qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-devel] [PATCH][RFC] Add compare subcommand for qemu-img


From: Miroslav Rezanina
Subject: Re: [Qemu-devel] [PATCH][RFC] Add compare subcommand for qemu-img
Date: Thu, 2 Aug 2012 01:19:20 -0400 (EDT)


----- Original Message -----
> From: "Eric Blake" <address@hidden>
> To: "Miroslav Rezanina" <address@hidden>
> Cc: address@hidden, "Paolo Bonzini" <address@hidden>
> Sent: Wednesday, August 1, 2012 3:21:03 PM
> Subject: Re: [Qemu-devel] [PATCH][RFC] Add compare subcommand for qemu-img
> 
> On 08/01/2012 04:03 AM, Miroslav Rezanina wrote:
> > This patch adds compare subcommand that compares two images.
> > Compare has following criteria:
> > - only data part is compared
> > - unallocated sectors are not read
> > - in case of different image size, exceeding part of bigger disk
> > has to be zeroed/unallocated to compare rest
> > - qemu-img returns:
> >    - 0 if images are identical
> >    - 1 if images differ
> >    - 2 on error
> > 
> > Signed-off-by: Miroslav Rezanina <address@hidden>
> > 
> 
> > +++ b/qemu-img.c
> > @@ -96,7 +96,9 @@ static void help(void)
> >             "  '-a' applies a snapshot (revert disk to saved
> >             state)\n"
> >             "  '-c' creates a snapshot\n"
> >             "  '-d' deletes a snapshot\n"
> > -           "  '-l' lists all snapshots in the given image\n";
> > +           "  '-l' lists all snapshots in the given image\n"
> > +           "Parameters to compare subcommand:\n"
> > +           "  '-g' Second image format (in case it differs from
> > first image)\n";
> 
> As written, this sounds like:
> 
> No -f, no -g => probe both
> -f, no -g => -f applies to both
> no -f, -g => probe first, use -g for second
> -f, -g => use given formats for both
> 
> Is that really what you meant, or do we actually get:
> 

Yes, this is what I meant.

> -f, no -g => -f applies to first, probe second
> 
> I think both interpretations could make sense, but I'd prefer having
> the
> omission of -g imply probing the second file type regardless of the
> presence or absence of -f, for consistency.
> 

I was evaluating both approach. For me only -f means 'Use this format for 
input'.
In case of compare input is equal to both files so the -f value will be used for
both of them. There's only one subcommand allowing to specify multiple input 
files - convert 
and it use -f value for all input commands. However, unlike the compare it does 
not allow
to specify different format for different files and has variable number of 
input files.

So I decided to use -f as specification for both files. I do not have problem 
with using
it for first image only but prefer this handling. I can switch handling in next 
version in case
of more votes for it.

> --
> Eric Blake   address@hidden    +1-919-301-3266
> Libvirt virtualization library http://libvirt.org
> 
> 

Regards,
Miroslav Rezanina



reply via email to

[Prev in Thread] Current Thread [Next in Thread]