Delete all directory and files on disc

[Cpcdos]

Hi!

I want to know without formatting, delete all files and directory on a disc (or directory)

Thank

[TJF]

• del *

or in FB

• SHELL("del *")

[fxm]

del *.*

[Roland Chastain]

Another proposition.

Code: Select all

@echo off

rem DelTree.bat

rem Deletes in a specified directory all subdirectories including any files
rem and then deletes the directory itself.

set root="c:\users\roland\thedirectory"

for /d /r %root% %%f in (*) do rd "%%f" /s /q

rd %root%

[Cpcdos]

Thank you
i know del *.* or * etc.. but i can't delete all directories, files and full subdirectories


Roland
I have this :


«CACA» contains files and directories and subdirectories

I use FreeDos *

[Roland Chastain]

Since you don't use a batch file, but type in the command directly, maybe it's %f and "%f" (instead of %%f and "%%f"), but I am not sure.

And why not the FreeDOS deltree command ?

[Cpcdos]

Ho!
I'm stupid
deltree I had not thought !

But if people found a freebasic code, this interest me ! :)


Thank you

[DOS386]

> But if people found a freebasic code, this interest me !

You just have to copy the behavior of FreeDOS DELTREE tool, look at http://www.freebasic.net/wiki/wikka.php?wakka=KeyPgDir DIR function. Run through the directory and inspect the "D" attribute of every found item, if file, delete it (try KILL or ASSASSINATE command), if directory, process it recursively, after you emptied a direcotry, you can delete it using RMDIR command. Note: with DOS target of FB, the KILL command apparently is able to delete even files having the "R" attribute set and empty subdirectories too (http://freebasic.net/forum/viewtopic.php?t=13128), but it definitely won't delete subdirectories with all content - you must do the recursive processing yourself.

[Cpcdos]

Ho thank you very much ! :)

[marcov]

Doesn't deleting files from a directory you are findfirsting make the ongoing findfirst unreliable? I'm not 100% sure how dos reacts to that.

If that is the case, do a findfirst only recording filenames till memory is full, then delete them all, and then do findfirst again.

That takes keeping track of some things though.

[TJF]

But if people found a freebasic code, this interest me ! :)

Code: Select all

#INCLUDE ONCE "dir.bi"

#IFDEF __FB_UNIX__
#DEFINE SLASH "/"
#ELSE
#DEFINE SLASH "\"
#ENDIF
#DEFINE NL !"\n"

SUB delete_all()
  VAR n = DIR("*", fbDirectory), t = ""
  WHILE LEN(n)
    IF n <> "." ANDALSO n <> ".." THEN t &= n & NL
    n = DIR()
  WEND

  VAR a = 1, e = a, l = LEN(t)
  WHILE a < l
    e = INSTR(a, t, NL)
    n = MID(t, a, e - a)
    IF 0 = CHDIR(n) THEN
      delete_all()
      CHDIR ("..")
      RMDIR(n)
    END IF
    a = e + 1
  WEND

  n = DIR("*")
  WHILE LEN(n)
    KILL (n)
    n = DIR()
  WEND
END SUB

IF 0 = CHDIR("X:\") THEN _' adapt your drive (or directory) here
  delete_all()

[Edit]
RMDIR added.
Rolands improvements added.
IF 0 = CHDIR(n) added.
[/Edit]

[Roland Chastain]

Hello TJF !

Your code looks fine but, unfortunately, with me (fbc 0.90.0, Windows 8) it crashes.

When I compile with -exx option, I get a "segment violation signal". I will see if I manage to find what happens.

I suggest a little modification, for security reasons :

Code: Select all

IF CHDIR("C:\Atelier\Basic\DelTree\tree") = 0 then
  delete_all()
END IF

Indeed, if I make a typo in the folder name, it could happen that the program does its job with the current directory, instead of the expected one, isn't it ?

[TJF]

Indeed, if I make a typo in the folder name, it could happen that the program does its job with the current directory, instead of the expected one, isn't it ?

Yes, it is. Good proposal!

And I forgot to remove the directories (fixed now in the code above).

When I compile with -exx option, I get a "segment violation signal". I will see if I manage to find what happens.

Sorry, for me this sentence is ambiguous. Does the compiler crash or the compiled executable?

I didn't test the executable, but compiling with fbc-0.24 is fine here (even with option -exx).

[Roland Chastain]

Sorry, for me this sentence is ambiguous. Does the compiler crash or the compiled executable?

Yes, it is ambiguous. I wanted to say that the executable crashes. As I saw that, I compiled with -exx in order to get more information. Here is the message I get (even with your last correction) :

Aborting due to runtime error 12 ("segmentation violation" signal) in delete_all_test.bas::DELETE_ALL()

Has someone else tested the program ?

[TJF]

Here is the message I get (even with your last correction) :

Aborting due to runtime error 12 ("segmentation violation" signal) in delete_all_test.bas::DELETE_ALL()

Thanks for testing and reporting!

I found the problem: scanning directories doesn't work properly on windows. When a filename has no extension, windows reports the file as a directory and the recursion runs endless.

You can fix it by checking the result of the CHDIR statement, see (fixed) code above.