| Re: [eigen] Dynamic exception specifications are deprecated in C++11 mode and other warnings |
[ Thread Index |
Date Index
| More lists.tuxfamily.org/eigen Archives
]
- To: eigen@xxxxxxxxxxxxxxxxxxx
- Subject: Re: [eigen] Dynamic exception specifications are deprecated in C++11 mode and other warnings
- From: Nicola Gigante <nicola.gigante@xxxxxxxxx>
- Date: Sat, 23 Aug 2014 16:26:05 +0200
- Cc: Nicola Gigante <nicola.gigante@xxxxxxxxx>
- Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=content-type:mime-version:subject:from:in-reply-to:date:cc :content-transfer-encoding:message-id:references:to; bh=q7xgjoRbU+genj5DaHCRy0gA1zEEM5jOFjmHIAlJnGE=; b=szGX4t0xK2W97QDAIWnyunvL6a747aUpeg8I23CAXlldY5+2uPal2mQ6WxizrLhJDG mqlBZjA9oVpFrfmi7LgdmyzlrT0n/bDRQlz9QuyUHkt+cPh6zhaQKCPDWpoxH855kG3t SwIXo5Z+DnaZ6fAsSrzDHZA5ZEWQGQIIHL17filoLJClU+N/Qh6aBnnL+qcorM5OReFO YqrSBwdRuJ4fBy4DNv1fVc/VkSVTnVJUI+5zbFBvq1RVOVcZA6S8Dxf9G0X0ICfyroD3 8TMtdO12Fvsjd9M+wKiCuYulU55BNB0s92muJhtzxo9LSo2J9oDDL3OY6SV9SubNP8MY kpIg==
Il giorno 23/ago/2014, alle ore 16:11, Christoph Hertzberg <chtz@xxxxxxxxxxxxxxxxxxxxxxxx> ha scritto:
> On 23.08.2014 15:28, Nicola Gigante wrote:
>> The last time I've looked, I've seen some move constructor conditionally
>> declared in C++11 mode, but now I can't find them. How do you support
>> move semantics in C++11 mode without testing the language version?
>
> For GCC we do check the value of __cplusplus, but we do not have any C++11 support in the 3.2 branch (that might be the reason, you couldn't find any ...)
> Here is were we determine if we have RValue (i.e. 'move') support:
> https://bitbucket.org/eigen/eigen/src/51c2effe19bd54186b8b39c34004ad3f28d7dbdd/Eigen/src/Core/util/Macros.h?at=default#cl-121
>
>
> Regarding your original question, we already define some compiler-dependent attributes/declspecs in src/Core/util/Macros.h and I would not object introducing EIGEN_NORETURN and EIGEN_NOTHROW macros (at least in the dev-branch).
> The 3.2 branch should generally only receive bug-fixes. We could still introduce these macros, but there are generally no plans to make it support C++1x features.
>
Ok.
I'm not tied to the 3.2 branch for any particular reason, I've just downloaded
the latest version. Do you suggest to switch to follow the development branch?
>
> Christoph
Bye,
Nicola