HTTP condition upload? It's stupid noone has thought of this yet

[vailixi]

I was just uploading files to drive and I was a little annoyed that my entire upload stopped to ask if I wanted to overwrite a duplicate file. The fact that firefox knows it's a duplicate file makes this a trivial point in programming.

Since it knows the file is duplicate it should know what files are already on the server. If the file you are uploading is duplicate it should wait until the end to ask what you want to do with the file and upload the unique files first that way the entire download isn't stalled until you manually click the dialogue box asking if you want to overwrite the file or skip it. Just skip the file until the end of the upload. If you don't click the dialogue box within X seconds it just defaults to only uploading the unique files.

I should code this option into a browser object and make it work just as an example. I'm just not really sure where to look for conditional http request. Anybody know of an arcane network programming book that covers topics like this? Better yet is there an implementation of this already somewhere that I can borrow code base from?

Anyway what do you guys thing about this idea?

Edited December 8, 2016 by vailixi

[digip]

Does it claim duplicate on file name, or actual MD5 sum or such? How is it sure it identical, and, if downloading the same file, could the new one be an updated version with more data, say a document file, log file, etc. These are things to think about and look into. FireFox is supposed to be opensource though, so you could probably find a Mozilla forum for FF dev's that would know where to look or how to hook into it, but I have a feeling it's not OS agnostic in that the operating system libraries dictate this vs the browser, but I could be wrong.

[0phoi5]

I feel this way about many, many installs, updates and file transfers; Ask me questions at the end, not part-way through, when I've gone to bed or left the PC on for the day, only to find the process I was trying to complete stalled after 5 minutes to ask a stupid question.

Edited January 26, 2017 by haze1434

[digip]

3 hours ago, haze1434 said:

I feel this way about many, many installs, updates and file transfers; Ask me questions at the end, not part-way through, when I've gone to bed or left the PC on for the day, only to find the process I was trying to complete stalled after 5 minutes to ask a stupid question.

I do a lot of syncing for backups on a remote server. File versioning takes care of overwrites in the cloud side of this. However, I have to copy files from the source folders, to my sync folder(I duplicate files locally in several places as well), and one thing I know from this, is that when it finds a duplicate, I DO want to overwrite, as I make updates to the files, such as PSD's that contain new layers, etc. This is why just stating its a duplicate does not always mean "identical", even if file size is the same. Save data and creation dates would be something you want to look at, as well as file size, but for me, I just check the box for overwrite all since I know i want to update with the newer files. Any upload system that DOENS'T prompt you to overwrite, is not doing it's job(other than cloud syncing which is automatically saving to same file path and handling backups on that end). Downloads are the same. If you don't want the prompt, save as a new file name before downloading and sort your shit later, but I'd rather have a prompt to warn,especially if what I download, is an outdated file.

[0phoi5]

23 hours ago, digip said:

I do a lot of syncing for backups on a remote server. File versioning takes care of overwrites in the cloud side of this. However, I have to copy files from the source folders, to my sync folder(I duplicate files locally in several places as well), and one thing I know from this, is that when it finds a duplicate, I DO want to overwrite, as I make updates to the files, such as PSD's that contain new layers, etc. This is why just stating its a duplicate does not always mean "identical", even if file size is the same. Save data and creation dates would be something you want to look at, as well as file size, but for me, I just check the box for overwrite all since I know i want to update with the newer files. Any upload system that DOENS'T prompt you to overwrite, is not doing it's job(other than cloud syncing which is automatically saving to same file path and handling backups on that end). Downloads are the same. If you don't want the prompt, save as a new file name before downloading and sort your shit later, but I'd rather have a prompt to warn,especially if what I download, is an outdated file.

I agree, but I feel the point here would be to ask the user at the end of doing all other file transfers, rather than part way through.

So, in moving 1000 files from one location to another, where file number 200 has a duplicate in the destination; instead of transferring 199 files and then stopping, transfer 999 files and then ask the question. Would be so much better.

Linux apt-get upgrade does this as well. Why install half of all upgrade components and then ask a question (such as do you want the cron service to restart?). Why not complete all actions that are possible, and then ask the questions at the end?

I would find this a blessing, as I can think of at least 24+ hours wasted where I thought a process would be complete by the morning, only to find it had completed 2% and then asked a question and stopped.

Edited January 27, 2017 by haze1434

[digip]

12 hours ago, haze1434 said:

I agree, but I feel the point here would be to ask the user at the end of doing all other file transfers, rather than part way through.

So, in moving 1000 files from one location to another, where file number 200 has a duplicate in the destination; instead of transferring 199 files and then stopping, transfer 999 files and then ask the question. Would be so much better.

Linux apt-get upgrade does this as well. Why install half of all upgrade components and then ask a question (such as do you want the cron service to restart?). Why not complete all actions that are possible, and then ask the questions at the end?

I would find this a blessing, as I can think of at least 24+ hours wasted where I thought a process would be complete by the morning, only to find it had completed 2% and then asked a question and stopped.

In the case of linux apt-get, might be dependencies in certain order, file a needs to be installed before file b, but for most cases, I gather it's first come first serve, in that what it sees and downloads first has no bearing on an order of items such as, let me scan all remote files first, then see if I have the same file locally. You won't know, until it's downloaded, unless you have a versioning system with a separate file of say, MD5 sums or hashes of the app/program/file to compare to at both ends, how does one system know the remote file is the same other than by name alone, it would need to download it first to know content is a duplicate.