[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-devel
Subject:    Re: PATCH: KTar Does Not Recognize Differently Enabled Checksums
From:       Robert Hogan <robert () roberthogan ! net>
Date:       2006-07-17 19:11:56
Message-ID: 200607172011.56728.robert () roberthogan ! net
[Download RAW message or body]

On Monday 17 July 2006 09:12, Brad Hards wrote:
> On Monday 17 July 2006 08:00, Robert Hogan wrote:
> > Done. The patch wouldn't hurt anyway though, would it?
>
> I haven't reviewed the patch or tested it yet.
>
>  My concern is that there may be something else that the tar variant does
> that might cause breakage / data loss under some circumstances. So I'd like
> to understand what else is changed.
>
> Brad

Just had a look at tar itself and it has a neat little function (from_header) 
for iterating through a checksum value and pulling out the checksum itself 
while ignoring leading zeroes and spaces. There's quite a lot of /* 
Accomodate buggy this and that*/ in it - quite a fraught little world that of 
tar checksums!

Given that there is a lot that can go wrong with a tar checksum would KTar be 
better off taking a best-effort approach to untarring files (in the absence 
of completist workarounds for other tar implementation's bugs)? Rather than 
giving up at the first dodgy checksum? A warning to stdout might suffice. Or 
a message box to the effect 'sort of worked' - inspect unrolled contents to 
see if they're usable?

It's worth nothing that for the dante tarball, tar itself takes this approach: 
it complains about the tarball but unrolls it anyway and creates a dante tree 
that compiles and works.

[The reason it complains is different to the reason KTar complains: it doesn't 
like the absence of an overall directory header I think. It also fails to 
match the first checksum: the tarball and KTar calculate 11327; tar 
calculates 11326.]

-- 

KlamAV - An Anti-Virus Manager for KDE - http://www.klamav.net
TorK   - A Tor Controller For KDE      - http://tork.sf.net
 
>> Visit http://mail.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<
[prev in list] [next in list] [prev in thread] [next in thread]