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

Re: [libvirt] [PATCH 08/13] Replace use of virStorageReportError with virReportError



On 07/18/2012 05:52 AM, Daniel P. Berrange wrote:
> From: "Daniel P. Berrange" <berrange redhat com>
> 
> Update the storage driver to use virReportError instead of
> the virStorageReportError custom macro
> 
> Signed-off-by: Daniel P. Berrange <berrange redhat com>
> ---

Still big, but not as big as 7/13.  Again, mostly mechanical, and I
didn't spot any blatant problems.

ACK.

> +++ b/src/storage/storage_backend_rbd.c
> @@ -145,8 +145,8 @@ static int virStorageBackendRBDOpenRADOSConn(virStorageBackendRBDStatePtr *ptr,
>                                pool->def->source.hosts[i].name,
>                                pool->def->source.hosts[i].port);
>          } else {
> -            virStorageReportError(VIR_ERR_INTERNAL_ERROR,
> -                                  _("received malformed monitor, check the XML definition"));
> +            virReportError(VIR_ERR_INTERNAL_ERROR,
> +                           _("received malformed monitor, check the XML definition"));

Another case of missing "%s", worth a separate cleanup patch.  [hmm,
wonder how hard it would be to make cfg.mk detect a message without a
'%' as part of syntax-check, rather than the current approach of
building with NLS disabled to trigger a gcc warning]

> +++ b/src/storage/storage_driver.c
> @@ -475,7 +475,7 @@ static int storagePoolIsActive(virStoragePoolPtr pool)
>      obj = virStoragePoolObjFindByUUID(&driver->pools, pool->uuid);
>      storageDriverUnlock(driver);
>      if (!obj) {
> -        virStorageReportError(VIR_ERR_NO_STORAGE_POOL, NULL);
> +        virReportError(VIR_ERR_NO_STORAGE_POOL, NULL);

While this still gives a working error message, it's pretty weak, and
I've noticed that at least on FreeBSD, the gcc version tends to issue
warnings about a NULL format string.

Maybe someday we should require all errors to have a message, rather
than our current practice of permitting NULL for a default message.

Or perhaps we could trivially silence those warnings if we make the
virReportError macro expand to 'format ? format : ""' instead of
'format', but then you get into issues that gcc also warns about passing
"" as a format string.  But that gets hairy fast to support a macro with
no format string by default (witness VIR_HAS_COMMA and other goop in
libvirt.c to allow VIR_DOMAIN_DEBUG() to be smart).

So for now, nothing to change here, at least without a real report of a
build failure somewhere.

-- 
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org

Attachment: signature.asc
Description: OpenPGP digital signature


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