|
Eddy Vluggen wrote: Not in Windows; replacing requires writing (and locking) to a file that is in use.
If the application starts the uninstaller and then terminates, the application executable is not in use. The uninstaller might also check for other instances running, and present a list to the user, offering the user to close down each instance in a controlled way, or optionally killing those processes forcefully. I am regularly handling several uninstallers that do that.
Now that I look it up, I see that with the delete/rename on reboot, the old file name is valid until reboot time. But at that time it certainly disappears. I mixed it up with another OS I worked with, where you could delete the name of a file to make it temporary: On Close() on a file that has no name, the file is deleted. (So you could make a temporary file by creating it with an empty name.)
So, with the Windows delet/rename at reboot mechanism, there is meaning to the question if you want to reboot now: With some uninstallers, some files may not disappear until reboot. But again: Scores of uninstallers manage to remove all relevant files, including itself an the directory where the uninstaller was residing, without a reboot. They are using other mechanisms. It happens, so it must be possible.
Eddy Vluggen wrote: That's two different things; the progress not being updated by the installer, or Windows not painting because the main-process is busy.
It is obviously possible to set your screen update thread to a very low priority and the worker threads doing the deletion and cleanup to a very high priority. That is a case of "You Asked For It, You Got It". But I wouldn't say that is a recommended way to make your uninstaller, or any program with a progress bar. You give the display update process a high enough priority that it is not significantly delayed - certainly not for minutes, as the original poster experienced - by the background worker process.
Eddy Vluggen wrote: Usually, it doesn't [know the files to delete].
In that case, I would think it could run into problems deleting them. Sure, at entry to "int main(int arc, char**argv)" it doesn't know, but it should create a list of files to be deleted quite early in the process.
Eddy Vluggen wrote: That's exactly what the Windows progressbar shows when copying files and where all those people complain about it moving from 10s to 30m to 15s.
First: The jumping up and down continues throughout, it doesn't stabilize as the uninstaller, or whatever, gains experience with how fast the disk and other system components are (although my gut feeling is that it was much worse in earlier versions of Windows). Second: While there may be some flucutations of the estimate, nothing can justify orders of magnitude jumps up and down! We frequently see estimates jumping up or down by a factor of ten or more, indicating a completely crazy estimate of remaining work.
Eddy Vluggen wrote: Given a FS works, the file may be "all over the place".
The estimate may jump all over the place, from 10 sec to 30 minutes to 15 sec. The files' disc locations doesn't influence that very much (more so on disks thirty years ago, but very little with modern disks).
Eddy Vluggen wrote: Lots of small files take more than than a single file of the same size.
So, as I wrote, there is a time cost per file delete. Lots of small files will each cost the time of a file delete. When the uninstaller digs up the directory entries to create the file lists, it should also be able to see the actual size of a file: Releasing the space for a gigabyte file takes more time than releasing 2 kbytes of file space. This is particularly true on file systems allocating on a page basis (such as DOS and "classical" *nix file systems); on NTFS (and other files with extent allocation), it is more a function of the number of extents than of disk pages. On a fully defragmented disk (every file has only a single fragment) there will be far less difference in time between deleting a 2k and a 1G file, although there still is some. But if the last three files of 1 Gbyte each took half a second to delete, the uninstaller may assume that number four will take roughly as much time. If the last twenty kilobytes files each were deleted in 10 ms, it may assume that the next fifty kilobyte files will require a similar amount of time each.
It seems like some programmers handling progress bars are not thinking at all! E.g. if you are given a "pause" option, and use it, the clock is not stopped when the operation is paused: If half of the files/megabytes were processed in 10 seconds, and then you pause everything for 30 seconds before continuing with the next file, the time estimate is calculated from the 30.01 seconds it took to handle the first file after the resume, completely ignoring that 30 of 30.01 seconds were not used for processing the file.
It is not easy to get a perfect pre-calculation that will be correct within a margin of, say, 5%. But getting within a factor of two should be easy, within 50% definitely also realistic - at least when the operation is 10% on its way.
I made myself an incremental, deduplicating backup system, that handles all my disks of various technologies and interfaces. In the backup management data, I keep information for each individual disk how much time it typically takes to open a file (in various modes), how long it typically takes to read 10 kbyte, 100 kbyte, 1 Mbyte, 10 Mbyte, how long a create or delete takes. After each backup is performed, these figures are adjusted. They might be somewhat off from the "historical" value, e.g. if I usually plug in the USB3 disk in an USB3 connection, but then one day plug it into an USB2 connection. It is detected quite rapidly during the backup that one disk is slower than the management data indicates, so the estimated time for completion is adjusted. The managed value is adjusted somewhat, being set to 75% old value + 25% new value (percentage was arbitrarily chosen) so that it remains fairly stable but will slowly adapt to e.g. new interfaces.
An uninstaller, being run once or maybe a couple of times, with no place to save former experience, cannot easily save historical data the way I do in my backup system. The application could, though: I have been using several applications that can tune itself by running timing tests on the speed of disks, display cards and other components. These data are saved a configuration file (or in the Registry) for use by the application itself, but it could also be used by any other function that displays a progress bar, such as an unistaller.
Getting a good progress indicator, a lot better than what is typical today, is actually not that difficult. Getting a perfect one may be difficult, but we are not asking for that. Only for something that is reasonably close to real world operations.
|
|
|
|
|
Member 7989122 wrote: If the application starts the uninstaller and then terminates, the application executable is not in use.
That's what we usually do, since you can't overwrite the executable while it is in use.
Member 7989122 wrote: In that case, I would think it could run into problems deleting them.
No, since it usually keeps those files in a folder it knows. Updates aren't downloaded to random places.
Member 7989122 wrote: The estimate may jump all over the place, from 10 sec to 30 minutes to 15 sec. The files' disc locations doesn't influence that very much (more so on disks thirty years ago, but very little with modern disks).
It was one example of things influencing speed. The amount of data a sector holds also plays a role; in my own Dokan-based FileSystem, it is visible that even the amount of folders plays a role, since for each "Get Directory contents" command Windows will look for some files it expects there. Also interfering with the correctness of the total time-calculation here is your virus-scanner.
So yes, the estimate is going to jump a lot. It is what it is, an estimate, where people act as if it is a promise.
Member 7989122 wrote: Getting a good progress indicator, a lot better than what is typical today, is actually not that difficult
Go ahead and write an article on how 
Bastard Programmer from Hell 
If you can't read my code, try converting it here[^]
"If you just follow the bacon Eddy, wherever it leads you, then you won't have to think about politics." -- Some Bell.
|
|
|
|
|
OriginalGriff wrote: Or has Corel employed the guy who wrote the original Windows "File Copy" progress indicator?
Corel has left such a bad taste in my mouth - ever since the Windows 3.1 days, when their software would take over your system and they'd replace user controls with their own version because they were so arrogant they honestly believed they were doing a better job of it than Microsoft. If you tried to uninstall their software, odds were pretty good you were left with an unusable system and the only solution was to pave over the entire thing...
I have no idea if they're still true to their roots to this day. I don't ever intend to go out of my way to find out.
|
|
|
|
|
Back in the days of dinosaurs, I bought Paint Shop Pro 5. This was before Corel bought it out. I still use it every day. Later I tried one of the Corel versions  , but hated it so I dumped it.
Rarely I want to do some fancy image editing, but then I use an old version of Photoshop (version CS2).
|
|
|
|
|
I stuck with the JASC Version up to V9 (the last pe-Corel version) for yeas, until it's problems with Aero on Win7 finally moved me to X6 in 2014. (It was just hassle to try and do screen shots: you had to take the shot, run PSP which disabled Aero, save / edit the image, close PSP which re-enabled Aero, take the next shot, ...
I am currently using PSP 2020 due to some excellent special offers over the years! It's pretty good, but not as "focussed" as the JASC version was - Corel is damn good at adding bugs and bloating, damn them.
"I have no idea what I did, but I'm taking full credit for it." - ThisOldTony
AntiTwitter: @DalekDave is now a follower!
|
|
|
|
|
Hi All,
I was wondering if others who had seen shared the opinion there were some ideas that were not developed or fully used. Plus X-Wings are really sealed again moisture ingress (mind you they are space ships...)
Glenn
|
|
|
|
|
Don't you mean "The Rise of Skywalker"? 
Anything that is unrelated to elephants is irrelephant
Anonymous
-
The problem with quotes on the internet is that you can never tell if they're genuine
Winston Churchill, 1944
-
Never argue with a fool. Onlookers may not be able to tell the difference.
Mark Twain
|
|
|
|
|
No, he's just running two years behind the rest of us.
That's about right for Bristol ... 
"I have no idea what I did, but I'm taking full credit for it." - ThisOldTony
AntiTwitter: @DalekDave is now a follower!
|
|
|
|
|
Oi, Howstha, something something sheep! :baa:
|
|
|
|
|
Hey! Most of Wales is still waiting for Windows 98 to be released ...
"I have no idea what I did, but I'm taking full credit for it." - ThisOldTony
AntiTwitter: @DalekDave is now a follower!
|
|
|
|
|
What's wrong with Windows 95?
Why are you youngsters always so fast to jump straight into all sorts of modernities before you know they'll be successful?
Anything that is unrelated to elephants is irrelephant
Anonymous
-
The problem with quotes on the internet is that you can never tell if they're genuine
Winston Churchill, 1944
-
Never argue with a fool. Onlookers may not be able to tell the difference.
Mark Twain
|
|
|
|
|
I'll tell you what was right about win '95: I used to start the computer, wait until the desktop was displayed (which happened sooner in the cycle than it does now), then click everything I wanted open, go off and get coffee, come back, and start working.
Since '98, that hasn't worked. You're lucky if even one of the apps opens.
I wanna be a eunuchs developer! Pass me a bread knife!
|
|
|
|
|
I haven't even seen the last jedi, yet. I got so bored with all the undeserved fuss.
I wanna be a eunuchs developer! Pass me a bread knife!
|
|
|
|
|
Yes I do, the Last Jedi was the train wreck that derailed the latest trilogy.
|
|
|
|
|
glennPattonWork wrote: Plus X-Wings are really sealed again moisture ingress (mind you they are space ships...)
Just made me think of Futurama:
Farnsworth: Dear Lord, that's over 150 atmospheres of pressure.
Fry: How many atmospheres can this ship withstand?
Farnsworth: Well it's a spaceship, so I'd say anywhere between zero and one.
"These people looked deep within my soul and assigned me a number based on the order in which I joined."
- Homer
|
|
|
|
|
|
It's working to contain flu outbreak! (10)
"I have no idea what I did, but I'm taking full credit for it." - ThisOldTony
AntiTwitter: @DalekDave is now a follower!
|
|
|
|
|
I like that one - but can't do tomoz
"We can't stop here - this is bat country" - Hunter S Thompson - RIP
|
|
|
|
|
"I have no idea what I did, but I'm taking full credit for it." - ThisOldTony
AntiTwitter: @DalekDave is now a follower!
|
|
|
|
|
Looks like you are up tomorrow - you've posted much more difficult ones than this and they've been solved - weird 
"We can't stop here - this is bat country" - Hunter S Thompson - RIP
|
|
|
|
|
It also looks like he has forgotten the answer... 
|
|
|
|
|
Nah, I had to take Herself to Hydrotherapy.
"I have no idea what I did, but I'm taking full credit for it." - ThisOldTony
AntiTwitter: @DalekDave is now a follower!
|
|
|
|
|
I have been a Rush fan for more years than I care to remember. I was 14 when I discovered Rush when I picked up their "Exit Stage Left" album. I'm now 50 and have adored this band all that time. I grew up listening to their music and played the drums because of Neil Peart. I have their tattoos on my arms dedicated to them. Gutted to find out he has passed away.
"Suddenly you were gone
From all the lives you left your mark upon...."
#RIPNeilPeart
Rush Drummer Neil Peart Dead at 67 - Rolling Stone[^]
"There are two ways of constructing a software design: One way is to make it so simple that there are obviously no deficiencies, and the other way is to make it so complicated that there are no obvious deficiencies. The first method is far more difficult." - C.A.R. Hoare
Home | LinkedIn | Google+ | Twitter
|
|
|
|
|
Nice guy from what I have heard too 
|
|
|
|
|
Neil was unlike any other rock drummer. He was quiet, softly spoken, introspective, articulate and loved reading books. The complete antithesis of a rock musician. He will be sorely missed
"There are two ways of constructing a software design: One way is to make it so simple that there are obviously no deficiencies, and the other way is to make it so complicated that there are no obvious deficiencies. The first method is far more difficult." - C.A.R. Hoare
Home | LinkedIn | Google+ | Twitter
|
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
