On Mon, Jul 17, 2017 at 10:02 AM, Aleksander Morgado <aleksan...@aleksander.es> wrote: >> What you included here as a >> snippet is a huge bug actually. > > Well, not that huge really... the only issue is that we're setting the > GSimpleAsyncResult result twice to TRUE. But, of course, if the > parent's implementation isn't using a GSimpleAsyncResult and is using > a GTask instead (as in your case), there really would be a bug as a > GTask isn't a GSimpleAsyncResult :) So, let me fix as many of those > issues as I find in a quick look, and then publish your full branch > and we'll get it reviewed and merged.
I did the following grep in the source code to find all places where we were casting explicitly to G_SIMPLE_ASYNC_RESULT when calling g_simple_async_result_set_op...(): $ grep -r "g_simple_async_result_set_op" . | grep "G_SIMPLE_ASYNC" And fixed all places reported. The rationale is that we usually explicitly use the GSimpleAsyncResult type instead of "gpointer" in the user_data passed to the async functions inside the async function we're implementing, so there usually is never the need to explicitly cast that to GSimpleAsyncResult again to call g_simple_async_result_set_op...(). Where there was an explicit cast, the code was buggy. Please rebase your branch on top of master, include the missing commits you took out, and repush for review. -- Aleksander https://aleksander.es _______________________________________________ ModemManager-devel mailing list ModemManager-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/modemmanager-devel
