My setupper gives the ERROR access denied dialog when overwriting one DLL fails,
its locked&in use sometimes.
what is recommended procedure for doing this..
Can I somehow make INNO to compare if its even necessary to overwrite this DLL (size match?)
InnoSetup automatically compares the files to see if it needs to be replaced by looking at the version resource in the file.
You should add the RestartReplace flag to the [Files] entry for your DLL (and any other file that might be in memory / locked). From the InnoSetup documentation:
restartreplace
When an existing file needs to be
replaced, and it is in use (locked) by
another running process, Setup will by
default display an error message. This
flag tells Setup to instead register
the file to be replaced the next time
the system is restarted (by calling
MoveFileEx or by creating an entry in
WININIT.INI). When this happens, the
user will be prompted to restart their
computer at the end of the
installation process.
To maintain compatibility with Windows
95/98/Me, long filenames should not be
used on entries with this flag. Only
"8.3" filenames are supported.
(Windows NT platforms do not have this
limitation.)
NOTE: This flag has no effect on
Windows NT platforms if the user does
not have administrative privileges.
Therefore, when using this flag, it is
recommended that you leave the
PrivilegesRequired [Setup] section
directive at the default setting of
admin.
Related
I am using NSIS to write an installer for my windows application. After installing the application, an uninst.exe is also generated in my program directory. Later on I need to uninstall my program but I failed to do that in control panel. Therefore I went to the file system and tried to delete the directory. Everything other than that uninst.exe was removed. I've tried changing permission of this file and other methods but it doesn't work.
WriteUninstaller does not set file permissions or any other attribute that might prevent you from deleting it. It sounds like the file might still be in use by something on your machine.
Things to try:
Use Task Manager or Process Explorer to see if there is a uninst.exe process still running.
Use the find handle feature in Process Explorer to find any open handles to the file.
Use Process Monitor to get detailed information about the failed delete operation.
Check %LOCALAPPDATA%\VirtualStore to make sure UAC Virtualization is not tricking you with "ghost" files.
Disable your Anti-Virus.
Reboot the machine and try to delete the file again.
My setup project performs a UAC check because it uses regserver. This is OK (and in fact necessary); however it makes all of the installed files be owned by Administrator , and not by the user that originally ran the installer.
This means that the user that ran the installer cannot edit any of the configuration files that are part of my installation.
I don't want to set permission on the files to be writable by Everyone; they just need to be writable by the user that ran the installation.
Does Inno Setup have any way of assigning ownership of certain files (or all of the files) to the user that ran the self-installing executable?
I am not using the line PrivilegesRequired=admin although it seems makes no difference whether or not that line is active. (Presumably because regserver implies this).
I have a setup that is installing a PDF program. It works perfectly on test machines. When there is other PDF software on the machine, errors pop up because of the large number of shared files that are inuse.
Inno has the /SUPPRESSMSGBOXES option. However I can not seem to configure this to ignore the file copy - it either wants to Abort or Retry. Abort - the setup will fail. Retry - will never succeed because the file is in use.
Does anyone know how to automatically ignore any copy file errors while setting up?
You can use the restartreplace flag for the files that may be in use, it is exactly to suppress those error dialogs.
To quote the documentation of the [Files] section:
restartreplace
When an existing file needs to be replaced, and it is in use (locked) by another running process, Setup will by default display an error message. This flag tells Setup to instead register the file to be replaced the next time the system is restarted (by calling MoveFileEx or by creating an entry in WININIT.INI). When this happens, the user will be prompted to restart their computer at the end of the installation process.
I have created an installation with inno setup. My app (among other things) after is run creates a pdf file inside a subfolder and then opens it. But windows 7 says access denied and exception pops up. What's wrong? How to grant access to subfolders using innosetup?
Here's the code snippet inside ino:
Source: "C:\Users\Someone\Desktop\NET\Animations\*"; DestDir: "{app}\Animations"; Flags: ignoreversion recursesubdirs createallsubdirs
Presumably because you're trying to copy the file from a different user's private folder. That's off-limits. You can only place files into the current user's folder (the one who is running the install). It's difficult to imagine a good reason why you'd want to do otherwise, anyway.
Try using the {userdocs} constant, instead. Use ExpandConstant to expand it to the full path.
If you need things to go in a location that will be accessible to all users, you need to run the installer with Administrative privileges. Then, you'll be able to read/write from the All Users profile directory.
EDIT: Ah, sorry. I totally missed the part of your question where you said you were trying to do this after installation. I just looked at the code and thought this was something you were having Inno Setup do during the setup process.
It's a completely different answer for something after the installation has completed. Windows 7 (thanks to UAC) doesn't let your app (or any app, for that matter) write to system folders. That includes the Windows directory, as well as the Program Files folder and any of the folders it contains. This is a security measure, intended to stop applications from running amok in places where they don't belong.
You have a couple of different options:
If you absolutely need write access to the Program Files folder, you can prompt the user to elevate your application's process. Basically, that means that you'll be requesting Administrative privileges, and they'll see a box from UAC asking them for a password.
I give more information about how you might go about doing that from a C# application in my answer to this question. You'll follow similar steps for an app written in any other language; you're just calling functions built into the Windows API.
The better option, though, is to modify your application so that it doesn't have to write to system folders. That way, you won't have to run with Administrative privileges. This is the intended model for all standard Windows applications. Microsoft has been recommending it at least since the early days of Windows 2000, but you weren't actually forced into following it until Vista.
I talk more about the various places that an app has write access to (and the various uses of each) in my answer to this question.
I want to delete a folder which contains the currently running application. How can i do it..? is there any way of doing it ? i.e the folder which contains the application should delete after the application has finished running ?
Your best bet is probably to use the Win32 API MoveFileEx. It has a flag that can be set for deleting files when they are in use on the next reboot called MOVEFILE_DELAY_UNTIL_REBOOT. Set the new filename parameter of MoveFileEx to NULL to perform this type of delete.
If dwFlags specifies
MOVEFILE_DELAY_UNTIL_REBOOT and
lpNewFileName is NULL, MoveFileEx
registers the lpExistingFileName file
to be deleted when the system
restarts.
Note: Normal files that are in use can be deleted normally using the Win32 API DeleteFile depending on if they were opened (Using the Win32 API CreateFile) with FILE_SHARE_DELETE permission. I don't think running programs by default on Windows have that permission though. When a file is specified to be deleted that is in use but that was opened with this flag, then the file will be removed when the last file handle is closed.
This is hairy. I had to implement this once for a self-patching app, where the patcher had to (by client request) delete itself after installing the patch. You can do this by launching a helper DLL which deletes your process, along with itself.
The full method for deleting your process can be found here: http://www.handcraftedbytes.com/articles/writing-install-and-uninstall
As others have pointed out, you're not going to be able to delete the folder that your executable resides in while it exists there. My suggestion is to:
Use MoveFileEx to move your executable off to a temporary directory,
delete your application's directory,
delete your executable using the self-deleting DLL method described in the link above.
You cannot delete an executable file that is currently running, however you can delete a batch file that is currently running (cmd.exe loads the whole file into memory and then you can delete it).
So the simplest solution would be to launch a batch that tries to delete the .exe in a loop (because it may not work the first time - until your .exe has been unloaded) and then exit your process - with the batch file still running.