Remove the slprj folder programmatically

There is an 'slprj' in the working directory that is from another Simulink version. Consequently, I receive the error

'The existing Simulink code generation folder in the current folder was created for a different version of MathWorks products. This project folder is not compatible with the current version. To continue you must manually remove the slprj folder and any generated code files it contains.'

when trying to compile my model. However, I do not want to manually remove the folder each time. The programmatic approach via

rmdir('slprj') 

from the MATLAB command line also issues an error.

So how can I programmatically remove the 'slprj' folder?

[Jan]

If you mention, that an error occurs, be so kind to post the error message also. It is easier to fix an error than to guess, what the error is.

[Bonpland]

Sorry, you're absolutely right.

The generic error, when trying

rmdir('slprj')

is

Error using rmdir
No directories were removed.

[Jan]

The directory migt be in use, e.g. when a contained file is open. Is 'slprj' a folder inside the current directory? Using absolute path names is more secure than relying on the current directory.

[Bonpland]

1. I can definitely exclude that some file is in use.
2. Absolute path names are no options, as the model callback is supposed to work on changing platforms. Therefore the rmdir command has to apply to the pwd.

[Image Analyst]

What does this do:

folder = 'slprj'
if exist(folder, 'dir')
      rmdir(folder)
else
	message = sprintf('The folder %s does not exist', folder)
	uiwait(errordlg(message));
end

[Bonpland]

Error using rmdir
No directories were removed.

[Image Analyst]

So it is a existing subfolder of your current folder but it won't let you delete it. The only reason I can think of is you don't have permission, either because the folder was being viewed in File Explorer, or you have a file from it open in some program, or the folder is readonly, or similar.

Can you delete it manually in WIndows Explorer (Windows 7), File Explorer (Windows 10), or Finder (mac)?

[Bonpland]

That would have been my first explanation. However, manually removing this folder is possible without issues.

[Mostaza Con huevos]

someone else can helpme whit the folder slprj all the folder.

Answer 1

[Sean de Wolski]

What about with the 's' option

rmdir('slprj', 's')

[Gordon Lai]

The full path version of this worked for me as 'slprj' is not empty (nor is it usually).

Note: I first added it and subfolders to the MATLAB path, using

addpath(genpath(pwd))

(unsure if that is necessary)

[Sean de Wolski]

Typically if you're calling addpath/genpath, you should be using a project. This also lets you customize code generation settings at a project level.

https://www.mathworks.com/help/matlab/matlab_prog/create-projects.html

Answer 2

[Jan]

This does not solve the problem, but helps to identify the problem: Catch the error message:

[status, msg] = rmdir('slprj')
if status ~= 1
    fprintf(2, '%s\n', msg);
end

By the way, of course you can use absolute path names even on different platforms. This does not mean a hard coded path name, but that you do not rely on the current directory to be, what you expect. Remember that each timer and GUI callback can use cd to change the current directory, and rmdir('slprj') might access an unexpected folder. Better:

baseFolder = cd;  % Or even better: define it with a meaningful folder
[status, msg] = rmdir(fullfile(baseFolder, 'slprj'))
if status ~= 1
    fprintf(2, 'Folder: %s\nProblem: %s\n', ...
        fullfile(baseFolder, 'slprj'), msg);
end

Now the error message should be more useful.

[Bonpland]

The error message still says

No directories were removed.

Also, the slprj folder contains nothing but empty subfolders, so write access restrictions on my Win 10 machine should not be the issue.

[Jan]

@Bonpland: This sounds strange, because then the replied status must be 1, which means a successful deleting. In case of a problem, the error message written by my code would contain the name of the folder.

Even empty folders can require different privileges for deleting, so it does not matter if the subfolders are empty.

Do you have any evidence, that the current directory is what you expect it to be? Is the folder really empty, when you call rmdir or did you test this some seconds later only in the Windows Explorer - with enabled display of hidden and system files? Remember that e.g. the virus scanner might access and block the files, so what happens with a fallback method:

ready  = false;
count  = 0;
folder = fullfile(baseFolder, 'slprj');
while count < 10 && ~ready
   count         = count + 1;
   [status, msg] = rmdir(folder);
   if status == 1
      ready = true;
   else
      fprintf(2, 'Folder: %s\nProblem: %s\n...retrying...\n', ...
          folder, msg);
      pause(0.5);
   end
end
if count == 10
   fprintf(2, 'Deleting folder failed: %s\n', folder);
end

There is still a detail you did not mention, because you are not aware of it. So please post as much details as you can to give the readers a chance to see, what you do not see.

[Bonpland]

Sorry, I do not unterstand what you're looking for, especially with the last paragraph of your last comment. So, let me reiterate the facts:

1) As expected, the output of your script is just a recurring

Folder: D:\dev\slprj
Problem: No directories were removed.
...retrying...

The status is '0'.

2) The folder 'slprj' may be removed manually from the MATLAB GUI or in the Windows Explorer.

3) Creating a new folder 'slprj' manually somewhere else and trying to delete it programmatically in MATLAB is also no issue.

From 1)-3) I infer that the automatic creation of the folder 'slprj' is imposing some kind of access restriction which cannot be circumvented from within MATLAB. However, this is in no way contained in the error message.

Any help would be greatly appreciated.

[Jan]

The next question: What does "There is an 'slprj' in the working directory that is from another Simulink version" mean? Is "the other Simulink version" still running?

Answer 3

[Bonpland]

Are there any news on this issue?

It would be good to have at least a feedback on how this problem is perceived by MathWorks.

[Jan]

Please do not post a pseudo-answer only to bump your question. This wastes the time of all readers. This is the public forum. If you want MathWorks to react to your question, contact them directly using the "Contact Us" link on top of this page.