On 4/6/23 02:27, Iain Sandoe wrote:
I spotted this problem while trying to collate my WIP patches (sorry
for not spotting it when the original patch was discussed).
Although we do not yet have any regression filed against this (AFAIK),
it's only a matter of time - so I'd suggest that we should apply to
13.
Tested on x86_64-darwin21 and x86_64_linux-gnu.
OK for trunk? (when?)
Iain
--- 8< ---
When the function contains no local vars and also no nested scopes, there
is no top-level bind expression. Because the rewritten coroutine body will
require both local vars and contain nested scopes, we add a bind expression
to such functions. When this was done the necessary scope blocks were
omitted which leads to disconnected function content.
Fixed by adding a new block to the added bind expression.
Signed-off-by: Iain Sandoe <i...@sandoe.co.uk>
gcc/cp/ChangeLog:
* coroutines.cc (coro_rewrite_function_body): Ensure that added
bind expressions have scope blocks.
---
gcc/cp/coroutines.cc | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/gcc/cp/coroutines.cc b/gcc/cp/coroutines.cc
index b307c8ca8b6..59a240ebd40 100644
--- a/gcc/cp/coroutines.cc
+++ b/gcc/cp/coroutines.cc
@@ -4113,6 +4113,10 @@ coro_rewrite_function_body (location_t fn_start, tree
fnbody, tree orig,
tree bind_wrap = build3_loc (fn_start, BIND_EXPR, void_type_node,
NULL, NULL, NULL);
BIND_EXPR_BODY (bind_wrap) = fnbody;
+ /* Ensure we have a block to connect up the scopes. */
+ tree new_blk = make_node (BLOCK);
+ BIND_EXPR_BLOCK (bind_wrap) = new_blk;
+ BLOCK_SUBBLOCKS (top_block) = new_blk;
fnbody = bind_wrap;
}
I wonder if you want to use c_build_bind_expr for various BIND_EXPR
creation? You'll still need to explicitly build a BLOCK in this
particular case, as in poplevel with nonzero functionbody, but that
might reduce the boilerplate in general.
The patch is OK.
Jason
