Tracking issue for migrating/deleting old codegen tests

[vext01]

We recently deleted the old LLVM JIT backend. We renamed all non-new-codegen C tests to *.old.c and they are currently ignored.

Some of those tests need porting, some can be deleted. This issue tracks this progress.

Need to be ported

• arg_mapping_callee.old.c
• awkward_unmappable.old.c
• bf.old.c
• call_args.old.c
• call_ext_in_obj.old.c
• call_ext_simple.old.c
• calls_double.old.c
• choice.old.c
• conditionals.old.c
• constexpr.old.c
• const_global.old.c
• fib.old.c
• funcptrarg_hasir.old.c (crashing, ignored)
• funcptrarg_noir.old.c
• indirect_branch.old.c (crashing, ignored)
• indirect_external_function_call.old.c
• internal_linkage_same_obj.old.c
• intrinsic_noinline.old.c (crashing, ignored)
• intrinsics.old.c
• loopy_funcs_not_inlined_by_default.old.c
• many_threads_many_locs.old.c
• many_threads_one_loc.old.c
• mutable_global.old.c
• nested_sidetrace.old.c
• nested_writetoptr.old.c
• noopts.old.c
• not_loopy_funcs_inlined_by_default.old.c
• no_trace_annotation2.old.c
• no_trace_annotation.old.c
• promote_expr.old.c
• ptr_global.old.c
• pt_zero_len_call.old.c
• qsort.old.c
• reentrant.old.c
• rel_path.old.c
• resume_and_branch.old.c
• side-trace.old.c
• sidetrace_phinode.old.c
• sidetrace_while.old.c
• simple_nested.old.c
• simple_non_serialised.old.c
• smmultisrc2.old.c
• smmultisrc.old.c
• stats1.old.c
• stats2.old.c
• stats3.old.c
• stats4.old.c
• strarray.old.c
• struct_simple.old.c
• trace_too_long_hwt.old.c
• trace_too_long.old.c
• unmapped_setjmp.old.c
• unroll_safe_implies_noinline.old.c
• unroll_safe_inlines.old.c
• varargs_inlined.old.c
• varargs.old.c
• void_ret.old.c
• yk_unroll_safe_vs_yk_outline.old.c

Duplicate tests that can be removed.

• debuginfo.old.c (dup of aot_debuginfo.c)
• direct_recursion.old.c (dup of outline_recursion.c)
• mutual_recursion.old.c (dup of outline_recursion_indirect.c)
• inline_asm_call.old.c (dup of inline_asm.c)

Irrelevant tests that can be deleted.

• interp_inside_call2.old.c (designed as blackholing test)
• interp_inside_call.old.c (designed as blackholing test)
• outlinebase_bug.old.c (tests an old outlining mechanism)
• promote_not_great.old.c (tests limitation of an old promote mechanism)
• lifetime_intrinsics.old.c

[vext01]

There's a lot of work to be done here. Is it worth spreading the load?

[vext01]

OK, I've migrated all the tests apart from the side-tracing tests.

I'll reassign this to @ptersilie for when he revisists side-tracing.