Skip to content
  1. Feb 26, 2009
  2. Dec 08, 2008
  3. Dec 07, 2008
  4. Aug 19, 2008
  5. Jul 23, 2008
  6. Jul 18, 2008
  7. Jul 17, 2008
    • Boris Smilga's avatar
      Fixed JSON-BIND to intern symbols at run-time instead of compile-time. · 414d2c86
      Boris Smilga authored
      JSON-BIND has not been updated after the introduction of *JSON-SYMBOLS-PACKAGE*
      on 2007-03-24.  If that variable were then set to some other package than
      KEYWORD, the expansion of JSON-BIND would include unquoted symbols (i.e.
      possibly unbound variables).  Even worse, the call to #'JSON-INTERN was
      in the matrix expansion code, and as such performed at compile time, thus
      invalidating the very idea of *JSON-SYMBOLS-PACKAGE* as a means of user
      customization.
      
      darcs-hash:3c0721c3fcb21de6c06e8d14bbc58c6d48ef81d3
      414d2c86
  8. Jul 07, 2008
  9. Dec 31, 2007
    • boris.smilga's avatar
      Fixed MAKE-OBJECT-PROTOTYPE (T) so that MAX-PACKAGE ignores STANDARD-OBJECT superclass. · 4feb7380
      boris.smilga authored
      When constructing an anonymous class, an empty list of superclasses is
      defaulted to '(STANDARD-OBJECT).  Now if we have all slot names in the
      package KEYWORD the correct behaviour would be to return that package,
      but STANDARD-OBJECT is not in KEYWORD.  (Another bug in
      MAKE-OBJECT-PROTOTYPE (T) led to the deduced maximum package being
      COMMON-LISP, and even if we fixed that, we'd have MAX-PACKAGE issuing
      warnings.)  The solution is to have STANDARD-OBJECT removed from the
      list of superclasses before it is passed to MAX-PACKAGE.
      
      darcs-hash:8314f5e6954fc5bf119f29d06bc1e8e955c25ca2
      4feb7380
    • boris.smilga's avatar
      Fixed handling of package which resulted in test failure. · 4ea7652b
      boris.smilga authored
      The program failed TEST*JSON-SYMBOLS-PACKAGE*.  That was due to the
      manner of operation of WITH-OLD-DECODER-SEMANTICS which worked
      essentially by supplying a canonical prototype, so that the decoder
      has a preconception of what kind of object to create--and where to
      intern names.  The last part was clearly wrong: there should be no
      such preconception, as the user might be setting *JSON-SYMBOLS-PACKAGE*
      (outside WITH-OLD-DECODER-SEMANTICS) to a non-default value, and it is
      this value rather than the one supplied in the prototype that should
      be obeyed.
      
      darcs-hash:918066bcfcc2fe6b90f2e89e297bb6fa404fa2e5
      4ea7652b
  10. Mar 18, 2008
  11. Jan 20, 2008
  12. Jan 14, 2008
    • Henrik Hjelte's avatar
      todo file · 8993ab1a
      Henrik Hjelte authored
      darcs-hash:bf67055f0fdcb81669d0b5967522f7fffd25edd0
      8993ab1a
  13. Jan 05, 2008
    • Henrik Hjelte's avatar
      restore list-decoder defaults · 16540867
      Henrik Hjelte authored
      for backwards compatibility, some
      of my old code broke with the new default
      variable values
      
      darcs-hash:479002faabab45f468b0c228bafbd8911be07d23
      16540867
  14. Dec 28, 2007
  15. Dec 23, 2007
  16. May 31, 2007
  17. Mar 25, 2007
  18. Mar 24, 2007