boost::sandbox::odeint 安装
我想使用 boost::sandbox 中的 odeint 模块,但我不确定如何将从 svn 检索到的沙箱代码集成到我的 boost 安装中。我已将 boost 安装到 $BOOST_ROOT=/usr/local/boost_1_45_0。将 odeint 的内容合并到我的 boost 路径中是否正确,这会将标头放置在 $BOOST_ROOT/boost/numeric/odeint.hpp 处。
这是正确的方法还是有获取和安装 boost::sandbox 代码的标准方法?
I would like to use the odeint module from the boost::sandbox but I am not sure how to integrate the sandbox code which I retrieved from svn into my boost install. I have boost installed to $BOOST_ROOT=/usr/local/boost_1_45_0. Is it proper to just merge contents of odeint into my boost path which would place the header at $BOOST_ROOT/boost/numeric/odeint.hpp.
Is this the right approach or is there a standard way of getting and installing boost::sandbox code?
如果你对这篇内容有疑问,欢迎到本站社区发帖提问 参与讨论,获取更多帮助,或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱,如果其他用户或者作者回复了您的评论,将不能在第一时间通知您!
发布评论
评论(3)
boost.odeint 完全是头文件。因此,您可以将 odeint/boost 目录的内容复制到 $BOOST_ROOT 目录中,一切都应该可以正常工作。或者,您可以将 odient 目录的位置添加到编译器的包含标志中(例如,对于 gcc -Ipath_to_odeint)。
boost.odeint is completely header-only. So, you can copy the content of the odeint/boost directory into your $BOOST_ROOT dir, and everything should work. Alternatively, you could add the location of the odient directory to the include flags of your compiler (for example for gcc -Ipath_to_odeint).
我不知道任何“正确”的解决方案,但我可以告诉你,如果你只是复制
boost/numeric/odeint.hpp
和boost/numeric/odeint,你不应该遇到任何问题/*.hpp
到该目录,因为不会覆盖任何文件(如果出现问题,只需删除它们即可完成)。为了它的价值,我为 boost::process 这样做了,它工作得没有任何问题。
I am not aware of any "proper" solution, but I can tell you that you should not meet any problems if you just copy
boost/numeric/odeint.hpp
andboost/numeric/odeint/*.hpp
to that directory since no file will be overwritten (if something fails just remove them and you're done).For what its worth I did that for boost::process and it worked without any problems.
目前我们完全重新设计了图书馆。界面将是相同的,但它将具有一些附加功能(cuda 支持、对结构和元组作为状态类型的支持)和更多步进器(刚性求解器、隐式求解器等)。错误跟踪器和单独的列表不存在,但您可以使用 [电子邮件受保护]< /a> 进行讨论或建议。
At the moment we completely redesign the library. The interface will be the same but it will have some additional features (cuda support, support for structs and tuples as state types) and some more steppers (stiff solvers, implicit solvers, etc.). A bug tracker and a separate list does not exist, but you could use [email protected] for discussion or suggestions.