Re: [boost] Re: Re: Moving Boost.Thread forward

From: Jeremy Maitin-Shepard (jbms_at_[hidden])
Date: 2003-12-17 11:50:55

"Michael Glassford" <glassfordm_at_[hidden]> writes:

> The main technical reason in addition to the ones you mention is that, with
> security settings continually tightening, I can imagine problems in the
> future with writing code into a file and executing it (which might be
> interpreted as virus-like behavior by virus protection programs, for
> instance). I'd have to look at the code in more detail than I have so far to
> see if this is a real problem.

> The main non-technical reason is that I can imagine people might be leery of
> accepting such a solution into Boost. Until they do (or don't), there's no
> way of knowing if this is a real problem.

I don't think it is reasonable to base decisions on possible incorrect
behavior from ``virus'' scanning software that might impede a
particular approach.

Instead of a writing to a file, another (less portable) method is to
modify the process's memory (in place), such that it appears to Windows
that some DLL has been loaded, and then this DLLs cleanup code is run
when the process exits.  I believe a technique similar to this is used
by various ``cheat''-prevention systems added to certain multiplayer
games.  This would have the advantage of not using a temporary file.

-- 
Jeremy Maitin-Shepard