Hi Paul, On Sat, 28 Jun 2025 14:32:13 +0200, Paul Gevers <[email protected]> wrote: > On 28-06-2025 11:57, Stephen Kitt wrote: > > The fix involves a small number of lines, and while it’s not simple, > > it makes sense to me. wine is a leaf package with no alternative. > > Did I spot it right that the first patch is undone by the second, even > though the patch offset would be huge in one of the two cases? I assume > you used the first patch to not have to amend the second? Might have > been worth mentioning as the amended patch would have been less review.
Yes, that’s correct, and it would indeed have been simpler:
diff --git i/dlls/ntdll/unix/virtual.c w/dlls/ntdll/unix/virtual.c
index 0c0a12c394d..947cf3d60f9 100644
--- i/dlls/ntdll/unix/virtual.c
+++ w/dlls/ntdll/unix/virtual.c
@@ -3115,10 +3115,11 @@ static NTSTATUS virtual_map_image( HANDLE mapping, void
**addr_ptr, SIZE_T *size
status = map_image_into_view( view, filename, unix_fd, image_info,
machine, shared_fd, needs_close );
if (status == STATUS_SUCCESS)
{
+ image_info->base = wine_server_client_ptr( view->base );
SERVER_START_REQ( map_image_view )
{
req->mapping = wine_server_obj_handle( mapping );
- req->base = wine_server_client_ptr( view->base );
+ req->base = image_info->base;
req->size = size;
req->entry = image_info->entry_point;
req->machine = image_info->machine;
I kept the two separate patches so they’d match the upstream commits.
Thanks,
Stephen
pgpJfqmHlaiGe.pgp
Description: OpenPGP digital signature

