From: Venu G. <gop...@ya...> - 2011-09-20 07:03:49
|
>> From: Charles Wilson <cwi...@us...> >>>> This is possible, but of course this is up to mingw.org's maintainer, >>>> if they want to add it. >> From: Charles Wilson <cwi...@us...> >> I think it is a good idea; the BSD-vs-LGPL license is a good argument in >> winpthread's favor (also, pthreads-w32 has seen relatively few updates >> in recent years). On the down side, winpthreads uses TLS callbacks so >> it doesn't support Win9x whereas pthreads-w32 does -- but that really >> isn't a drawback, because our (mingw.org) native thread cleanup code now >> employs TLS too, so we no longer support win9x EITHER. >> However, I'm not the mingw gcc maintainer, so it's really up to Cesar. >> (And, as we just this week had our first 4.6.x release with ONE >> pthreads-related change, we might want to let that settle for a bit >> before we contemplate introducing ANOTHER major change to our threading >> infrastructure!) Chuck, JonY, Ruben, K. Frank et al: Thank you all for your replies. Chuck, Ruben: Would appreciate if y'all could keep it on you higher priority list for the future (don't know how to contact Cesar). It is about time the world had a standardized threading API across multiple platforms in C/C++. As a user of the development environment, I don't really like patching it (esp. if I have to get and compile the sources) due to configuration management concerns. So I will wait till C++0X support shows up fully in mingw in an offical release in binary form. In the meantime, I think I am just going to more forward with my own bare-minimum (i.e. containing only what I need immediately) wrapper class that wraps Win32 APIs on Windows and pthread APIs on Linux. (Don't have a Linux or Mac machine, so I'll have to leave the other OS's implementation untested for now). I'll if I can make it look like std::thread as much as possible. Cheers, Venu Gopal |