CL:DIRECTORY with wildcard in name

[Mark Evenson]

(directory "/tmp/**")

The value NIL is not of type FIXNUM.
   [Condition of type TYPE-ERROR]

Restarts:
 0: [RETRY] Retry SLIME REPL evaluation request.
 1: [*ABORT] Return to SLIME's top level.
 2: [ABORT] Abort thread.

Backtrace:
  0: (INVOKE-DEBUGGER #<TYPE-ERROR {7ED158A9}>)
  1: org.armedbear.lisp.Lisp.error(Lisp.java:382)
  2: (POSITION-IF #<LOCAL-FUNCTION IN COMPONENT-MATCH-WILD-P {7BF6BC72}> "**" :START NIL)
  3: (SYSTEM::COMPONENT-MATCH-WILD-P "com.apple.launchd" "**" NIL)
  4: (SYSTEM::COMPONENT-MATCH-P "com.apple.launchd" "**" NIL)
  5: (PATHNAME-MATCH-P "com.apple.launchd.cyc79CMFEO" "**")
  6: (DIRECTORY "/tmp/**")
  7: (SYSTEM::%EVAL (DIRECTORY "/tmp/**"))

[Mark Evenson]

Broken at least back to abcl-1.3.2

[Mark Evenson]

And DIRECTORY on a WILD-INFERIORS is also borked:

(directory "/tmp/**/")

[Mark Evenson]

Also, we don't return the truename of all entries like we should.

[Mark Evenson]

Replying to Mark Evenson:

Also, we don't return the truename of all entries like we should.

Actually, we decided not to return TRUENAME for DIRECTORY in order to be able to get the contents of a directory with dangling symlinks.

The extension to ANSI via :RESOLVE-SYMLINKS was added as
<https://abcl.org/trac/ticket/340>, in which it was argued that
symlinks should be considered contents of a directory, and that in
any event, performing a DIRECTORY on a dangling symlink should not
signal an error.

See <https://abcl.org/trac/changeset/14624> for additional information on implementation decision.

[Mark Evenson]

Addressed in r15369.

[Mark Evenson]

Milestone renamed