On 16/4/25 10:28, Daniel P. Berrangé wrote:
On Wed, Apr 16, 2025 at 05:14:06PM +0900, Kohei Tokunaga wrote:
On emscripten, function pointer casts can cause function call failure.
This commit fixes the function definition to match to the type of the
function call using g_list_sort_with_data.
Signed-off-by: Kohei Tokunaga <ktokunaga.m...@gmail.com>
---
hw/core/loader.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/hw/core/loader.c b/hw/core/loader.c
index 2e35f0aa90..93a8b45d28 100644
--- a/hw/core/loader.c
+++ b/hw/core/loader.c
@@ -1410,7 +1410,7 @@ typedef struct RomSec {
* work, but this way saves a little work later by avoiding
* dealing with "gaps" of 0 length.
*/
-static gint sort_secs(gconstpointer a, gconstpointer b)
+static gint sort_secs(gconstpointer a, gconstpointer b, gpointer d)
{
RomSec *ra = (RomSec *) a;
RomSec *rb = (RomSec *) b;
@@ -1463,7 +1463,7 @@ RomGap rom_find_largest_gap_between(hwaddr base, size_t
size)
/* sentinel */
secs = add_romsec_to_list(secs, base + size, 1);
- secs = g_list_sort(secs, sort_secs);
+ secs = g_list_sort_with_data(secs, sort_secs, NULL);
I don't see what the problem is with the original code.
IIUC the rationale is in patch #10 where Kohei poisons g_list_sort():
On emscripten, function pointer casts can cause function call
failure. g_list_sort and g_slist_sort performs this internally
so can't be used on Emscripten. Instead, g_list_sort_with_data
and g_slist_sort_with_data should be used.
The commit message says we have a bad function cast, but the original
method decl is
GList *g_list_sort(GList*list, GCompareFunc compare_func);
where the callback is
typedef gint (*GCompareFunc)(gconstpointer a, gconstpointer b);
Our code complies with that GCompareFunc signature.
For comparison the new code uses:
GList *g_list_sort_with_data(GList *list, GCompareDataFunc compare_func,
gpointer user_data);
where the callback is
typedef gint (*GCompareDataFunc)(gconstpointer a, gconstpointer b, gpointer
user_data);
which the new code complies with, but it is undesirable since we
have no use for the data parameter.
With regards,
Daniel
