Diff bus naming

On 09/22/2016 07:38 AM, Morten Leikvoll wrote:

 

Currently, bus names may contain a single end bus of signal names

ending with the index#, like D[0..7],A[0..15]

 

However, diff signals needs to end with _P or _N for eagle to detect

them as diff, hence they can not be added to a bus using the index

range. D[0..7]_P,D[0..7]_N doesnt work. This means, diff buses must be

written with very long names.

 

I suggest that bus names should at support one of the following,

ranging from simple to complex effort:

 

1-Index in the middle of the name, such as D[0..7]_P,D[0..7]_N

One could use DP[0..7],DN[0..7], or D_P[0..7],D_N[0..7].

 

2-D[0..7]_[PN] for even more compactness

 

3-Multiple indexes like D[A..C][0..7]_[PN], in good old regexp style.

I see your point, multiple iterators in any position rather than a

single iterator at the end.

Odd that you mention regexps.  Every time I type in a bus name index I

use a dash to specify a range, and have to go back and correct it to two

dots.  Creatures of habit, I guess.

Jorge, I'm in for this one.

On 22.09.2016 14:41, Chuck Huber wrote:

On 09/22/2016 07:38 AM, Morten Leikvoll wrote:

 

Currently, bus names may contain a single end bus of signal names

ending with the index#, like D[0..7],A[0..15]

 

However, diff signals needs to end with _P or _N for eagle to detect

them as diff, hence they can not be added to a bus using the index

range. D[0..7]_P,D[0..7]_N doesnt work. This means, diff buses must be

written with very long names.

 

I suggest that bus names should at support one of the following,

ranging from simple to complex effort:

 

1-Index in the middle of the name, such as D[0..7]_P,D[0..7]_N

 

One could use DP[0..7],DN[0..7], or D_P[0..7],D_N[0..7].

I guess you mean that the diff definition has to change instead of bus

naming definitions. DPn/DNn will not be routed as diff as it is now.

Yes thats an option, to allow something after the P or N.. Or even make

a general regexp for P/N pairs, like "^.*_[PN]$". I think it was

suggested back at the days just before diff pair routing appeared.

2-D[0..7]_[PN] for even more compactness

 

3-Multiple indexes like D[A..C][0..7]_[PN], in good old regexp style.

 

I see your point, multiple iterators in any position rather than a

single iterator at the end.

 

Odd that you mention regexps.  Every time I type in a bus name index I

use a dash to specify a range, and have to go back and correct it to two

dots.  Creatures of habit, I guess.

 

Jorge, I'm in for this one.

Regexp could be used more in Eagle. I use the strxstr() often in ULP's.

I would like to see it used for show commands too, but thats for a

different thread.

I'll support this suggestion too, I use busses a lot and diff pairs quite

often, it would be very nice to be able to have a nice compact notation to

specify the diff pairs within a bus. Happy with either option 1 or 2 for

this. Option 3 looks like a great idea too, would have to have a think

about the possibilities for how the extra flexibility could be used but in

general I like all these ideas.

--

Web access to CadSoft support forums at www.eaglecentral.ca.  Where the CadSoft EAGLE community meets.

On 09/22/2016 08:52 AM, Morten Leikvoll wrote:

On 22.09.2016 14:41, Chuck Huber wrote:

On 09/22/2016 07:38 AM, Morten Leikvoll wrote:

 

Currently, bus names may contain a single end bus of signal names

ending with the index#, like D[0..7],A[0..15]

 

However, diff signals needs to end with _P or _N for eagle to detect

them as diff, hence they can not be added to a bus using the index

range. D[0..7]_P,D[0..7]_N doesnt work. This means, diff buses must be

written with very long names.

 

I suggest that bus names should at support one of the following,

ranging from simple to complex effort:

 

1-Index in the middle of the name, such as D[0..7]_P,D[0..7]_N

 

One could use DP[0..7],DN[0..7], or D_P[0..7],D_N[0..7].

 

I guess you mean that the diff definition has to change instead of bus

naming definitions. DPn/DNn will not be routed as diff as it is now.

Yes thats an option, to allow something after the P or N.. Or even

make a general regexp for P/N pairs, like "^.*_[PN]$". I think it was

suggested back at the days just before diff pair routing appeared.

 

2-D[0..7]_[PN] for even more compactness

 

3-Multiple indexes like D[A..C][0..7]_[PN], in good old regexp style.

 

I see your point, multiple iterators in any position rather than a

single iterator at the end.

Hmmm...  So would you want the entire bus to route in the same manner as

a single diff pair routes today (the follow-me routing feature), or

would you prefer to route individual pairs separately?

I kind of like this.  An entire bus, differential pairs or not, could be

routed in this manner.

Later,

    - Chuck

On 23.09.2016 17:58, Chuck Huber wrote:

On 09/22/2016 08:52 AM, Morten Leikvoll wrote:

On 22.09.2016 14:41, Chuck Huber wrote:

On 09/22/2016 07:38 AM, Morten Leikvoll wrote:

 

Currently, bus names may contain a single end bus of signal names

ending with the index#, like D[0..7],A[0..15]

 

However, diff signals needs to end with _P or _N for eagle to detect

them as diff, hence they can not be added to a bus using the index

range. D[0..7]_P,D[0..7]_N doesnt work. This means, diff buses must be

written with very long names.

 

I suggest that bus names should at support one of the following,

ranging from simple to complex effort:

 

1-Index in the middle of the name, such as D[0..7]_P,D[0..7]_N

 

One could use DP[0..7],DN[0..7], or D_P[0..7],D_N[0..7].

 

I guess you mean that the diff definition has to change instead of bus

naming definitions. DPn/DNn will not be routed as diff as it is now.

Yes thats an option, to allow something after the P or N.. Or even

make a general regexp for P/N pairs, like "^.*_[PN]$". I think it was

suggested back at the days just before diff pair routing appeared.

 

2-D[0..7]_[PN] for even more compactness

 

3-Multiple indexes like D[A..C][0..7]_[PN], in good old regexp style.

 

I see your point, multiple iterators in any position rather than a

single iterator at the end.

 

Hmmm...  So would you want the entire bus to route in the same manner as

a single diff pair routes today (the follow-me routing feature), or

would you prefer to route individual pairs separately?

 

I kind of like this.  An entire bus, differential pairs or not, could be

routed in this manner.

I was more on schematic level, but you could suggest bus routing in a

new thread. Sometimes I have missed it, even for single end buses. It

would have to define a way to select multiple airwires for parallel routing.