Topic: lisp don't work!!!why? (non-ANSI characters)

[userzhl]

(setq file "D:\\窗帘_F9AAFA6A-5E9B-4C46-8283-F10E8F70936A.atc")
(setq new "c:\\《工具》\\《工具》")
(setq fn (open file "r"))
(setq data (read-line fn))
(close fn)
(SETQ DD (x$sub new "D:\\《工具》" DATA))
(setq Fn2 (open file "w"))
(write-line DD Fn2)
(close Fn2)
;lisp don't work!!!why?

[Kerry]

What is this function ?
x$sub


ps:
Poll is not needed to post question

[userzhl]

(defun x$sub (n$ o$ str / n)
  (setq n 0)
  (while (setq n (vl-string-search o$ str n))
   (setq str (vl-string-subst n$ o$ str n))
   (setq n (+ n (strlen n$)))
  )str
)

[Keith™]

The application appears to work fine on my system, with the exception of the asian characters. I cannot test that because they do not copy correctly to my computer.

I saved your files, created a test lisp and successfully changed "<GeneralProperties>" to "<SpecialProperties>" without issue.

I did make one change to your code and that was to eliminate the starting point for the string search and string substitute. Because you are editing every occurence in the file, you can safely omit that parameter.

Code - Auto/Visual Lisp: [Select]

1. (defun x$sub (n$ o$ str / n)
2.   (while (vl-string-search o$ str)
3.    (setq str (vl-string-subst n$ o$ str))
4.   )str
5. )

[Lee Mac]

I did make one change to your code and that was to eliminate the starting point for the string search and string substitute. Because you are editing every occurence in the file, you can safely omit that parameter.

I would be inclined to retain the position argument to account for the case in which the replacement string contains the search string pattern, e.g. replacing 'is' with 'this'. For such an example, the ommission of the incrementing position argument would cause the code to loop infinitely.

[userzhl]

Keith™
Lee Mac
still don't work?

[Keith™]

I think it may be the asian characters. Can you verify that vl-string-subst and vl-string-search accurately finds the asian characters?

[userzhl]

;and why this is all right?
(defun x$sub (n$ o$ str / n)
  (setq n 0)
  (while (setq n (vl-string-search o$ str n))
   (setq str (vl-string-subst n$ o$ str n))
   (setq n (+ n (strlen n$)))
  )str
)
(setq file "D:\\窗帘_F9AAFA6A-5E9B-4C46-8283-F10E8F70936A.atc")
(setq new "c:\\")
(setq fn (open file "r"))
(setq data (read-line fn))
(close fn)
(SETQ DD (x$sub new "D:\\" DATA))
(setq Fn2 (open file "w"))
(write-line DD Fn2)
(close Fn2)

[irneb]

As Keith has hinted at ... it's due to the non-ANSI characters. Remember the vl-* functions were made originally in the mid-late 90's (Vital Lisp): long before ACad became UniCode compliant. So it's no wonder it has some strange quirks when using unicode characters on these functions.

Try to rewrite your x$sub function to only use AutoLisp function calls instead of Visual Lisp. I've found that the direct ARX functions like substr/wcmatch/strcat/etc. seem to work with strings which confuse vl-string-* functions.

Edit, perhaps this (though it's not very efficient):

Code - Auto/Visual Lisp: [Select]

1. (defun str-subst-all (new old string / search return)
2.   (setq search (strcat old "*") return "")
3.   (while (not (eq string ""))
4.     (if (wcmatch string search)
5.       (setq string (substr string (1+ (strlen old))) return (strcat return new))
6.       (setq return (strcat return (substr string 1 1)) string (substr string 2))))
7.   return)

[userzhl]

还是没解决

[irneb]

还是没解决 <Still not resolved>

OK, I've created a file named test.txt on my C drive in a folder with this path:

Code: [Select]

C:\\TEMP\\测试\\test.txtIf I type the folling into the command prompt:

Code: [Select]

Command: (setq fn (findfile "C:\\TEMP\\测试\\test.txt"))
"C:\\TEMP\\测试\\test.txt"As you can see it works fine for unicode letters in the path.

What might have an issue is the lisp write-line as it saves to file in ANSI format. Perhaps look into something Lee's done before: http://www.theswamp.org/index.php?topic=40709.0

Edit: It seems the open function causes an issue already:

Code: [Select]

Command: (setq fn (findfile "C:\\TEMP\\测试\\test.txt"))
"C:\\TEMP\\测试\\test.txt"

Command: (setq f (open fn "r"))
nilSo I'm guessing you might have to use a FileSystem object to obtain the 8.3 names for the path and then send that to open instead of the unicodes.

[irneb]

It seems even the FileSystem object gives issues:

Code: [Select]

Command: (setq fs (vlax-get-or-create-object "Scripting.FileSystemObject"))
#<VLA-OBJECT IFileSystem3 00000000006d66d0>

Command: (setq fn (findfile "C:\\TEMP\\测试\\test.txt"))
"C:\\TEMP\\测试\\test.txt"

Command: (setq f (vlax-invoke fs 'GetFile fn))
; error: Exception occurredPerhaps you need to use DosLib.

Code: [Select]

Command: (setq fn (findfile "C:\\TEMP\\测试\\test.txt"))
"C:\\TEMP\\测试\\test.txt"

Command: (setq sfn (dos_shortpath fn))
"C:\\TEMP\\D147~1\\test.txt\\"

Command: (setq sfn (vl-string-right-trim "\\" sfn))
"C:\\TEMP\\D147~1\\test.txt"

Command: (setq f (open sfn "r"))
#<file "C:\\TEMP\\D147~1\\test.txt">

Command: (read-line f)
"C:\\TEMP\\还是没解决"

Command: (close f)
nilNotice though that read-line reads the UTF8 file as if it's an ANSI encoded file. So stuff inside that file needs to be converted properly - see the link in my previous post for working with unicode files properly.

[userzhl]

thank you irneb!
read the file is OK!
(setq file "d:\\窗帘_F9AAFA6A-5E9B-4C46-8283-F10E8F70936A.atc")
(setq data (vk_ReadTextStream file "utf-8"))

But,how to replace data?