Re: [Boost-users] Boost::Process doesn't work with #define new (boost 1.69)

Subject: Re: [Boost-users] Boost::Process doesn't work with #define new (boost 1.69)
From: Gavin Lambert (boost_at_[hidden])
Date: 2019-02-12 05:16:56

On 12/02/2019 17:52, Vivek Subramanian wrote:
> Well, okay, but Win32 code doesn #undef new at the start, and Microsoft 
> themselves say to #define new for using CRT!
> 
> https://docs.microsoft.com/en-us/visualstudio/debugger/finding-memory-leaks-using-the-crt-library?view=vs-2017

Actually, no, they don't -- at least not on that page.  (In fact it 
specifically un-recommends that.)

The technique shown there is to define a DBG_NEW macro.  You then 
replace "new" with "DBG_NEW" whenever it appears in your application 
code.  Hopefully this included the source of the leak.

(You're doing this manually or via find/replace, not by #define new 
DBG_NEW.)

It is true that in the past (and with MFC) they did recommend redefining 
new, but that has caveats.  To minimise problems you were always 
supposed to do it only after all #includes, as I said in my other reply.