How to add a new Mesa card to pncconf?

naw that's pretty carazy to try to read.

pncconf would looks like this

cmorley wrote:

Personally I would really dislike using joint names for everything. When I look at the HAL code I want to first know what axis I am looking at then which motor (x or x2) on that axis.

But if someone wanted to code a more joint-centric pncconf I would consider their proposal.

Chris

But with V 2.8/9 you will know what joint is attached for each axis as you have to allocate them.

I've got no history before V 2.8 so I've seen it with new eyes. But I've never used pncconf other than for that one test for you. I hand coded everything becasue the 7i76e was not well supported and back then I would have needed to create a XYZA and and edit.
It was frustrating back then becasue the support for everything I needed to do was still unfolding. Now it would be a piece of cake by comparison...

Oops and Chris sorry for calling you Cam, I knew a Cam Morley once....

But with V 2.8/9 you will know what joint is attached for each axis as you have to allocate them.

Well first I could reverse that argument for you as well - you allocated the Axis so you know what the joints are.
By having axes names in the HAL file I don't need to reference the INI file to know what parts are what.
If everything said joints I would need to reference the INI to see what part of the machine we are talking of.
Keep in mind we can also be just looking at someones HAL file online - we don't always have the machine handy.
It's probably just how I think ...

Since I don't use joints in my personal machines (lathe and regular xyz mill) I don't even think in joints I think axes.
In fact it drives me nuts that I get error messages that mention joints and axes.

Anyways AFAICT it's a huge change that I'm not interested in tackling - heck I was very close to pulling pncconf from master...
If someone put in a reasonably equivalent config program in linuxcnc I would pull it for sure.
Or better yet someone interested in continuing pncconf's development.

Chris

Ps I'm sure Cam is a nice guy too

cmorley wrote:

Since I don't use joints in my personal machines (lathe and regular xyz mill) I don't even think in joints I think axes.
In fact it drives me nuts that I get error messages that mention joints and axes.

I do know the feeling when I get a Joint n is on limit or something, but I count it down!

cmorley wrote:

Ps I'm sure Cam is a nice guy too

He had to be as he was an Aussie! Anyway, I renamed you while I was thinking of this!

I think it would be a shame to see pncconf pulled and I know you have not enjoyed maintaining it. I did look at the source code once when you said you had had enough but I just don't know Python well enough and lets face it 5000 lines or so in one file is not really the easiest code to maintain.. Gmocappy is similar.

phillc54 wrote:

The never ending problem for developers, rocks and hard places...

Yeh, I try not to write code these days and leave it for people like you and Chris but I do know that if I ever took a shortcut, it would come back to bite me!. Mostly with database design when you had a little lookup table and you never took the time to add a unique key field. Eventually I redid them all.

I think it would be a shame to see pncconf pulled and I know you have not enjoyed maintaining it. I did look at the source code once when you said you had had enough but I just don't know Python well enough and lets face it 5000 lines or so in one file is not really the easiest code to maintain..

Not to whine too much ... The linuxcnc project is not like an organized business - things just happen or don't you never know. Pncconf is one of those apps that always sits at the cutting edge of what is new in the project - usually things not documented or mentioned right away. Put that together with devs don't use it (I wish they would try it once in a while) so the little critical bugs aren't usually found until it's released... crazy stuff like PID; if you connect a signal to one specific pin that isn't driven -it is bad- it must be driven or no signal connected. Peter of Mesa figured that one out and let me know.
But that was in a already released version. It just sucks that the program that is supposed to help new users can actually lead them down the wrong path with little bugs until (usually) a year after release of a new version of linuxcnc.

And ya the code is ugly and big. and only one guy knows much about it and he is much more interested in qtvcp ....
Whine off...

Chris