Hi Denis,
I rethink about my second question (parse_dataobj() shouldn't eat the
cake that not belongs to it) on IRC, and may find a case to defense my point.
The example is ber-tlv launch browser. The similarity to setup_menu is
that they all contain a list, while the difference is the list in lanuch
browser is optional. Below are the objects within launch browser: browser
id(M), URL(O), bearer(O),provisioning file 1(O), provisioning file
2(O),...,provisioning file N(O), text string(O), ...
Then we may call parse_dataobj(browser id, URL, bearer) to parse the
objects before provisioning files, then use parse_provisioning_file_list() to
parse provisioning files, and at last call parse_dataobj() again to parse the
rest. If there is no provisioning file at all, the situation is same to call
parse_dataobj() one after another. If so, we may read object text string in
first parse_dataobj(), and at the beginning of second parse_dataobj(), we would
move forward again, which could be wrong.
Thus I think the solution is parse_dataobj() needs to restore the iter
it can't consume. What's your opinion?
Regards,
-Yang
_______________________________________________
ofono mailing list
[email protected]
http://lists.ofono.org/listinfo/ofono
