Re: File Corruption copying to a DMG
Re: File Corruption copying to a DMG
- Subject: Re: File Corruption copying to a DMG
- From: Dave Camp <email@hidden>
- Date: Thu, 7 Oct 2010 17:19:04 -0700
What are you looking for people to say? If you have code that you think illustrates a bug, you should file a bug.
Beyond that, your email doesn't have much in the way of details about the code you are using, how you are making the DMG, if you've reproduced it on multiple machines and multiple different drives, etc...can you post more details?
Dave
On Oct 7, 2010, at 3:27 PM, Steve Gehrman wrote:
> No one has any thing to say about this data destroying bug in the OS?
>
> I'm simply copying an application to a dmg, and doing a filemerge on the original and the copy on the dmg and it randomly finds a corrupted file. I'm ejecting the dmg and remounting it which may or may not be related. Normally I've been making a dmg, the installing it and then realize that the install is corrupted and I go back and compare the original to the contents of the dmg.
>
> -steve
>
> On Oct 7, 2010, at 10:48 AM, Eric Tamura wrote:
>
>> If you have a reproducible case of filesystem corruption, whether in file payload or file metadata, please file a radar so that we can investigate.
>>
>> --Eric
>>
>> On 7 Oct 2010, at 5:00 AM, Steve Gehrman wrote:
>>
>>>
>>> I've been seeing some random corruption of files when copying to a DMG. At first I thought maybe my DMG was corrupted, so I created a new one, and I still get random corruption. Looking in a hex editor, I can see that the files size looks OK, but at some point the contents become garbage, or a recognizable contents of another file.
>>>
>>> One clue is that in the cases I've seen, it's always a single file that it bad, not multiple files. So that might help diagnose it.
>>>
>>> Anyone have ideas on what is going on? One thought was that maybe the DMG wasn't being flushed out to disk when it's ejected. I'm using FSEjectVolumeAsync to eject the DMG when it's done copying. Do I need to unmount it first and then call FSEjectVolumeAsync to be safe?
>>>
>>> I've been seeing this off and on for at least a year, but I wasn't sure what was causing the corruption until I was able to narrow it down to the copy to the DMG.
>>>
>>> Any comments?
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Filesystem-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden