[libvirt] consistency of virsh help output

[Eric Blake]

On 02/06/2011 02:29 AM, Zdenek Styblik wrote:
> I didn't say we should take one way or another; neither I've said I'm
> going to. I merely wanted to point out at differences, which I think,
> create a bit of confusion.
> 
> As suggested in my hasty reply, perhaps we should have a vote or poll.
> It doesn't seem to be easy decision to me.

Well, I'll start by casting my vote :)

> 
> 
>> attach-disk [--domain] <string> [--source] <string> [--target] <string>
>> [--driver <string>]
> 
> 
> This is domain of man pages and might be (the best) understood. But it's
> damn-o-damn long. And although we are way past 80 char terminals, I
> still tend to keep within this limit etc. :)
> 
> 
>> attach-disk <domain> <source> <target> [<driver>]
> 
> 
> When understood, it's ok. And the majority of virsh commands have this
> format; except 19; I would stick with it.

I also like the shorter form.

> 0001-Unify-virsh-help-synopsis-output.patch>From 2ce39caeb82a9a0b2dccc409ac986648c4017191 Mon Sep 17 00:00:00 2001
> From: Zdenek Styblik <stybla at turnovfree.net>
> Date: Sun, 6 Feb 2011 08:50:43 +0100
> Subject: [PATCH] Unify virsh help synopsis output
>  Signed-off-by: Zdenek Styblik <stybla at turnovfree.net>
> 
> ---
>  tools/virsh.c |    4 ++--
>  1 files changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/tools/virsh.c b/tools/virsh.c
> index 1f820e8..cc8efcf 100644
> --- a/tools/virsh.c
> +++ b/tools/virsh.c
> @@ -10596,11 +10596,11 @@ vshCmddefHelp(vshControl *ctl, const char *cmdname)
>                  case VSH_OT_INT:
>                      /* xgettext:c-format */
>                      fmt = ((opt->flag & VSH_OFLAG_REQ) ? "<%s>"
> -                           : _("[--%s <number>]"));
> +                           : _("[%s]"));
>                      break;
>                  case VSH_OT_STRING:
>                      /* xgettext:c-format */
> -                    fmt = _("[--%s <string>]");
> +                    fmt = _("[%s]");
>                      break;

Actually, I think these would be _("[<%s>]") (note the <> to signify the
meta-syntactic term).  But I like the idea; this shortens the synopsis
form, while still leaving the usage form to explain the meta-syntactic term.

-- 
Eric Blake   eblake at redhat.com    +1-801-349-2682
Libvirt virtualization library http://libvirt.org

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 619 bytes
Desc: OpenPGP digital signature
URL: <http://listman.redhat.com/archives/libvir-list/attachments/20110209/fe4a343d/attachment-0001.sig>