On Mon, 09/18 10:17, Eric Blake wrote: > On 09/17/2017 09:43 PM, Fam Zheng wrote: > > This will be used by setup test user ssh. > > Shouldn't the commit message have more details about WHY we are checking > in a private key? The conversation last time included useful gems such as: > > Some guests require the key to exist prior to provisioning the guest, > and reusing a pre-built key avoids consuming entropy every time the > testsuite is run. Because the private key is well-known, care must be > taken to use the key ONLY in situations that cannot be compromised by > external network clients.
Yeah. Not everyone looks at the commit message, I'll add a tests/keys/README to document that. Fam
