Hi On Thu, Feb 7, 2019 at 3:37 PM Alberto Garcia <be...@igalia.com> wrote: > > On Thu 07 Feb 2019 03:24:40 PM CET, Marc-André Lureau wrote: > > On Thu, Feb 7, 2019 at 3:23 PM Marc-André Lureau > > <marcandre.lur...@redhat.com> wrote: > >> > >> On Thu, Feb 7, 2019 at 2:23 PM Alberto Garcia <be...@igalia.com> wrote: > >> > > >> > After g_source_attach() the GMainContext holds a reference to the > >> > GSource, so the caller does not need to keep it. > >> > > >> > pty_chr_state() and qio_task_thread_worker() are not doing this, so > >> > the GSource is being leaked in both cases (pty_chr_open_src_cancel() > >> > is the exception here because it does remove the extra reference > >> > correctly). > >> > > >> > Signed-off-by: Alberto Garcia <be...@igalia.com> > >> > >> You could mention this is a fix for regression from > >> a17536c594bfed94d05667b419f747b692f5fc7f > > > > and 938eb9e9c83d34d90cac6ec5c5388e7998840e4e > > Ok, feel free to add this to the commit message if I don't need to > resend the series.
Well, the series conflicts with the currently queued chardev series. I think I will send the pullreq and let you rebase. (I also have a related series pending review "[PATCH v2 0/6] chardev: various cleanups and improvements") -- Marc-André Lureau
