Re: [Simulavr-devel] Python again ;-) and some other things

[ThomasK]

Hi Michael,

> The LIBIBERTY variable never gets set.

You are right! I've trusted LIBS variable and havn't checked, what's set 
there. And the best variant to check it: ... try it! So I've seen, that 
libiberty.a will not be linked. But this isn't my problem: I get:

> python -c "import _pysimulavr"
Traceback (most recent call last):
  File "<string>", line 1, in <module>
ImportError: ./_pysimulavr.so: undefined symbol: _ZN11SystemClock8InstanceEv

And also, if you look for the link call, then it's sure, that all 
objects from simulavr are not linked. Your make snippet uses instead of 
$(libsimulavr_la_OBJECTS) $(libsim_la_OBJECTS) and that's the problem on 
my site. (btw. I use standard CVS checkout, maybe you use some patches 
more, which I havn't)

Btw. is libm and libncurses necessary? And where it is used in simulavrxx?

> I'm not so much interested in GUIs as the ability to write unit tests,
> preferably on multi-processor systems.

Hm ... sounds for me, that we have nearly the same goals. My primary 
goal now is to get a possibility to use simulavr in unit tests for my 
coding. And for implementing ATMega16 support too to verify, that 
hardware like timers and so one run correctly.

> That said, I've no objection to others using my work to write GUIs.

You wrote: "... using my work ..." - is pysimulavr.i from you (and in 
this case, I think, also make rules to build _pysimulavr.so)? (this 
question only to know, who I could ask, if there are some questions)

> He needs a grandmother clock.
> He could call it Mrs. Ogg.

Nice! :-))

> The address below is fine.

I'll see, if I've finished this. Next days ...

cu, Thomas