Ubuntu Bugs steppgen in 10.04

I dont mean a graph I just mean a text list of what works and what does not work like this:

OS VER EMC VER HARDWARE OK SYMPTOM

U8.04 2.4.3 PARALLEL PORT YES none

U10.04 2.4.6 PARALLEL PORT NO FOLLOWING ERROR

etc

U8.04 no problem even with the latency nor stepgen. It works both software and hardware stepgen.

U10.04 a strange latency. Using the Base period + hostmot raises a fatal conflict.
Stepgens I work weird. Using EMC version 2.5.0 and above, no used at all, evokes folloving error.

Can you give more specific information including max latency in working and non working cases"
When you say base thread causes fatal conflict, what is the error?
and is the basethread actually used? (is an addf xxx basethread statement present)
There is a known bug with _unused_ basethreads and HostMot2
Have you tried setting the stepgens maxaccell value?
What motherboard are you using?

The reason I ask these questions is that its not easy to duplicate this problem
I have EMC 2.5pre on Ubunto 10.04 and testing with a 5I20 and the supplied hm2-stepper
configuration works fine on my system

In my case I have tried the hm2 stepper sample config.
I tried it with master it follow errors quickly.
on 2.4.6 it seems fine.
both use 8.04.

I will edit my personal config to run on 2.4.6 and try it later.

Is this with the same config files in both cases? as the sample config may have changed

Possible it's different. i didn't check.
but I modified my personal config to remove the mux16 (which 2.4 doesn't have) and it seems to jog fine.
I should make a nice clean usable in both versions config, to test with. I'll get back to you.

ok made a plain one in 2.4 then tried it on master (though a slightly old version)

seems to jog fine in 2.4 but in master the X axis follow errors, usually in the negative direction.
The other axis seem to jog ok but thats cause the follow error values are different (larger)

Yes I agree, I have other results ...

Here are the results of today's tests.
Anyway EMC ver. > 2.4 has a bug, sorry

it is ver 2.4.1 all right

and it is master...
(same CNC program)