Subject:
|
Re: bug in setting lnp host address?
|
Newsgroups:
|
lugnet.robotics.rcx.legos
|
Date:
|
Wed, 5 Jun 2002 17:51:18 GMT
|
Viewed:
|
2132 times
|
| |
|
|
And, you'll probably hate my reply...
"Just because you can do something, doesn't mean you should."
The very idea of a discontinuous array of bits representing a network
address in something as rudimentary as LNP makes my hair stand on end. The
idea here was to be simple to use and flexible in implementation. All I did
was make the legOS API and dll command-line parameters a little less
geek-like and more user friendly. The shifting you object to is not part of
any critical execution path, its only when the host address is stored or in
the command-line decoding within dll.
In case you haven't noticed, the port handler dispatch arrary assumes
monotonic port numbers too.
As you stated and as I suspect, 16 nodes X 16 ports should hold up for this
generation of RCX and host based LNP implementations. Other combinations of
splitting the addressing byte are easily built and used without penalty.
peace, bro... /ted
"Albert Huang" <sfitw@hotmail.com> wrote in message
news:Gx7K30.862@lugnet.com...
> you'll probably hate me for being a nitpicker....
>
> > I will parameterize this shift-value based on the CONF_LNP_HOSTMASK
> > definition. Unless others object strenuously, I prefer to leave this
> > interface using small valued numbers.
>
>
> technically, you should not do any bitshifting at all. There is always • the
> odd case when someone decides to be clever and have a hostmask with
> interleaved zeros. for example, a hostmask of 10101010 (0xAA), while
> confusing, is still a valid hostmask. If that host is listening on port
> 01010101, and someone sends a message addressed to 11111111, then that • host
> should pick it up and direct the message to the right port. In this case,
> bitshifting will never work.
>
> This is also why I would like to see the addressing scheme finalized. I
> think it would make more sense to have a series of address-only bits
> followed by a series of port-only bits. Unless anybody wants to
> significantly change LNP, I think it would make sense to settle on 4 bits
> for the host and 4 for the port, as this is what most people are using • right
> now and is most likely adequate for the future, unless someone plans on
> throwing a big legOS LNP festival...
>
> cheers,
> albert
>
>
|
|
Message has 1 Reply:
 | | problem using emulegos
|
| hi all, im having problem using emulegos 1.2.5.0 with legOS 0.2.4 i have used legOS 0.2.4 for over 1 year now withouth any problems (including lnp), but for some reason when i type make in ~/hovedfag/emulegos/...ples/rover i get: makefile:8: (...) (22 years ago, 10-Jun-02, to lugnet.robotics.rcx.legos)
|
Message is in Reply To:
| | Re: bug in setting lnp host address?
|
| you'll probably hate me for being a nitpicker.... (...) technically, you should not do any bitshifting at all. There is always the odd case when someone decides to be clever and have a hostmask with interleaved zeros. for example, a hostmask of (...) (22 years ago, 5-Jun-02, to lugnet.robotics.rcx.legos)
|
6 Messages in This Thread:
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
This Message and its Replies on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
