emc2 Starting Error

Dears Linuxcnc

I've just installed Slackware 2.6.10 with RTAI 3.3 and kernel patch 2.6.10-r12.
After an infinite number of trials I succeeded on make the realtime tests work.
I installed also the X11 environment with a lot of problems with the mouse, the video card, the recorder and so on, that are solved now!!!
I've also suceeded on compiling the emc2 without errors and have learned that the order of "insmoding" the modules is very important, because if there are no defined symbols the initialization will abort.

On that net of troubles I found the folowing:
There is an undefined symbol "ceil" that appears on the start of emc.
If I invert the options and the target in the "make" sentence I should solve the problem.
But I can't insert that in an infinite and recursive configure file!!!

The question is: Are there any recommended parameters I should give in the ./configure (--xxxxx_xxxx_xxx) to avoid this?

Or maybe I should insert a module before...

Yours:

Eng. Alejandro L. de la Barra.

Hi

I am not really sure what you are describing, but a couple of pointers if they help.

ceil() and floor() are C++ functions which round up and down respectively whatever value is passed to them.
There are many over-riden types and if you do not have for instance C99++98 maths support and the code is using ceil() on float values instead of ceilf() it will have problems

Regards changing the Makefile, not sure exactly what you mean by

If I invert the options and the target in the "make" sentence

but look at Makefile.in and see what can be changed there, rather than configure.

regards

Thanks for your answer!!
This is a very specific topic and I had nobody to ask. Even a little clue is important.

ArcEye wrote:

Hi

I am not really sure what you are describing, but a couple of pointers if they help.

ceil() and floor() are C++ functions which round up and down respectively whatever value is passed to them.
There are many over-riden types and if you do not have for instance C99++98 maths support and the code is using ceil() on float values instead of ceilf() it will have problems

Regards changing the Makefile, not sure exactly what you mean by

If I invert the options and the target in the "make" sentence

but look at Makefile.in and see what can be changed there, rather than configure.

regards

I didn't want to bother with details, but this is what I found in the WEB:

"Compile it with:

$ gcc -lm -o ceil ceil.c this shouldn't work
$ gcc -o ceil ceil.c -lm this works because I tried it on my machine.

stackoverflow.com/questions/8266183/link...efine-ceil-on-ubuntu
This is the link if you want to dive in the deep!!

OK. I'll try again to modify the Makefile.in

Last time I tried, the "make" detected it, "repaired it", and also made a new "vmLinuz".
I was frightened!!

It certainly does seem to be linkage in the maths library that would cause the problem

I came across this which gives one solution at makefile level

stackoverflow.com/questions/13326723/lm-...ary-in-makefile?rq=1

regards