build report/bug report/suggestions

[John Jones <jj@asu.edu>]

Hi,

I built pari-2.0.10.beta fine with gcc, but I had some problems with
g++.  After a make clean, reran Configure, make all, and make bench,
it reported a bug in "program".  I attatched out.program.dyn.diff at
the end (out.program.sta.diff had exactly the same problem).  Related
files are also at the end.

While the build with gcc was clean, the build with g++ had a few
warnings (also at the bottom of this message).

Now for a few suggestions on the build process.

 - the INSTALL guide mentions building with g++, but it doesn't come
   out and say how to do it.  It seems that "Configure -a" is the only 
   way.  It would be better if the C/C++ compiler can be given as an
   argument to Configure (like prefix).

 - setting of prefix on the command line with Configure does not seem
   to be standard.  I am not an expert on configure programs, but I
   seem to recall --prefix=/usr/local/test as being the more standard
   syntax.

 - since the C and C++ versions of libpari are not compatible, maybe
   they should be given different names.  Then it would be easier for
   someone to have both installed.  I admit that not very many users
   use libpari.a directly, and even fewer would want both the C and
   C++ versions, but this seems like it would make more sense to give
   them different names.

Now for the promised attatchments.  First, the build warnings when
using g++:

/usr/bin/g++ -c -O3 -fexpensive-optimizations -malign-loops=2 -malign-jumps=2 -malign-functions=2   -I. -I../src/headers  -o mpqs.o ../src/modules/mpqs.c
../src/modules/mpqs.c: In function `long int * mpqs(long int *)':
../src/modules/mpqs.c:2706: warning: implicit declaration of function
`int unlink(...)'

/usr/bin/g++ -c -O3 -fexpensive-optimizations -malign-loops=2 -malign-jumps=2 -malign-functions=2   -I. -I../src/headers  -o es.o ../src/language/es.c
../src/language/es.c: In function `int term_width_intern()':
../src/language/es.c:269: warning: implicit declaration of function
`int ioctl(...)'

/usr/bin/g++ -c -O3 -fexpensive-optimizations -malign-loops=2 -malign-jumps=2 -malign-functions=2   -I. -I../src/headers -I../src/language -I/usr/local/include -o gp_rl.o ../src/gp/gp_rl.c
../src/gp/gp_rl.c: In function `int change_state(char *, int *, int)':
../src/gp/gp_rl.c:73: warning: implicit declaration of function `int _rl_save_prompt(...)'
../src/gp/gp_rl.c:76: warning: implicit declaration of function `int
_rl_restore_prompt(...)'

Now for the diff file from the BUG warning.

*** ../src/test-bench/32/program	Wed Jun  3 05:55:03 1998
--- out.program.sta	Wed Jul 22 09:57:19 1998
***************
*** 114,125 ****
  350 1.632105905172986668189652273
  400 1.632424285532931448171405619
  ? ? install(addii,GG)
  ? addii(1,2)
! % = 3
  ? kill(addii)
  ? ? getheap
! % = [20, 2986]
  ? print("Total time spent: ",gettime);
! Total time spent: 2010
  ? \q
  Good bye!
--- 114,128 ----
  350 1.632105905172986668189652273
  400 1.632424285532931448171405619
  ? ? install(addii,GG)
+   ***   can't find symbol 'addii' in dynamic symbol table of process.
+ 
  ? addii(1,2)
!   ***   unknown function or error in formal parameters: addii(1,2)
!                                                         ^---------
  ? kill(addii)
  ? ? getheap
! % = [19, 2979]
  ? print("Total time spent: ",gettime);
! Total time spent: 1080
  ? \q
  Good bye!

Now for dft.Config.in.

Attachment: dft.Config.in
Description: Binary data

Finally, just in case it is not in that file, my system is Linux
2.0.34, and gcc -v gives:

gcc version egcs-2.90.29 980515 (egcs-1.0.3 release)
 

                                 John Jones


---------------------------------------------------------------
Department of Mathematics               Office:  PSA 729
Arizona State University                Phone:   965-3725
Tempe, AZ    85287-1804                 FAX:     965-8119
URL: http://math.la.asu.edu/~jj         e-mail:  JJ@asu.edu
---------------------------------------------------------------