Hi!
Gerry> In the following example mib , it seems as though the Gerry> parser is ignoring the "IMPLIED" keyword. I can not find any Gerry> reference to it in the .xsd output. Any thoughts?
Good point. Torsten, how do you think about adding an according construct to such entries' <appinfo> elements?
Torsten> Good idea. I have added an <appinfo> element called <implied> Torsten> that has the value <implied>true</implied> if the IMPLIED Torsten> keyword is present and <implied>false</implied> otherwise.
Thanks for the quick update. I spent a look on the currently generated output and more thoughts on this kind of flag information (and I'm sorry for not doing this in the first place before you spent work on it).
First, I propose to supply this new information only on the elements that represent a row type, i.e. not for columnar or other scalar elements.
Furthermore, there is another kind of flag information that is still missing (which I did not recognize earlier): we do not yet signal in the XSD output whether writable columnar elements (i.e., those with <maxAccess>read-write</maxAccess>) are just writeable or also creatable. Hence, I suggest to denote these two kinds of flags together as <flags [implied="yes|no"] [create="yes|no"]/>, where the default "no" can be omitted and if both flag values are "no" the whole <flags> element can be omitted. I think this can be mapped quite straight forward from smi.h::SmiNode.
-frank