bug#18037: 24.3; term.el cannot track non-ascii directories

[Johan Claesson]

Ok, good catch.

--- a/term.el   2014-07-16 19:22:09.348823919 +0200
+++ b/term.el   2014-07-16 21:30:28.272894153 +0200
@@ -3424,7 +3424,8 @@
                           (substring string (1+ first-colon) second-colon))))
           (setq term-pending-frame (cons filename fileline))))
        ((= (aref string 0) ?/)
-        (cd (substring string 1)))
+        (cd (decode-coding-string (substring string 1)
+                                   locale-coding-system)))
        ;; Allowing the inferior to call functions in Emacs is
        ;; probably too big a security hole.
        ;; ((= (aref string 0) ?!)

Regards,

/Johan



Eli Zaretskii <eliz@gnu.org> writes:

>> From: Johan Claesson <johanclaesson@bredband.net>
>> Date: Wed, 16 Jul 2014 20:09:48 +0200
>> 
>> --- a/term.el   2014-07-16 19:22:09.348823919 +0200
>> +++ b/term.el   2014-07-16 19:25:33.920825786 +0200
>> @@ -3424,7 +3424,7 @@
>>                            (substring string (1+ first-colon) 
>> second-colon))))
>>            (setq term-pending-frame (cons filename fileline))))
>>         ((= (aref string 0) ?/)
>> -        (cd (substring string 1)))
>> +        (cd (string-as-multibyte (substring string 1))))
>>         ;; Allowing the inferior to call functions in Emacs is
>>         ;; probably too big a security hole.
>>         ;; ((= (aref string 0) ?!)
>
> string-as-multibyte is not right.
>
>> (Motivation: in term-exec-1 the coding-system-for-read is set to 'binary
>> with the following comment:
>> 
>>         ;; The process's output contains not just chars but also binary
>>      ;; escape codes, so we need to see the raw output.  We will have to
>>      ;; do the decoding by hand on the parts that are made of chars.
>> 
>> I think term-command-hook is one place where such decoding is needed.)
>
> Indeed; but string-as-multibyte is not the way to decode the string.
> You need to use decode-coding-string and locale-coding-system.