ISO files are not the same

[dmaker2]

I am using version 2.5.8.0.  When I use the “Create image file from files/folders” mode I observe a behavior that I did not expect.  I create an ISO (call it ISO1), then I mount it and do a binary comparison to the source data and it verifies correct.  Then I create another ISO from the same source files and folders (call it ISO2), then mount it and it also verifies to the same source files correctly.  BUT when I hash ISO1 and ISO2 they have different MD5.  I also compared the file attributes between the ISO files and they are the same.  I also checked the case of the files names and folders and they are the same too.  I tried UDF and ISO9600+UDF and ISOs created are not identical.  If I binary compare the ISOs files to each other there is a difference but it appears not to be with the file and folder data but in some other area of the ISO files. I also tried different source files and folders and still they are not the same.  I have the ImgBurn setting to the default values.  I always assumed that ImgBurn would create ISO files that would be identical.   Can anyone help me understand why ImgBurn doesn’t create identical ISO files?

[dbminter]

This is just a guess, but it could be different metadata causing the different contents.  The 2nd ISO's files and contents, though the same, may have different time stamps since those files would have been added later.  I don't know if ImgBurn uses the file metadata on the source device or if it creates its own time stamps based on when the files were added to the ISO.  Again, this is just a guess, and it's pretty much all I've got on what may be happening here.

 

When you said you compared the file attributes, does that mean the date and time stamps are the same for the files in both ISO's?  If they are, then that blows the above theory out of the water, and I've got nothing else.

[LIGHTNING UK!]

I'm pretty sure some fields in the file system are calculated using the current systrem time. So it'll basically never be 100% identical.

What would be the same though is building an iso, burning it and then reading it back to a new iso. It's only the 'building' bit that makes a different iso each time.

[dmaker2]

4 hours ago, dbminter said:

When you said you compared the file attributes, does that mean the date and time stamps are the same for the files in both ISO's?  If they are, then that blows the above theory out of the water, and I've got nothing else.

File attributes, meaning file names, read-only, archive bits, etc.  all are identical to the source.

 

3 hours ago, LIGHTNING UK! said:

I'm pretty sure some fields in the file system are calculated using the current systrem time. So it'll basically never be 100% identical.

What would be the same though is building an iso, burning it and then reading it back to a new iso. It's only the 'building' bit that makes a different iso each time.

Yes, I would agree the difference seams to develop in the building process.  Unfortunately my knowledge of the ISO structure is limited so I can't easily figure out what is the variable element in the ISO to understand this further.  

I guess my question about all this was is this normal behavior when using ImgBurn and I think that is what you are suggesting.  Although to me it I would not expect the system time to play any role in creating the UDF file structure.  What would be the logic to do so?

Thanks for your replies.

 

 

 

[LIGHTNING UK!]

There are timestamps all over the file system - and not just creation / modification ones for files / folders, they're the obvious ones.

Unless you actually want to read the osta.org udf specs and find out exactly where the date and time is used, that's probably all you need to know.