> >> And I just realized after submitting that all errors are ignored >> during all the steps... I think the logic I suggested still applies, >> though, to keep clear who owns the 'retries' object. >> >> But now, shouldn't we fail with an error if e.g. none of the retries >> checks succeeds? > > My idea here is to avoid skipping one or more checks of the retries only > because the previous one failed, this is why I used only logging messages to > signal the failures (and this is why I used mm_err, in place of mm_warn), > but yes, if none of the checks succeeds, return with a failure would be the > right thing to do. >
Better mm_warn(); mm_err() is actually used only in 2 single places to flag programmer errors, not logic errors. > I would put a flag in the Context struct to take trace that at least one > check succeeded and at the last step (PUK2) I would check this flag and > decide whether return with failure or not. What do you think about that? Given that you already have the LOAD_UNLOCK_RETRIES_STEP_LAST step, you could do the final error check there. If none of the checks succeeded, you could then g_simple_async_result_set_error() instead of set_op_res_gpointer(). -- Aleksander https://aleksander.es _______________________________________________ ModemManager-devel mailing list ModemManager-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/modemmanager-devel