Giving up on Vista

littleguru wrote:



Sven Groot wrote:  littleguru wrote: Back it up - hide that from the user! Make it better. Find ways... Congratulations! You've just reinvented System Restore.

soooo... why doesn't this work on his side?

Probably because the back up also got corrupted.

EDIT: A second copy is exactly what FAT did with the file tables. It's not a good solution. When the backup and the real copy don't match, how do you tell which one is corrupt?

littleguru wrote:



Bas wrote:  littleguru wrote: Another point is that this registry thing needs to fixed in one way or another. It can't be that you need to reinstall your whole system, because one component failed. This is not possible!! Isn't this kind of unavoidable? I mean, there's always going to be an essential component.

Back it up - hide that from the user! Make it better. Find ways...

Backing it up might work.. but then you'd run into differences between your restored registry and what happened on the rest of your harddisk in the mean time. System restore isn't that great for that reason either.

Hiding it from the user would just result in threads by some people here going "Windows is not allowing me acces to my own computer!"

Making it better though... I agree completely. As long as we don't end up with endless amounts of xml files in thousands of subfolders.


Anyway, the real answer to this, of course, would've been WHS. Because then you'd just boot from a CD and return your PC into the exact state it was in a day before the registry got corrupted.

Correct me if I'm wrong, but didn't Phreaks already try reinstalling Vista at some point, and it didn't help? Or am I confused with someone else?

phreaks wrote:



Sven Groot wrote:  phreaks wrote: I can't convey how utterly disappointed I am with Vista. I'm sorry you feel that way, especially since my own experience has been so completely different.

Yea, I guess its hit or miss.

I suspect something went wrong with my install, because its been one issue after another and it's really unfortunate that it has gotten to this point; where it just doesn't make sense for me to continue losing time trying to work around registry rights issues.

I've tried hard to solicite feedback from MS regarding my problems using various avenues; but they just seem to ignore me.

After about 3 months, I think I've put enough of an effort in, I really wanted to like Vista.

All my hardware is Vista Approved, the one piece that wasn't, I replaced the day after I installed.

Why not reinstall Vista instead of reinstall XP? If your installation of Vista is broken, this would be more logical than downgrading. Otherwise you might see yourself upgrading to Vista again in a couple of months if SP1 delivers on the promises.

So far I've only had one install out of three show instability, and this was due to; what I believe to be an instability in my graphics card that got more pronounced with the 'virgin' graphics drivers of Vista..

You noted that this was multiple machines you were going to revert, that seems extremely odd that you have so many issues.. I would perhaps worry about the interactions with the OS, are you doing a lot of tweaking?

All the same specification? perhaps its a hardware combination issue?

littleguru wrote:

I don't get this... really. There is tons of people around who are doing backups of databases on a daily base. Why shouldn't this be different with the registry. But let's say Windows does that on a second partition (which might be tiny) or to another file. Or even to a set of files, like you do with database backups.

But that's what system restore does. It still doesn't help if the corruption has crept into the backups too. Or if the last correct backup is so old that you might as well reinstall.

And I believe Windows does keep at least one more copy of at least the boot driver configuration around, which is used for the "Last known good configuration" boot option.

littleguru wrote:

Shouldn't the system be aware if something went wrong and immediately restore?

Explain to me how it knows something went wrong? The only thing I can think of is a CRC for the registry, which would need recomputing every time the registry is changed. That'll be nice for performance.

Sven Groot wrote:



littleguru wrote: Shouldn't the system be aware if something went wrong and immediately restore?

Explain to me how it knows something went wrong? The only thing I can think of is a CRC for the registry, which would need recomputing every time the registry is changed. That'll be nice for performance.

Well, you could do it for each element, and then if you read a corrupt element you fire a "whoa! it's busted" event.

Which would be slightly less of a headache than the entire registry.

I'm wondering what, exactly, can go wrong in the registry. The only thing I often hear is that it's 'messed up' or 'corrupted'. I'm wondering what exactly is going on, and why. Why didn't this happen with all those separate .ini files, for instance?

I think its just referring to reg entries state being left in a bad way.. Imagine I had a program I wrote as a service, and it needed to write a registry update when it closed, and I'd stupidly made the program utterly rely on this, and one day the computer crashed or whatever, now the service doesn't work until someone manually repairs the bad (or lack of) entry in the registry..

I'm sure it's possible for a mass corruption on the underlaying file, but I think in most cases its poor design choices that rely too much on the registry..

Plus, It occurred to me recently just how bad the registry can be on a system, when I was thinking about 'Wouldn't it be nice to have an app installed on a portable harddrive, and move it from system to system when I wanted to use it?'..

Self contained apps would be a nice thing, but I think there's a balance to be had for the sake of the systems awareness and such..

Bas wrote:

Why didn't this happen with all those separate .ini files, for instance?

It most certainly happened there as well!

Sven Groot wrote:



Bas wrote: Why didn't this happen with all those separate .ini files, for instance?

It most certainly happened there as well!

But what? What exactly happened?

Massif wrote:



Well, you could do it for each element, and then if you read a corrupt element you fire a "whoa! it's busted" event.

Which would be slightly less of a headache than the entire registry.

How do you checksum a DWORD?

AndyC wrote:



Massif wrote:  Well, you could do it for each element, and then if you read a corrupt element you fire a "whoa! it's busted" event. Which would be slightly less of a headache than the entire registry.

How do you checksum a DWORD?

Approximately. (you could... gasp... just use a checksum bit (which would at least catch half the corruptions). Heck you could just store the damn thing twice and then if they differ you know something's gone wrong, and it's time to look for a backup. I'm not suggesting you make it such that you can restore from the check data or anything.)

Massif wrote:



Approximately. (you could... gasp... just use a checksum bit (which would at least catch half the corruptions). Heck you could just store the damn thing twice and then if they differ you know something's gone wrong, and it's time to look for a backup. I'm not suggesting you make it such that you can restore from the check data or anything.)

But that really doesn't work, because it's impossible to know whether it's the checksum that's corrupt or the data itself. It's just too small a datastructure to try and implement that sort of validation. Not to mention the actual overhead of trying to compute that in any reasonable way.

What's more, it's not a very effective strategy. Registry corruption is usually caused by one of two things:

1) A crash during read/write operations, leading to an on-disk corrupt structure. At that point you're looking a significant overhead on every read/write operation for the sake of a tiny possibilty of data recovery in the event of a crash.

2) Applications writing invalid data. No amount of checksumming helps there because what's getting written is technically what the application wants to write. Enforcing type-safety on the registry might help, but again it's a lot of overhead when applications are supposed to be doing the right thing.