[Tarantool-patches] [RFC v5 4/5] limbo: order access to the promote terms

[Cyrill Gorcunov]

On Thu, Jul 15, 2021 at 02:48:48PM +0300, Serge Petrenko wrote:
> 
> I think all panic_on(!cond) invocations may be replaced with assert(cond)
> 
> Even if we use some functions incorrectly, this may be catched by an
> assertion.

You know, initially I thought of using assert as well but then
realized that our test's coverage is far from being perfect
so that if we manage to miss locking, the results are really
awfull - data become inconsistent as far as I understand. So
I would prefer to have panic_on here. But if you still prefer
assert then sure, I'll drop panic_on.