Re: [eigen] Eigen 3.1 release plans |
[ Thread Index |
Date Index
| More lists.tuxfamily.org/eigen Archives
]
- To: eigen@xxxxxxxxxxxxxxxxxxx
- Subject: Re: [eigen] Eigen 3.1 release plans
- From: Gael Guennebaud <gael.guennebaud@xxxxxxxxx>
- Date: Wed, 2 May 2012 11:24:11 +0200
- Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=mime-version:in-reply-to:references:from:date:message-id:subject:to :content-type:content-transfer-encoding; bh=er7OVmv6rT4z5lHmSIudndR6Vei/VrhYoQiwhS2deuM=; b=Cs3EKqegSHSz5mZdW7T+YK7kCHY2vnLM/TN3j3bz+nG3HoaEgLOl1WQ3L7QlPSX/bO dv6GK38mFGduwQ26MF1+RJaRS1kiHLBuibH5C+VnCl2DfUMuJSMtW++rMKREbP146AYH InvEdd2so4+OAXfgDDanBOkDOxxNGrVT1SLP9mSppFQyOPPR58KNxL8tFcHZMx2vnElP kCBRaJ1Ay0rB+Zivb61NuU2LSwgqrmSX0a2NdPgtMUHg4RS7VRb1nX9ng/tFlv8Im0KE ZVKI59ivKOvt6eIf5A9f2lLfN24d7ldgHiVoPjWZ5pyiJP+5rc9tyCzFMAcOLOuma70l mbbg==
The problem is that I'm pretty sure some users are already using it,
so we cannot simply remove or rename it. Updating its documentation
should at least discourage new users to rely on it. For old users I
see you proposed to raise a warning if EDTRM is defined and add a new
more explicit token. Perhaps it would be enough to add a warning in
the release note, which is, I think, what is the most likely to be
read by developers.
gael
On Wed, May 2, 2012 at 10:23 AM, Christoph Hertzberg
<chtz@xxxxxxxxxxxxxxxxxxxxxxxx> wrote:
> On 02.05.2012 10:03, Gael Guennebaud wrote:
>>
>> Regarding bug 422, I would simply update the documentation to state
>> it's mostly for internal use and it is not recommended to use it.
>
>
> Internal use such as what?
> And more important, for people who specifically require column major
> matrices, is it save enough to assume EDTRM is not defined, should they
> check whether it really is not defined, or are they required to always
> specify ColMajor-ness?
>
> I find it quite annoying if I'm required to write things such as:
>
> template<int n, int m>
> class MyClass {
> typedef Eigen::Matrix<double, n, m, n!=1 ? Eigen::ColMajor :
> Eigen::RowMajor> MyMatrix;
> // ...
> };
> Additionally, bug 416 makes this slightly worse here ...
>
> So unless someone comes up with a good reason for keeping EDTRM, I would
> vote for deprecating it, saving us a lot of trouble in the future.
>
>
> Christoph
>
>
>
> --
> ----------------------------------------------
> Dipl.-Inf. Christoph Hertzberg
> Cartesium 0.049
> Universität Bremen
> Enrique-Schmidt-Straße 5
> 28359 Bremen
>
> Tel: +49 (421) 218-64252
> ----------------------------------------------
>
>