Re: [boost] A summary of testing with sanitizers

Subject: Re: [boost] A summary of testing with sanitizers
From: Niall Douglas (s_sourceforge_at_[hidden])
Date: 2015-05-05 20:05:35

On 5 May 2015 at 10:58, Ben Pope wrote:

> The summary now looks like this:

I don't know about other people, but I'd prefer these marked against 
specific libraries so we can more easily see which are at fault.

We should be *absolutely* running these as part of the Boost 
regression matrix. Moreover, ubsan should ALWAYS be turned on by 
default including on release optimised builds. Individual library 
maintainers should turn it off if needed, but the default should be 
always on.

Niall

-- 
ned Productions Limited Consulting
http://www.nedproductions.biz/ 
http://ie.linkedin.com/in/nialldouglas/