Grumble about re-doing failed update

Hi folks.

Even in a VM if updating a new build via WU and it fails for something silly like non supported CPU / graphics / no TPM etc etc (all of which you can change via the "Virtual Hardware" in the VMs configuration on most VM platforms) -- on retry the update process re-downloads the whole update again -- surely after fixing the error and doing the "retry" the system shouldn't have to re-download the whole dogs dinner again -- must be really annoying for those on slow Internet --I'm lucky there. But surely are developers that bonkers -- anybody who has any experience with computers - even 5 mins worth want to be able to re-start a failed process without going right back to the start where possible. I can understand if you run out of disk space - but that's obvious enough - but for the other stuff --especially in a VM re-download of the entire update shouldn't be necessary.

Imagine if you get a glitch many pages through a PhD thesis and you'd have to start the whole thing from scratch again !!!!.

Cheers
jimbo

I think the reasoning is that... the cause of the failure may also be the actual download.
So... when trying to fix/redo the update... it's best to redo it all.

Ghot said:

I think the reasoning is that... the cause of the failure may also be the actual download.
So... when trying to fix/redo the update... it's best to redo it all.

Click to expand...

Sorry -- I don't buy that one -- if the failure is something like "Not supported CPU, no sec boot, no TPM etc" then that's definitely not a failed download. In any case wouldn't a serious developer on a project where 100,000's of users or more are going to use the facility simply after download run a verify and check the SHA1 or whatever the key sums are used these days are correct before starting the install.

Cheers
jimbo

jimbo45 said:

something like "Not supported CPU, no sec boot, no TPM etc

Click to expand...

Well... even those failures could be caused by a bad download.
Errors in a download could hit anywhere.

The people that code the updates.. have to take that into consideration, I would think.

Your post reminds me of the days when I still had dial-up and any failed download had to start all over. The worst was when trying to download a new version of Mandrake Linux. It took days to download and several times I had to start over. Eventually Best Buy started selling the disk.

Not to mention in the time that you ran WU and it downloaded the package(s) it intends to install, one or more of those packages could have fundamentally changed.

However, I agree that it would be child's play to very verify the packages via file hashes instead of downloading the whole thing again, and only download if something has changed.

BITS may be background, but it sure as hell is not Intelligent, in this case.