On 07.05.2013 00:00, Eric Blake wrote:
On 05/03/2013 08:53 AM, Michal Privoznik wrote:
--- src/cpu/cpu_generic.c | 8 ++++---- src/cpu/cpu_map.c | 3 ++- src/cpu/cpu_powerpc.c | 13 ++++++------- src/cpu/cpu_x86.c | 10 +++++----- 4 files changed, 17 insertions(+), 17 deletions(-)
+++ b/src/cpu/cpu_powerpc.c @@ -333,9 +333,8 @@ ppcDecode(virCPUDefPtr cpu, goto cleanup; }
- if (!(cpu->model = strdup(model->name)) || - (model->vendor && !(cpu->vendor = strdup(model->vendor->name)))) { - virReportOOMError(); + if (VIR_STRDUP(cpu->model, model->name) < 0 || + (model->vendor && VIR_STRDUP(cpu->vendor, model->vendor->name) < 0)) {
Another case where not requiring ATTRIBUTE_NONNNULL(2) on VIR_STRDUP would be handy. I think I've convinced myself we should go that route. Maybe it would help if I post a patch showing what I'm thinking?
Yep. you're completely right. However, the 1st patch is already pushed with ATTRIBUTE_NONNULL(2). So do you want me to post a fix dropping the useless attribute and pushing this without it, or ..,?
ACK.