On 06/21/2016 10:05 AM, Peter Krempa wrote:
On Mon, Jun 13, 2016 at 20:27:45 -0400, John Ferlan wrote:
Move to virsecret.c and rename to virSecretParseSecret. Also convert to
NACK to the naming. It implies that it parses <secret> xmls.
The "Secret" just replaced the "StorageAuthDef" since no longer was a <secret> going to be only an element of <auth>.
usage xmlNodePtr and virXMLPropString rather than virXPathString.
Signed-off-by: John Ferlan <jferlan@redhat.com> --- po/POTFILES.in | 1 + src/libvirt_private.syms | 1 + src/util/virsecret.c | 44 +++++++++++++++++++++++++++++ src/util/virsecret.h | 5 +++- src/util/virstoragefile.c | 71 ++++++++++++----------------------------------- 5 files changed, 67 insertions(+), 55 deletions(-)
diff --git a/src/libvirt_private.syms b/src/libvirt_private.syms index 32d5179..ca65885 100644 --- a/src/libvirt_private.syms +++ b/src/libvirt_private.syms @@ -2211,6 +2211,7 @@ virSecurityLabelDefNew; # util/virsecret.h virSecretLookupDefClear; virSecretLookupDefCopy; +virSecretParseSecret;
It should be named similarly to the above or to the original name.
ACK to the code as long as you name it reasonably
OK - is virSecretLookupParseSecret more palatable? Or an even longer name virSecretLookupParseDomainSecret. Up through this point it's only a subelement of <auth>, but adding to <encryption> short term and to <serial> for TLS secrets in another series. Also, As I was working in the TLS code - I began to realize using 'secdef' for variable names became confusing, so I changed to 'seclookupdef'... That was posted as a separate series I also note that patch 5 doesn't yet have any review, but this patch relies on that. Tks - John