I think the "vsi_search_string be an array" problem could be solved as following steps:
  1. An item such as (vsi_support_key_string="{supported forwarding mode: (0x40) reflective relay, supported capabilities: (0x7) RTE ECP VDP}";) is configured well in the .conf file.
  2. The application read this value from .conf file as a long string.
  3. The application executes supporting function about VSI and check whether the output returned is a substring of vsi_support_key_string. If the output contains in the vsi_support_key_string, supporting of VSI is proved.

Sincerely yours,
  Xu Wang



>On 04/09/2013 06:40 AM, Wenchao Xia wrote:
>> Hi, John
>> Sorry to interrupt but still need you to review this patch, which
>>make it work more flex on different version.
>> By the way, I got only 3 fail and they seems small problems in test
>> suit's code, do you think it is ready to release 0.6.2(with this patch
>> pushed)?
>> I don't have 'lldptool' installed on my system, so I cannot verify the output or the fix.
>
>What concerns me though is the reliance on something that could (and does) change between versions. >What you have is just "two" versions of differences. What if there's a 3rd version that had/has >something different? Is the output different on different archs/OS's?
>
>Is there no way to make the vsi_search_string be an array in the configuration file? Rather than two >constant strings.
>
>I suppose what you have works, but it seems there has to be a better way to do this. I also think the >comments in the .conf file could be beefed up so someone would know what to look for. Is that a >specific field in the output or is that string what is returned. Essentially how would someone know >what to look for in order to add a new/different string.
>
>Call this a "luke-warm" ACK. It works, but I think it's a shortcut.
>
>John
>
>
>
>>> In RH6.4 lldptool query command and output changed a bit,
......