zhekov commented on this pull request.
@@ -967,6 +999,23 @@ static gboolean spawn_read_cb(GIOChannel *channel, GIOCondition condition, gpoin
sc->cb.read(buffer, input_cond | failure_cond, sc->cb_data); } + /* Check for continuous activations with G_IO_IN | G_IO_PRI, without any + data to read and without errors. If detected, switch to timeout source. */ + else if (sc->empty_gio_ins < MAX_EMPTY_GIO_INS && status == G_IO_STATUS_AGAIN)
"status" can be different from G_IO_STATUS_NORMAL only if we attempted to read. This means a (G_IO_IN | G_IO_PRI) condition, possibly combined with G_IO_FAILURE condition. We filter out the failures in the "if" clause, so a G_IO_STATUS_AGAIN means only (G_IO_IN | G_IO_PRI) condition at this point. But if there is no failure, spawn_read_cb() reads only once, for reasons explained in the source, and thus AGAIN is not a result of a subsequent read.
In summary, we received a G_IO_IN | G_IO_PRI, attempted to read, and immediately got G_IO_STATUS_AGAIN, which shoudn't happen...