Take the 2-minute tour ×
Programmers Stack Exchange is a question and answer site for professional programmers interested in conceptual questions about software development. It's 100% free, no registration required.

Do sync programs like Dropbox typically track file changes by doing byte by byte comparisons, or using hashes, or using diff / keeping local commit logs like version control, or what?

share|improve this question

3 Answers 3

Ultimately to compare files you need to compare every byte - how else would you notice a single byte change?

In reality you read blocks of bytes and compute a hash value, you then check against a list of hashes. A good example is "rsync"

As far as I know dropbox only dedupes entire files, so will compute a hash of the entire file to check fro the same file

share|improve this answer
1  
pst on some file systems there's a date modified –  ratchet freak Oct 2 '12 at 20:08
1  
Wouldn't hashing create a small but real risk of collisions, resulting in a file not being synced? Dropbox apparently uses a diff-like implementation. dropbox.com/help/8/en –  mcandre Oct 2 '12 at 20:47
1  
@mcandre that sounds like it'd be better of as its own question. –  GrandmasterB Oct 2 '12 at 21:07
    
@ratchetfreak: date modified, on some systems, isn't necessarily reliable for this kind of problem. A simple touch would cause the date modified to be different, where a sync may not actually be required. –  Steve Evers Oct 2 '12 at 22:28
1  
@SnOrfus then double check the changes when the date modified gets changed –  ratchet freak Oct 2 '12 at 22:33

On Windows there's a mechanism to have the OS alert you when there's a change to a 'watched' directory structure - FindFirstChangeNotification(). When that indicates a file has changed, an application can then go about comparing files in the changed directory to find the actual files that have changed by looking at size, modified date, hash, etc.

This (as Michael points out below) is something that each platform would provide in some manner. I wasnt saying this sort of thing was unique to Windows.

share|improve this answer
1  
And there's inotify on Linux. –  Michael Oct 2 '12 at 20:22
    
But what if the sync program wasn't running at the time of the file change? Wouldn't it miss the opportunity of being notified by the hook? –  mcandre Oct 2 '12 at 20:46
    
You used the specific example of dropbox, which generally is running in the background. Obviously if the program isnt running, it wont get notified. Then it has to use other methods (modified date, perhaps). I dont think you're going to get a concrete answer about what the specific methods used ARE, because different programs do things differently. Better to ask the creators of the specific programs in question. –  GrandmasterB Oct 2 '12 at 21:03
    
@mcandre Pretty much what GrandmasterB said - if you missed the notification, you'd have to scan the folder. Depending on how "accurate" you want to be, this can mean simply looking for new files and modified timestamps / file sizes (these are relatively inexpensive to do), or in the worst case, comparing the entire file. Programs like rsync typically hash the file in chunks, so changes early on in the data can be detected earlier, but in the worst case (files are identical), you're going to read the whole thing in. –  Daniel B Oct 3 '12 at 6:57

.NET for instance has a FileSystemWatcher class. I'm sure other low-level languages and runtimes can provide similar capabilities.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.