[bug #63495] File operations fail in the presence of non-UTF-8 pathnames

[Taylor R. Campbell]

URL:
  <https://savannah.gnu.org/bugs/?63495>

                 Summary: File operations fail in the presence of non-UTF-8
pathnames
                 Project: MIT/GNU Scheme
               Submitter: riastradh
               Submitted: Fri 09 Dec 2022 03:27:02 PM UTC
                Category: runtime
                Severity: 3 - Normal
                Priority: 5 - Normal
              Item Group: Suboptimal behavior
                  Status: None
                 Privacy: Public
             Assigned to: None
         Originator Name: 
        Originator Email: 
             Open/Closed: Open
         Discussion Lock: Any
                Keywords: 


    _______________________________________________________

Follow-up Comments:


-------------------------------------------------------
Date: Fri 09 Dec 2022 03:27:02 PM UTC By: Taylor R. Campbell <riastradh>
% touch "$(printf 'h\351llo')"    # héllo, encoded in ISO-8859-1
% ls | hexdump -C
                                                        1.  68 e9 6c 6c 6f 0a   
                              |h.llo.|
00000006

(directory-read ".")
;Ill-formed UTF-8 sequence: #u8(233 108 108 111)

(define the-file-name
  (let ((v (make-vector-8b 5)))
    (vector-8b-set! v 0 #x68)
    (vector-8b-set! v 1 #xe9)
    (vector-8b-set! v 2 #x6c)
    (vector-8b-set! v 3 #x6c)
    (vector-8b-set! v 4 #x6f)
    v))
;Value: the-file-name

(string-from-primitive the-file-name)
;Ill-formed UTF-8 sequence: #u8(233 108 108 111)

(string-for-primitive the-file-name)
;Value: #u8(104 195 169 108 108 111)

(call-with-input-file the-file-name (lambda (port) 'ok))
;The primitive new-file-open-input-channel, while executing the open system
call, received the error: No such file or directory.


Is there any way to get at this file in MIT Scheme now?  Is it intentional
that file system operations will raise errors when confronted with such file
names?  Should string-from-primitive only do utf8->string if the input is
valid UTF-8?  Should string-for-primitive preserve legacy strings instead of
re-encoding them as UTF-8?







    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?63495>

_______________________________________________
Message sent via Savannah
https://savannah.gnu.org/