Backup To Zip Problem

I’m running Scrivener for Windows beta 1.6 on an up-to-date installation of Windows 7 64-bit and I am experiencing the same issues described in this post: [url]https://forum.literatureandlatte.com/t/backup-project-to-zip-problem-resolved/8371/1]

To sum again, the zipped backup files contain 0-byte files with the same name of all the vital folders, thereby preventing them from being unzipped (without deleting said blank files). While the original poster solved the problem with Windows Vista’s Service Pack 2, such a solution does not exist in my case. There are no updates available for my installation of Windows.

EDIT: As an addendum, if the folder in question is empty, only the blank file exists in the zip file.

Microsoft seem to cover this issue here support.microsoft.com/kb/952249

"This issue occurs in special .zip files. If the .zip file contains folder items that do not have the directory flag set in the external file attributes field, the computer treats the items as files that are 0 bytes. Therefore, the issue occurs.

In Windows Vista SP1 and in Windows Server 2008, the computer parses files first. The 0-byte files are created before the folders. This behavior makes the extraction fail because the folder structure cannot be created."

So it was solved in Vista SP2 but maybe is still a problem in W7 64bit?

More concerning for me is why is Scrivener creating a 0 byte file? Especially if it is a backup and the restore fails.

Also what is a “special .zip file” and could scrivener not just create a normal zip file?

According to your quote, it’s not. But unzip utilities treat it as such because the folder items are not flagged as being directories (though I think it’s also partly the fault of said utilities for not being able to figure it out from the fact that they have files inside). No doubt there’s part of the zip architecture that I’m unfamiliar with as to this particular behaviour.

I have to admit, Microsoft’s wording here is peculiar. Until this point, I was under the assumption that a zip file is a zip file is a zip file. I’ve never before seen this behaviour before and it begs the question of what manner Scrivener is using to produce its zip files. Without knowing that, I couldn’t say one way or the other.

There are a lot of different compression methods and other factors that could change a .zip file, but I suspect it’s just poor wording. What makes these files “special” is that they contain folders without the directory flag set.

This exact thing is happening to me. The automatic .zip backups cannot be unzipped, I get an error and any files that happen to make it through are useless or empty. I can work around it, but I thought it warranted mentioning.

I’m on Windows 7 64-bit with all of the latest updates installed.

DavidCrake–Thanks for the report, though I’m sorry to hear you’re still having trouble with this! Did you have any beta versions installed earlier that weren’t fully uninstalled prior to upgrading to 1.0? Where are your projects and the backups stored? If you go via Windows Explorer, can you right-click on a project’s .scriv folder and create a zipped copy that way which can be successfully extracted?

I too was having the same problem as DavidCrake, but I found a way around it: Open with Explorer.

I use 7-zip (v9.2) and I couldn’t unzip anything made by Scrivener. But I thought about opening the zip with the built-in windows tool, and it works fine.

Hope you find this works for you as well, DavidCrake.