[modularization] Moving <boost/math/common_factor_ct.hpp> to TypeTraits
Hi,
|moving boost/math/common_factor_ct.hpp to TypeTraits would remove
the dependency from Interprocess, Pool and Variant to Math.
|
/interprocess/
http://www.pdimov.com/tmp/report-6d1f271/interprocess.html
|
Vicente J. Botet Escriba wrote:
Hi,
moving boost/math/common_factor_ct.hpp to TypeTraits would remove the dependency from Interprocess, Pool and Variant to Math.
common_factor_ct.hpp defines static_gcd and static_lcm, and these seem candidates for 'integer' rather than type traits. I wondered what Ratio uses, and you have your own implementations there, with mpl lambda support and everything.
Le 06/06/14 03:43, Peter Dimov a écrit :
Vicente J. Botet Escriba wrote:
Hi,
moving boost/math/common_factor_ct.hpp to TypeTraits would remove the dependency from Interprocess, Pool and Variant to Math.
common_factor_ct.hpp defines static_gcd and static_lcm, and these seem candidates for 'integer' rather than type traits. Yes this will be better as Integer is at a lower level than Math.
I wondered what Ratio uses, and you have your own implementations there, with mpl lambda support and everything. Yes. ratio manage in a specific way the overflow detection.
Vicente
Vicente J. Botet Escriba wrote: Le 06/06/14 03:43, Peter Dimov a écrit :
common_factor_ct.hpp defines static_gcd and static_lcm, and these seem candidates for 'integer' rather than type traits.
Yes this will be better as Integer is at a lower level than Math.
No, it will be better because 'integer' already contains integer operations such as static_log2 and static_min_max. :-)
On 06/06/2014 02:43, Peter Dimov wrote:
Vicente J. Botet Escriba wrote:
Hi,
moving boost/math/common_factor_ct.hpp to TypeTraits would remove the dependency from Interprocess, Pool and Variant to Math.
common_factor_ct.hpp defines static_gcd and static_lcm, and these seem candidates for 'integer' rather than type traits.
I'm not saying you're wrong to move these - they are compile time functions rather than the usual Math lib stuff, but we shouldn't get too carried away moving everything "just because". Users still have to be able to find their stuff... John.
John Maddock wrote:
I'm not saying you're wrong to move these - they are compile time functions rather than the usual Math lib stuff, but we shouldn't get too carried away moving everything "just because".
+1 on not getting carried away. We're almost done with the initial flurry of moving things around; this was temporary and should not be our usual procedure. Ideally, breaking a dependency from X to Y should be done or initiated by the maintainer of X. Or, failing that, by the maintainer of Y. :-) Still, ...
Users still have to be able to find their stuff...
... 'integer' would be the first place I look for static_gcd due to it already having static_this and _that.
Le 06/06/14 13:04, Peter Dimov a écrit :
John Maddock wrote:
I'm not saying you're wrong to move these - they are compile time functions rather than the usual Math lib stuff, but we shouldn't get too carried away moving everything "just because".
+1 on not getting carried away. We're almost done with the initial flurry of moving things around; this was temporary and should not be our usual procedure. Ideally, breaking a dependency from X to Y should be done or initiated by the maintainer of X. Or, failing that, by the maintainer of Y. :-)
Still, ...
Users still have to be able to find their stuff...
... 'integer' would be the first place I look for static_gcd due to it already having static_this and _that.
Integer is ok for me. Could we decide this point? Best, Vicente
El 06/06/2014 3:17, Vicente J. Botet Escriba escribió:
Hi,
|moving boost/math/common_factor_ct.hpp to TypeTraits would remove the dependency from Interprocess, Pool and Variant to Math. |
/interprocess/ http://www.pdimov.com/tmp/report-6d1f271/interprocess.html
|
| * from |
|
This include is not needed. Seems like I forgot to drop the dependency. I'll remove that. Thanks for the report, Best, Ion
Le 06/06/14 14:08, Ion Gaztañaga a écrit :
El 06/06/2014 3:17, Vicente J. Botet Escriba escribió:
Hi,
|moving boost/math/common_factor_ct.hpp to TypeTraits would remove the dependency from Interprocess, Pool and Variant to Math. |
/interprocess/ http://www.pdimov.com/tmp/report-6d1f271/interprocess.html
|
| * from |
| This include is not needed. Seems like I forgot to drop the dependency. I'll remove that.
Thanks, Vicente
participants (4)
-
Ion Gaztañaga -
John Maddock -
Peter Dimov -
Vicente J. Botet Escriba