may only be specificed implicitly via the named common block" } +!$omp allocate (x) align(128) ! { dg-error "'x' at .1. is part of the common block '/com1/' and may only be specified implicitly via the named common block" } !$omp allocate (a, b, a) allocator (omp_pteam_mem_alloc) ! { dg-error "Duplicated variable 'a' in !.OMP ALLOCATE" } contains --- gcc/testsuite/gfortran.dg/gomp/allocate-14.f90.jj 2026-03-27 10:17:20.578225585 +0100 +++ gcc/testsuite/gfortran.dg/gomp/allocate-14.f90 2026-06-02 08:24:30.819473365 +0200 @@ -79,7 +79,7 @@ subroutine common use m integer :: a,b,c(5) common /my/ a,b,c - !$omp allocate(b) allocator(omp_cgroup_mem_alloc) ! { dg-error "'b' at .1. is part of the common block '/my/' and may only be specificed implicitly via the named common block" } + !$omp allocate(b) allocator(omp_cgroup_mem_alloc) ! { dg-error "'b' at .1. is part of the common block '/my/' and may only be specified implicitly via the named common block" } end subroutine c_and_func_ptrs --- gcc/testsuite/gcc.target/sparc/struct-ret-check-1.c.jj 2026-03-27 10:17:19.828237823 +0100 +++ gcc/testsuite/gcc.target/sparc/struct-ret-check-1.c 2026-06-02 08:24:14.712691414 +0200 @@ -3,7 +3,7 @@ /* Test that GCC follows the SPARC 32-bit psABI with regards to structure return checking in a callee. When -mstd-struct-return - is specificed then gcc will emit code to skip the unimp insn. */ + is specified then gcc will emit code to skip the unimp insn. */ /* Origin: Carlos O'Donell <[email protected]> */ /* { dg-do run { target sparc*-*-solaris* sparc*-*-linux* sparc*-*-*bsd* } } */ Jakub
