Re: [PATCH 1/3] Documentation/CodingGuidelines: explain why assert is bad

From: Jonathan Nieder <jrnieder@gmail.com>
Date: Nov 22, 2017, 23:54 UTC
Message-ID: <20171122235425.GK11671@aiede.mtv.corp.google.com>
In-Reply-To: <CAGZ79kY7yrgYSM7jLQD3Q4sRpjovXY=HZwU4M-Ai_XRPseda9A@mail.gmail.com>

Stefan Beller wrote:

> On Wed, Nov 22, 2017 at 2:59 PM, Jonathan Nieder <jrnieder@gmail.com> wrote:

Show 8 quoted lines

>> In a certain ideal world, the preference would be reversed: you'd want
>> to use assert() wherever you can and require the compiler to check
>> that all assert()s are verifiable at compile time.  A check that a
>> static analyzer can verify is more valuable than a run-time check.
>> When a compile-time check is not possible, you'd have to fall back to
>> BUG_ON().
>
> Linux has BUILT_BUG_ON as well, which we may desire?

I thought so until I tried to use it: https://public-inbox.org/git/20170830065631.GH153983@aiede.mtv.corp.google.com/

So I think we are stuck with run-time checking for now.

Thanks, Jonathan