Luotao Fu wrote:
> Hi Oliver,

>> Nice idea! Especially for some kind of *integrated* (GUI-based) applications.
>>
> 
> That's what I had in mind as I started this.


8-)


>>>>>> int canhw_get_ctrlmode(const char *name, struct can_ctrlmode *cm);
>> This leads also to 'libcanhw' as library name.
>>
>> "SocketCAN" and "Netlink" is just state-of-the-art for CAN hardware in Linux.
>> No need to put it into the library name ...
>>
> 
> interesting idea. I'll think about this.


Fine.


>> IMHO there's also no real benefit to put this lib together with the helpers 
>> in
>>
>>    trunk/can-utils/lib.[ch]
>>
>> These are addressing different topics.
>>
> 
> I personally don't have opinions on this. The canconfig tool is partly
> supposed to demonstrate the usage of the library too. Intergrate the lib
> into the canutils package would work the same, as long as the library is
> build as dyn. so and can be loaded separatedly.


Yep. That's right.


>>> Also, keep it simple having in mind that it's especially useful on
>>> embedded system (the busybox's "ip" utility does not yet support CAN).
>>
>> Which should be fixed inside busybox :-)
>> Fortunately ptxdist allows the iproute2 package to be build instead of using
>> the built-in 'ip' from busybox.
>>
>>
>> Regards,
>> Oliver
>>
>> ps. Did you recognize the updated can_if script in
>>
>> http://svn.berlios.de/svnroot/repos/socketcan/trunk/etc/can_if
>>
>> that can deal with the new 'ip' tool also?
>>
>> # The syntax for the CAN devices is: devna...@bitrate][,restart-ms]
>> # example CAN_IF="can0 c...@500000 c...@500000,200 can3,500"
>>
>> Just another idea.
>>
> 
> Yeah I know the script. Indeed I looked into it every time I was
> wondering for how the ip parameter should look like. ;-) The script is
> nice to bring up/turn down or reload the interfaces, but the
> possibilities are limited,Besides there're some natural limitations in
> the iproute tool, as I mentioned in the last mail, and not at least the
> big footprint and complex infrastructure. So I think canconfig comes in
> quite some occassions more handy than iproute + script.


To be more handy is indeed a good point. Having canconfig in the embedded
environment and to have a expressive example for the lib is a good idea.

But having a nice handy canconfig probably makes it harder to maintain the
iproute2 tool for Wolfgang ;-))

Thanks,
Oliver

_______________________________________________
Socketcan-users mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/socketcan-users

Reply via email to