Ehtan,
On 10/30/09 00:36, Ethan Quach wrote:
> Sriman,
>
>
> be_utils.c
> ----------
> 2908 - Is this correct zfs terminology? I thought by definition
> a snapshot is a dataset. So perhaps this should be changed to:
>
> Target snapshot/filesystem/volume already exists.
>
> or just
>
> Target dataset already exists.
I don't have much idea about zfs terminology. But, after browsing
a bit through onnnv source, it looks like your argument is correct.
/usr/src/lib/libzfs/common/libzfs.h:64
EZFS_EXISTS, /* pool or dataset
already exists */
/usr/src/lib/libzfs/common/libzfs_util.c:86
case EZFS_EXISTS:
return
(dgettext(TEXT_DOMAIN, "pool or dataset exists"));
I don't think pool creation is relevant in beadm code.
So, I changed the message as "Target dataset already exists."
>
>
> libbe.h
> -------
> Per the comment at 67, can you move this to the end of the enum
> and make the corresponding change to messages.py
I changed it as per your suggestion. But, I would like to understand the
rationale behind the comment at line 67. Why is it necessary to add messages
at the end of enum to keep them in sync.
Is it not Ok if we add new messages at same position in both libebe.h
and messages.py?
I am asking this for my own understanding.
Updated Webrev : http://cr.opensolaris.org/~sb229247/caiman-8006/
Thanks,
-Sriman
