NAND Patch

[Owen]

This patch adds signal type 5, the NAND pre signal.

A NAND pre signal has the exact ooposite properties of a normal combo signal. In other words, it is red when a normal signal would be green.

Currently, the signal uses the PBS graphics. I do not have enough NewGRF knowledge to chnage this.

I am not sure if I am smashing the semaphore bitflags in rail_map.h. Could a developer check that for me, so I can fix it if I am.

When building a NAND signal, care should be taken to make sure the signals do not form a loop, else OpenTTD will hang changing signal state constantly. You have been warned. I was wondering if the devs have any information on how to avoid this?

Notes:
* Someone check rail_map's bitflags please
* Fix the hang issue (Delay before changing?)
* Custom graphics

Enjoy!

EDIT: Oh yeah, can someone check the code I nicked from PBS applys? Thanks

[Graphite]

You've posted a lot of info about that NAND signal, but you seem to have missed one that interests me: why? What do you gain from being able to build these signals? I wouldn't mind seeing a practical application of them in a screenshot.

[Owen]

They are the basic building blocks of all other gates. From a NAND, you can build an AND, NOR, NOT, OR, etc.

I have a practical application also:
http://res1.humgun.com/ttdnand/statpref.png

This somewhat models a really busy station in the OpenTTD Coop Sandbox, where we want a futher away station to be prefered. With a NAND gate setup as shown, it would be possible.

The trains prefer the extension, which we don't want them to. In the example, the station on the right is the overflow station. While in this example, trains prefer the overflow station, the signaling dictates which way they go anyway.

NANDs can be extremely useful.

EDIT: Made image a link

[Graphite]

Hmmm... to be honest, you completely lost me on that image. I've built my fair share of complicated stuff, but that one surpasses my current ability to figure it out (though it's about bedtime, so that may have something to do with it).

I guess it does have its uses, though I doubt I'll ever use it myself. I am a bit worried that the signal builder is getting a bit overloaded though. If PBS is added back to the mix, you'd be cycling through five signals. If these are added to the game, then the signal gui could do with a make-over as well to resolve that issue.

P.S. That image is rather wide to be placed inline in the post. You might want to either crop it or put it in an attachment.

P.P.S. Transistor Tycoon anyone?

[Wolf01]

some pics, by popular demand

EIDT: updated the diff to avoid game hangs on NAND loops

[Nickman]

I'll be messing around with this one if I get the code to compile

[Owen]

Update, to remove the hang issue. A presignal change can no longer affect more than 10 blocks however. I hope that this becomes configurable

[Owen]

I created a minaturzied (Well, as possible ) version of the RS Flip-Flop:

[Wolf01]

now i can reset my memory cell

[Owen]

Ensuring all tracks are free:

EDIT: 99% Perfect balancer (Trains will go to A, B, A, B 99% of the time. The 1% being when a train sits on a flipper)

[sidew]

If I find my old book of microelectronics, maybe 'll try an small binary adder with carry bit

Good old times when I attend the campus...

[Owen]

Hehe. Might need to increase the cascade value, I doubt it will work on 10 hops...

[Frostregen]

I don't know if you fixed it already, but you need to change the Set/GetSignalVariant functions to make Semaphores work again.

Code: Select all

static inline SignalVariant GetSignalVariant(TileIndex t)
{
	return (SignalVariant)GB(_m[t].m4, 3, 1);
}

static inline void SetSignalVariant(TileIndex t, SignalVariant v)
{
	SB(_m[t].m4, 3, 1, v);
}

They can be found in rail_cmd.c.
Just modify them to get/set bit 3 instead of 2.

This gave me some serious headache while trying to use them with my copy&paste patch

[Owen]

Version 4, incorporating Frostregen's fixes although they were developed independently. I also bumped the cascade limit to 20 since 10 seemed limiting.

[Owen]

NAND Path version 5

Changes:
* Only counts NAND signals now
* When an overflow occurs, displays an error message. While it doesn't tell you the location, chances are this will only occur while building anyway[/b]

To apply, gunzip the diff and apply. Under Linux, do the following:

Code: Select all

cat nandv5.diff.gz | gunzip | patch

Edit: Added MiniIN patch

[Wolf01]

miniIN diff with configurable nand cascading in configure patches

[Owen]

A relay, much like it's mechanical counterpart, takes an input at one half and produces an output at the other.

[TrueBrain]

Some code-wise tips/trucks you might want to alter:

- Follow the OpenTTD Coding Guidelines
-> if() is wrong, it should have a space between if and ()
-> Don't printf stuff, use DEBUG()
- The define in variables.h might not be the best place. Also, capitals for defines are more common

I am sure there are some more things (nobody is perfect ), but that is what I noticed after a quick look I have to say I like it very much, it is nice to have a full logic-set of signals

Ps: post it at http://bugs.openttd.org, so devs will notice it

[Owen]

-> if: Thats my normal way, ill fix it to follow the guidelines
-> printf: Shouldn't (normally) escape my hands
-> deifne: It was temporary; I'm moving it to a patch option

[CobraA1]

Using trains as logic circuits .

Now that's something I've never seen done before.