On 07/17/10 - 06:07:30PM, Matthias Bolte wrote:
2010/7/8 Chris Lalancette <clalance(a)redhat.com>:
> With the change to make vshError() responsible for printing
> all errors, there were some places in the code that would no
> longer properly print errors. The good news is that the vast
> majority of virsh was already printing errors, so this patch
> just cleans up the rest of the users to make them consistent.
>
> Signed-off-by: Chris Lalancette <clalance(a)redhat.com>
> ---
> tools/virsh.c | 228 +++++++++++++++++++++++++++++++++++++++++++--------------
> 1 files changed, 174 insertions(+), 54 deletions(-)
>
The patch as is look good, but I wonder how you spot all those places
that missed a vshError call. Did you manually go over the whole
virsh.c file and checked all libvirt API calls?
I just spot calls to virDomainMigrate and virDomainMigrateToURI
(around line 3040) that are not matched with a call to vshError, you
might want to fold that into your patch.
Yeah, I did just go over the file visually, which is why I probably missed a
couple :). Dan's suggestion for a catch-all case is not bad; it's a
programming error when it is hit, but at least we will get *something* out on
an error. I'll look at doing that.
--
Chris Lalancette