[sdcc-devel] [ sdcc-Bugs-657351 ] sdcc fails to compile Knuth's rng.c

[<no...@so...>]

Bugs item #657351, was opened at 2002-12-21 15:40
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=100599&aid=657351&group_id=599

Category: msc51(8051) target
Group: None
Status: Closed
Resolution: Rejected
Priority: 5
Submitted By: Nobody/Anonymous (nobody)
Assigned to: Nobody/Anonymous (nobody)
Summary: sdcc fails to compile Knuth's rng.c

Initial Comment:
rng.c:84: error: FATAL Compiler Internal Error in file
'gen.c'
line number '261'  : getFreePtr should never reach here
Contact Author with source code



----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2002-12-29 13:18

Message:
Logged In: NO 

Fantastic--the current cvs now can compile the Knuth rng
code and get the correct result when run. 


----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2002-12-29 09:05

Message:
Logged In: NO 

Ok, if I have some extra time I'll try to pin it down some more.
The two random number generators can be built as stand
alone programs for testing--see the url's for info.
Thanks for taking a look and the suggestions.


----------------------------------------------------------------------

Comment By: Johan Knol (johanknol)
Date: 2002-12-29 08:32

Message:
Logged In: YES 
user_id=63512

Please isolate the problem in a 10 line code sample and 
repost if it is really a bug.

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2002-12-28 10:21

Message:
Logged In: NO 

I have Knuth's rng:

http://www-cs-faculty.stanford.edu/~knuth/programs.html#rng

And another ("usenet"):

http://www.lns.cornell.edu/spr/1999-01/msg0014148.html

...in a program compiled with sdcc. Both compile and produce
output, but the output does not agree with the output
that each program should produce, according to the values 
provided to check the program output contained in the program
source. The "usenet"  sample program should produce seven
zeros but does not, for example. 

I've been able to use sdcc to build your typical "hello
world" style
monitor and wanted to include a random facility as part of a
memory test routine. I'm thinking there's a tricky cast
problem I
haven't noticed yet or perhaps a math bug? What kind of math
should mcs51 be expected to handle at this point? 

The monitor code is available at:

http://www.wraith.sf.ca.us/mon/mon.c
http://www.wraith.sf.ca.us/mon/mon.h





----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2002-12-26 10:34

Message:
Logged In: NO 


Ok, use --model-large --stack-after-data and I've got it
compiled. I'd forgotten about this (blush).

Not A Bug.



----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=100599&aid=657351&group_id=599