In bug 343682 comment 8 Wan-Teh said: The SECItem derTemp in crmf_encode_popoprivkey may suffer from the same problem and should also be initialized.
Created attachment 232526 [details] [diff] [review] Patch v1 Bob, can you please review?
Comment on attachment 232526 [details] [diff] [review] Patch v1 r=wtc. When I looked at the code closely, I found that this patch isn't necessary because when the crmf_init_encoder_callback_arg call in crmf_encode_popoprivkey fails, derTemp.data has been set to NULL. This patch is a good idea because we should avoid depending on the side effect of a function on failure. But this means we don't need to check this patch in on the NSS_3_11_BRANCH.
Since the severity of this bug has been lowered to "code cleanup to improve maintainability", it should not block any bug. I also suggest that we just check it in on the NSS trunk.
fixed on trunk Checking in crmfpop.c; /cvsroot/mozilla/security/nss/lib/crmf/crmfpop.c,v <-- crmfpop.c new revision: 1.8; previous revision: 1.7 done