On 5/26/06, John Landahl <john@landahl.org> wrote:
So I'd like to try out the new! improved! Celtk2 under Linux/SBCL, but
I'm having some installation problems.  First I found that I had a
really old version of CFFI (thanks, Ubuntu!).  After installing the
latest CFFI tarball, I still get the following error:

* (require :celtk)

; loading system definition from /usr/local/lib/sbcl/site-systems/cffi.asd into
; #<PACKAGE "ASDF0">
; registering #<SYSTEM CFFI {A9D14D1}> as CFFI
; loading system definition from /usr/local/lib/sbcl/site-systems/cells.asd
; into #<PACKAGE "ASDF0">
; registering #<SYSTEM :CELLS {AD05B69}> as CELLS

debugger invoked on a SIMPLE-ERROR in thread
#<THREAD "initial thread" {B29D0D1}>:
  When attempting to read the slot's value (slot-value), the slot
  CFFI::TO-C-DYNAMIC-EXPANDER is missing from the object
  #<CFFI::FOREIGN-BUILT-IN-TYPE :UNSIGNED-LONG>.

Type HELP for debugger help, or (SB-EXT:QUIT) to exit from SBCL.

restarts (invokable by number or by possibly-abbreviated name):
  0: [RETRY ] Retry performing #<ASDF:LOAD-OP NIL {B48D421}> on
              #<ASDF:CL-SOURCE-FILE "tk-structs" {B46B491}>.
  1: [ACCEPT] Continue, treating #<ASDF:LOAD-OP NIL {B48D421}> on
              #<ASDF:CL-SOURCE-FILE "tk-structs" {B46B491}> as having been
              successful.
  2: [ABORT ] Exit debugger, returning to top level.

((SB-PCL::FAST-METHOD SLOT-MISSING (T T T T))
#<unavailable argument>
#<unavailable argument>
#<unavailable argument>
#<CFFI::FOREIGN-BUILT-IN-TYPE :UNSIGNED-LONG>
CFFI::TO-C-DYNAMIC-EXPANDER
SLOT-VALUE
NIL)


Any hints on where to go next?

Hmmm. (require :celtk)? Well, it got all the way up to tk-structs so i guess SBCL is smart enough to know what to do. To rule out an issue here, hop into celtk/load.lisp and start evaluating forms one by one. But really...

That error has me concerned as to whether you have managed to get a clean CFFI environment. CFFI normally complains intelligently if I am off base, and that looks like an internals trip up. Might be a mismatch of old and new code.

Meanwhile I will go stare at tk-structs and the CFFI source to see if I notice anything funny about :unsigned-longs.

kt