Cannot copy files with archive a to file with archive bit ah
Cannot copy files with archive a to file with archive bit ah
Hi Igor,
SI 1644 cannot copy files with archive bit "a" to a destination file with archive bit "ah".
Also SI does not copy sparse files, it copies the entire file in its full length, which can cause some real surprises (like suddenly having 10GB disk space taken, because your virtual machine's 10 GB win drive has changed from sparse to real).
Could you please fix this?
Thanks.
Kind Regards,
Brahman
SI 1644 cannot copy files with archive bit "a" to a destination file with archive bit "ah".
Also SI does not copy sparse files, it copies the entire file in its full length, which can cause some real surprises (like suddenly having 10GB disk space taken, because your virtual machine's 10 GB win drive has changed from sparse to real).
Could you please fix this?
Thanks.
Kind Regards,
Brahman
Hi Igor,
here you can find the source code of a NATIVE (NOT cygwin!) version of cp.exe, which can copy sparse files:
http://unxutils.sourceforge.net/
Regards,
Brahman
here you can find the source code of a NATIVE (NOT cygwin!) version of cp.exe, which can copy sparse files:
http://unxutils.sourceforge.net/
Regards,
Brahman
Hi Igor,
great, it really looks like you are on the ball!
While you are already working on this, maybe you could also look into enabling the duplication of an existing hard linked file structure on the target.
The same utility mentioned above (cp) can do it and rsynch also does a great job with hard links.
Right now I need to run
"rsync.exe -aHtiv source target "
to duplicate my hard link structure after synching with SI.
Since I use hard links extensively, it would be a very nice timesaver to have that ability too.
Have a great Monday!
Regards,
Brahman
great, it really looks like you are on the ball!
While you are already working on this, maybe you could also look into enabling the duplication of an existing hard linked file structure on the target.
The same utility mentioned above (cp) can do it and rsynch also does a great job with hard links.
Right now I need to run
"rsync.exe -aHtiv source target "
to duplicate my hard link structure after synching with SI.
Since I use hard links extensively, it would be a very nice timesaver to have that ability too.
Have a great Monday!
Regards,
Brahman
Well, creating new hard link is not a problem. What is a problem - there is no way to tell if this file is just a hard link. So if you know it is - you can copy it. But for software there is no way to distinguish between hard link and plain file. The only thing that I can think of is some kind of dictionary that user can fill, telling that this specific file(or masked files in folder) is a hardlink and should be copied as hardlink.
If that were true, then the two apps I mentioned (CP and rsync, but especially rsync) could not do what they do, namely duplicating an existing hard link structure from source to target without being specifically pointed to the hard file or told they are copying hard links and without any additional user interaction.
Here is a screenshot of a completed rsync copy operation using my special "SynchroTestBed" where I combined difficult copy/synching situations to see which synching program is capable of doing what.
On the top is my source, the bottom part is the target after the copy operation with rsync (command line on bottom) is finished. I color coded with TC several file types (like sparse, hard, and soft links) and where you see a "2" in the "HrdLnks" columns, it is a hard linked file. In this picture you can see that that rsync successfully copied the hard linked files AS HARD LINKS even when I gave them different names (the hard link counterparts are in the compressed subfolder)! So when TC can show that a file is a hard link and rsync can duplicate them, then there is a way to tell.
Also LinkShell Extension is another native Win32 tool that can duplicate hard link structures from source to target (scroll far down to section "Smart Copy").
I am sure if they can do it, you can do it too, because you are at least as good as those guys.
So I am crossing my fingers and hope you will find a way to implement it ...
Kind Regards,
Brahman
Here is a screenshot of a completed rsync copy operation using my special "SynchroTestBed" where I combined difficult copy/synching situations to see which synching program is capable of doing what.
On the top is my source, the bottom part is the target after the copy operation with rsync (command line on bottom) is finished. I color coded with TC several file types (like sparse, hard, and soft links) and where you see a "2" in the "HrdLnks" columns, it is a hard linked file. In this picture you can see that that rsync successfully copied the hard linked files AS HARD LINKS even when I gave them different names (the hard link counterparts are in the compressed subfolder)! So when TC can show that a file is a hard link and rsync can duplicate them, then there is a way to tell.
Also LinkShell Extension is another native Win32 tool that can duplicate hard link structures from source to target (scroll far down to section "Smart Copy").
I am sure if they can do it, you can do it too, because you are at least as good as those guys.
So I am crossing my fingers and hope you will find a way to implement it ...
Kind Regards,
Brahman
To show information about links and junctions in TC:
http://flint-inc.ru/eng/info/nlinfo.html
To easily work with links in TC:
http://flint-inc.ru/eng/info/ntfslinks.html
The developer is a Russian (of course ).
Re. copying HLs:
Yes, I thought about that too. When experimenting I figured that rsync, which I use to copy the hls, duplicates the existing hard link structure exactly as it appears to the target, but NOT LINKING the two copies.
If it were to copy hls AS hls, you would have all files hard linked back to each other (even though not hled now, i.e. suddenly in column 1 HL would become everywhere on both sides 2 HLs) - this is what the ntfslinks utility above does.
So it works like duplicating a drive to another drive even if you do the operation within the same drive (respecting NTFS 3.01 restrictions).
Regards,
Brahman
http://flint-inc.ru/eng/info/nlinfo.html
To easily work with links in TC:
http://flint-inc.ru/eng/info/ntfslinks.html
The developer is a Russian (of course ).
Re. copying HLs:
Yes, I thought about that too. When experimenting I figured that rsync, which I use to copy the hls, duplicates the existing hard link structure exactly as it appears to the target, but NOT LINKING the two copies.
If it were to copy hls AS hls, you would have all files hard linked back to each other (even though not hled now, i.e. suddenly in column 1 HL would become everywhere on both sides 2 HLs) - this is what the ntfslinks utility above does.
So it works like duplicating a drive to another drive even if you do the operation within the same drive (respecting NTFS 3.01 restrictions).
Regards,
Brahman
Here you can download my TC config file to replicate that column setup:
http://www.box.net/shared/nlbay0g0ny
And here you can download my Synchronization Testbed (mount it with Drive Snapshot - the trial version allows mounting even after trial ends - and you can copy out of it to your hearts content ):
http://www.box.net/shared/bpgsbbj1aq
Kind Regards,
Brahman
http://www.box.net/shared/nlbay0g0ny
And here you can download my Synchronization Testbed (mount it with Drive Snapshot - the trial version allows mounting even after trial ends - and you can copy out of it to your hearts content ):
http://www.box.net/shared/bpgsbbj1aq
Kind Regards,
Brahman