devinfo(3) problem...

[Daniel Rudy]

At about the time of 10/26/2005 11:25 AM, John Baldwin stated the following:
> On Wednesday 26 October 2005 01:17 pm, Daniel Rudy wrote:
> 
>>At about the time of 10/25/2005 12:50 PM, victor cruceru stated the
>>
>>following:
>>
>>>Hi Daniel,
>>>Yes, this is the fix. And yes, I think that the bug was reported (at
>>>least on a mailing list...)
>>>BTW: I dare to suggest to fully upgrade your system to the "latest"
>>>6.0, it is a huge step forward from 5.4. Before doing this you may
>>>want to give it a try by booting from a CD and check  that your hw is
>>>fully functional (and detected).
>>
>>For stability reasons, I don't run the absolute latest software.  When
>>6.x comes out, I'll wait awhile for the bugs to be worked out before I
>>migrate to it.  As for hardware, I tend to run FreeBSD on older
>>hardware, so I don't think that there will be a problem with the
>>hardware being properly detected.
> 
> 
> FYI, I merged that fix back to RELENG_5 a while ago, so it will be fixed in 
> 5.5 as well as 6.0.

Oh good, now I don't have to keep track of it when I go to 5.5 then.
I'll still have to keep an eye on it for the security releases through
the remainder of 5.4 though.

>>>To check if it is your mistake or another bug in libdevinfo, you may
>>>want to run the associated tool (man -k devinfo).
>>>Hope this helps.
>>
>>Actually, the devinfo tool was working the whole time.  I tried the
>>devinfo tool before I decided to use the libdevinfo.so library.  I've
>>tested it with my code and it is working now with that one line of
>>additional code.
> 
> 
> Yes, devinfo(1) only traverses the tree once, so it didn't run into this bug.
> 

I analyzed the source code for devinfo(8) and used as an example of how
to use the devinfo(3) library.  So I knew it transversed the device tree
once, and my code does it twice, first time to locate (is it there?),
the second time to actually extract the data and use it.

-- 
Daniel Rudy