Re: [PATCH] remote-curl: add testing for intelligent retry for HTTP

Johannes Schindelin [off-list ref] writes:

We should also consider using `test-tool genrandom <seed>` instead (where
`<seed>` would have to be predictable, but probably would have to change
between `gen_nonce()` calls).

Yup, that is exactly why I asked Sean about randomness requirement.

It turns out that they care only about uniqueness, so the comparison
is between keeping an ever-incrementing counter and (1) echoing its
current contents and/or (2) feeding it to "test-tool genrandom" as
the seed.  The complexity of the code _we_ need to write anew is the
same, but echo would probably be a win in both the number of forks
and cycles departments.

Thanks.