A few moments ago, I checked in some changes to linalg. I hope
that these are the last major changes to the linalg code before
Maxima 5.9.3. It would be a big help to me if those of you that
can build maxima on Monday or later with something other than GCL
would try running the following script. By the way, my code will not
work under 5.9.2 -- lots of things have changed since October.
------------------------
load(linalg);
mytest(file_search("test-linalg"));
mytest(file_search("test-lu"));
mytest(file_search("test-cholesky"));
mytest(file_search("test-eigens-by-jacobi"));
mytest(file_search("test-polynomialp"));
compile_file(file_search("linalg"));
compile_file(file_search("linalg-utilities"));
compile_file(file_search("lu"));
compile_file(file_search("linalgcholesky"));
compile_file(file_search("eigens-by-jacobi"));
-------------------------------
The output of the tests should look like:
(%i8) batch("C:/maximacvs/maxima/share/linearalgebra/test.mac");
batching #pC:/maximacvs/maxima/share/linearalgebra/test.mac
(%i9) load(linalg)
(%o9) c:/maximacvs/maxima/share/linearalgebra/linalg.mac
(%i10) mytest(file_search(test-linalg))
189/189 tests passed.
(%o10) []
(%i11) mytest(file_search(test-lu))
48/48 tests passed.
(%o11) []
(%i12) mytest(file_search(test-cholesky))
34/34 tests passed.
(%o12) []
(%i13) mytest(file_search(test-eigens-by-jacobi))
24/24 tests passed.
(%o13) []
(%i14) mytest(file_search(test-polynomialp))
16/16 tests passed.
(%o14) []
Let me know if the compiler finds any problems (most likely,
the compiler will complain that listp and matrixp have too
many arguments).
Barton