Unconditional delete - Why does this not exist????

[Nothinman]

I guarantee that if MS designed the system to allow you to delete a file whenever, and possibly bring the process that was using it down, you would bash them for that too.

Unix has had the best of both worlds for years now. When you delete a file the directory entry is removed so it can't be used by anything else and the on-disk data is freed as soon as the last file handle is closed. It makes upgrading services in place extremely simple as well, you just replace the files and the running service uses the old versions until you restart it.

Ignore that, he just bashes MS whenever he can

Yes, but it's also true. And the move on boot crap isn't a good solution, if I wanted to reboot to delete the file I could just boot into safemode or a rescue CD and delete it that way.

 

[jadinolf]

Originally posted by: Markbnj
BCWipe from Jetico. Just hit "Delete with Wiping" in the explorer context menu and say goodbye to the data.

But the "holy grail" of Windows? Did someone find your pr0n in the recycle bin?

Probably mommy.

 

[bsobel]

Unix has had the best of both worlds for years now. When you delete a file the directory entry is removed so it can't be used by anything else and the on-disk data is freed as soon as the last file handle is closed. It makes upgrading services in place extremely simple as well, you just replace the files and the running service uses the old versions until you restart it.

Windows has the same, it's called delete on close. However, you can not open the file to set that if the file is exclusively locked (which is the issue we are discussing)

Yes, but it's also true. And the move on boot crap isn't a good solution, if I wanted to reboot to delete the file I could just boot into safemode or a rescue CD and delete it that way.

Oh please, it's EASIER to you to reboot into safemode or a rescue CD (that requires a SECOND reboot to get back to your normal environment) than to run that tool and reboot when you get around to it? Come on, I think of you as a friend, but your reaching on this one.

 

[Nothinman]

Windows has the same, it's called delete on close. However, you can not open the file to set that if the file is exclusively locked (which is the issue we are discussing)

Then it's worthless since virtually everything opens files with an exclusive lock, hell even explorer does it when it generates the preview of a file.

Oh please, it's EASIER to you to reboot into safemode or a rescue CD (that requires a SECOND reboot to get back to your normal environment) than to run that tool and reboot when you get around to it? Come on, I think of you as a friend, but your reaching on this one.

Yea, easier wasn't the right word. But both solutions suck, IMO.

 

[PELarson]

Originally posted by: bsobel

Originally posted by: PELarson

Originally posted by: Doomer
This is the holy graill of Windows as far as I'm concerned, Why does it not exist????

When I tell windows to delete a file or folder, I would like it to do so NO MATTER WHAT !!! unconditionally and irretrievably.

If anyone knows of such a utility, you'll forever be my hero.

10-15 minutes, google search, words "registry tweak bypass recycle"

answer -

[HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\Explorer\BitBucket]
;Do Not Send Files to Recycle Bin
"NukeOnDelete"=dword:00000001

Reading is not comprehension

More like dropping the intellectual level as low as the original poster was impossible!

Because if the idiot wants something that will delete loaded files from a running program he is either trying to do something nasty to someone or just plain stupid. I suspect the first but see the second.

The only real method would be to use or write a program that wrote zeros X amount of times to the exact file location and tie it into the delete key.

 

[xtknight]

http://www.codeguru.com/Cpp/W-P/files/fileio/article.php/c1287/
http://www.bleepingcomputer.com/files/killbox.php

They will kill file handles. Also if it's a DLL you want to delete such as spyware, then you can also force an unload of that module.

Function ZwClose() in ntdll.dll.

Best way is to just kill the process that's using it though.

You can also close handles and mutexes with SysInternals Process Explorer.

 

[oog]

you could try forcedel, but i will echo other people's suggestion of using the SysInternals process explorer to find out what process is holding onto the file.

 

[Malak]

Unlocker you mean?

 

[Markbnj]

Unix has had the best of both worlds for years now. When you delete a file the directory entry is removed so it can't be used by anything else and the on-disk data is freed as soon as the last file handle is closed.

I agree that is a better solution now. At the risk of opening a whole new can of worms, though, most of what people consider weaknesses in the design of Windows are the result of maintaining nearly complete backwards compatibility across twenty years of operating system releases. The original operating systems were weak, and many of those weaknesses have been passed along so old code wouldn't break. It's both a vulnerability, and a large part of why Windows has been commercially successful.

 

[xtknight]

Originally posted by: Malak
Unlocker you mean?

That's a great prog.

 

[Malak]

Originally posted by: xtknight

Originally posted by: Malak
Unlocker you mean?

That's a great prog.

Majorgeeks is a great site.

 

[Nothinman]

I agree that is a better solution now. At the risk of opening a whole new can of worms, though, most of what people consider weaknesses in the design of Windows are the result of maintaining nearly complete backwards compatibility across twenty years of operating system releases. The original operating systems were weak, and many of those weaknesses have been passed along so old code wouldn't break. It's both a vulnerability, and a large part of why Windows has been commercially successful.

So? Most unixes maintain backwards compatibility with all of their previous releases and at least source level cross-compatibility in a lot of areas with other unixes.

 

[Markbnj]

So? Most unixes maintain backwards compatibility with all of their previous releases and at least source level cross-compatibility in a lot of areas with other unixes.

That hasn't been true in my area dealing with enterprise applications. Very often a new Sun release broke apps and forced rebuilds. Same with HPUX or AIX. I don't have a lot of experience with the BSDs of the world. But when I say backward compatibility in Windows I mean that programs written 15 years ago can in 99% of cases be dropped onto a current SP2 version of Windows XP and run without any changes. I have a ton of old Windows 3.1 and DOS stuff that I wrote years ago that still runs fine under XP.

 

[Nothinman]

I'm sure everyone could pick an example of some app that failed to run after some upgrade or config change, the bad things stick out in people's mind more than the good. And IME 'enterprise' apps are usually the worst engineered. They're usually only certified to run under one configuration at one patch level and if you change it and it breaks there's nothing support can do to help you. Heaven help you if you want to put two 'enterprise' apps on the same box because they all seem to be certified against different patch levels and configurations. Hell just recently we had Oracle telling us that we had to change kernels on a Linux box in order to fix a problem we were having, turns out the problem was that /etc/sqlnet.ora file was corrupt. So much for enterprise support.

I have a ton of old Windows 3.1 and DOS stuff that I wrote years ago that still runs fine under XP.

And there are games that were written 5 years ago that don't run properly in XP, it's not like MS has perfect compatibility that no one else can match.

 

[Markbnj]

I didn't say perfect, but I don't think it is arguable that MS has maintained broader backward compatibility across a much wider array of applications, over twenty years, than any version of Unix that has ever been. That compatibility has often constrained what they could do with the operating system, and it is also unarguable that most commercial Unix vendors have never felt so constrained. I sat in an audience fifteen years ago and heard Bill Joy say that Sun considered the maximum lifecycle of an application to be five years. Those of us who were still running ten year-old core systems could only shake our heads. Of course, no flavor of Unix has ever been a consumer market O/S, and that means they haven't had to strive in this direction near as much.