TestSuite and other Failures in today's CVS head

I'm running on a linux (slackware-12.1) box with clisp-2.47.  The build
went without complaint, but make check reveals:

Running tests in rtest15: 
********************** Problem 237 ***************
Input:
ev(Y2, numer)


Result:
[.5422366033071744, 1.943092630269517,
1.969205044088926, .5360853616153969, 
                                          .4610193427194349,
1.716205223757663]

This differed from the expected result:
[.5422366033071744, .4814049171519623, .4750213033941648,
1.645105736576537, 
                                            1.91297202083972,
0.51387624945997]

********************** Problem 238 ***************
Input:
                         8         1                            1
8
[inverse_jacobi_dc(float(-), float(-)), inverse_jacobi_dn(float(-),
float(-)), 
                         7         7                            7
7
                        1         8                            8
9
inverse_jacobi_ds(float(-), float(-)), inverse_jacobi_nc(float(-),
float(-)), 
                        7         7                            7
7
                        8         9                            8
9
inverse_jacobi_nd(float(-), float(-)), inverse_jacobi_ns(float(-),
float(-))]
                        7         7                            7
7


Result:
[.5422366033071744, 1.943092630269517,
1.969205044088926, .5360853616153969, 
                                          .4610193427194349,
1.716205223757663]

This differed from the expected result:
[.5422366033071744, .4814049171519623, .4750213033941648,
1.645105736576537, 
                                            1.91297202083972,
0.51387624945997]


I also ran make install and noticed the following run-time errors:

(%i1) load("eigen");
Could not find `eigen' using paths in
file_search_maxima,file_search_lisp.

Sure enough, find -name "*eigen*" in /usr/local/lib/maxima
and /usr/local/share/maxima don't show any matches.

Finally (I've posted this before):

I've set up my .emacs to support maxima and imaxima as:

        (autoload 'imaxima "imaxima" "Image support for Maxima." t)
        (autoload 'maxima-mode "maxima" "Major mode for writing Maxima
        programs" t)
        (autoload 'maxima "maxima" "Run Maxima interactively" t)
        
        ;;; add autoload of imath.
        
        (autoload 'imath-mode "imath" "Imath mode for math formula
        input" t)
        
        
        ;;; Make the line effective if you want to use maxima mode with
        imaxima.
        
        
        ;(setq imaxima-fnt-size "large")
        (setq imaxima-fnt-size "Huge")
        ;(setq imaxima-print-tex-command "latex %s; dvipdf %s.dvi
        imax.pdf; open imax.pdf")
        (setq imaxima-use-maxima-mode-flag t)
        
        (autoload 'maxima-mode "maxima" "Major mode for writing Maxima
        programs" t)
        (autoload 'maxima-minibuffer "maxima" "Run Maxima from the
        minibuffer" t)
        
        (setq load-path (cons "/usr/local/share/maxima/current/emacs"
        load-path))
        (setq auto-mode-alist (cons '("\\.ma[cx]" . imaxima)
        auto-mode-alist))
        ;(setq auto-mode-alist (cons '("\\.ma[cx]" . maxima)
        auto-mode-alist))
        
This has worked reliably for a long time; however, when I try to run, I
get:

Maxima encountered a Lisp error:

 
SYSTEM::SET-STREAM-EXTERNAL-FORMAT on #<IO TWO-WAY-STREAM
  #<INPUT UNBUFFERED FILE-STREAM CHARACTER #P"/dev/fd/0" @1>
  #<OUTPUT UNBUFFERED FILE-STREAM CHARACTER #P"/dev/fd/1">> is illegal

Automatically continuing.
To reenable the Lisp debugger set *debugger-hook* to nil.

*** - THROW: there is no CATCHer for tag RETURN-FROM-DEBUGGER
Break 1 [1]>
block(load(("/usr/local/share/maxima/current/emacs/imaxima.lisp")),
linenum:0)$

*** - SYSTEM::READ-EVAL-PRINT: variable BLOCK has no value
The following restarts are available:
USE-VALUE      :R1      Input a value to be used instead of BLOCK.
STORE-VALUE    :R2      Input a new value for BLOCK.
ABORT          :R3      Abort debug loop
Break 2 [2]> 

I've been talking with Yasuaki Honda about this, but I'm wondering if
the problem might be a missing packages issue instead?