[modularization] Removing iterator->conversion dependency
Hi, in order to remove the dependency from iterator to conversion I suggest to move |boost/implicit_cast.hpp to MPL. | |Vicente | /conversion/ <http://www.pdimov.com/tmp/report-6d1f271/conversion.html> |<boost/implicit_cast.hpp>| * from |<boost/pending/iterator_tests.hpp>| |//boost/implicit_cast.hpp| |template <typename T> inline T implicit_cast (typename mpl::identity<T>::type x) { return x; } |
On 06 Jun 2014, at 1:39, Vicente J. Botet Escriba <vicente.botet@wanadoo.fr> wrote:
Hi,
in order to remove the dependency from iterator to conversion I suggest to move |boost/implicit_cast.hpp to MPL.
That makes no semantic sense at all. Breaking dependencies is nice, but not at the cost of moving something to a place it doesn’t belong at all. Sebastian
On Friday 06 June 2014 10:08:59 Sebastian Redl wrote:
On 06 Jun 2014, at 1:39, Vicente J. Botet Escriba <vicente.botet@wanadoo.fr> wrote:
Hi,
in order to remove the dependency from iterator to conversion I suggest to move |boost/implicit_cast.hpp to MPL. That makes no semantic sense at all. Breaking dependencies is nice, but not at the cost of moving something to a place it doesn’t belong at all.
+1
On 6 June 2014 09:08, Sebastian Redl <sebastian.redl@getdesigned.at> wrote:
On 06 Jun 2014, at 1:39, Vicente J. Botet Escriba <vicente.botet@wanadoo.fr> wrote:
in order to remove the dependency from iterator to conversion I suggest to move |boost/implicit_cast.hpp to MPL.
That makes no semantic sense at all. Breaking dependencies is nice, but not at the cost of moving something to a place it doesn’t belong at all.
I think there's a better solution. Splitting lexical cast into its own module might make conversion a much lighter dependency. I'm wary of creating too many submodules, but I think that one might be justified.
2014-06-06 12:15 GMT+04:00 Daniel James <dnljms@gmail.com>:
On 6 June 2014 09:08, Sebastian Redl <sebastian.redl@getdesigned.at> wrote:
On 06 Jun 2014, at 1:39, Vicente J. Botet Escriba <
vicente.botet@wanadoo.fr> wrote:
in order to remove the dependency from iterator to conversion I suggest
to move |boost/implicit_cast.hpp to MPL.
That makes no semantic sense at all. Breaking dependencies is nice, but not at the cost of moving something to a place it doesn’t belong at all.
I think there's a better solution. Splitting lexical cast into its own module might make conversion a much lighter dependency. I'm wary of creating too many submodules, but I think that one might be justified.
+1 See my previous post about moving part of MPL into the Boost.Core (identity, bool_, integral_constant_, or_, and_). This will remove the dependency of implicit_cast from MPL -- Best regards, Antony Polukhin
Daniel James wrote:
I think there's a better solution. Splitting lexical cast into its own module might make conversion a much lighter dependency. I'm wary of creating too many submodules, but I think that one might be justified.
lexical_cast already has its own entries in libraries.htm and maintainers.txt, so I think that it makes perfect sense for it to have its own module.
Le 07/06/14 16:04, Peter Dimov a écrit :
Daniel James wrote:
I think there's a better solution. Splitting lexical cast into its own module might make conversion a much lighter dependency. I'm wary of creating too many submodules, but I think that one might be justified.
lexical_cast already has its own entries in libraries.htm and maintainers.txt, so I think that it makes perfect sense for it to have its own module.
Please could we decide on this point? Best, Vicente
On 9 June 2014 09:00, Vicente J. Botet Escriba <vicente.botet@wanadoo.fr> wrote:
Le 07/06/14 16:04, Peter Dimov a écrit :
Daniel James wrote:
I think there's a better solution. Splitting lexical cast into its own module might make conversion a much lighter dependency. I'm wary of creating too many submodules, but I think that one might be justified.
lexical_cast already has its own entries in libraries.htm and maintainers.txt, so I think that it makes perfect sense for it to have its own module.
Please could we decide on this point?
Someone has already created a lexical_cast module in develop, it's not in master yet though.
Le 06/06/14 10:08, Sebastian Redl a écrit :
On 06 Jun 2014, at 1:39, Vicente J. Botet Escriba <vicente.botet@wanadoo.fr> wrote:
Hi,
in order to remove the dependency from iterator to conversion I suggest to move |boost/implicit_cast.hpp to MPL. That makes no semantic sense at all. Breaking dependencies is nice, but not at the cost of moving something to a place it doesn’t belong at all.
Why this function could not be in MPL? Vicente
On Friday 06 June 2014 23:40:27 Vicente J. Botet Escriba wrote:
Le 06/06/14 10:08, Sebastian Redl a écrit :
On 06 Jun 2014, at 1:39, Vicente J. Botet Escriba <vicente.botet@wanadoo.fr> wrote:
Hi,
in order to remove the dependency from iterator to conversion I suggest to move |boost/implicit_cast.hpp to MPL.> That makes no semantic sense at all. Breaking dependencies is nice, but not at the cost of moving something to a place it doesn’t belong at all. Why this function could not be in MPL?
Because it has no relation to metaprogramming?
Le 06/06/14 23:47, Andrey Semashev a écrit :
On Friday 06 June 2014 23:40:27 Vicente J. Botet Escriba wrote:
Le 06/06/14 10:08, Sebastian Redl a écrit :
On 06 Jun 2014, at 1:39, Vicente J. Botet Escriba <vicente.botet@wanadoo.fr> wrote:
Hi,
in order to remove the dependency from iterator to conversion I suggest to move |boost/implicit_cast.hpp to MPL.>
That makes no semantic sense at all. Breaking dependencies is nice, but not at the cost of moving something to a place it doesn’t belong at all. Why this function could not be in MPL? Because it has no relation to metaprogramming?
Ok I see :( Vicente
On Friday 06 June 2014 01:39:24 Vicente J. Botet Escriba wrote:
Hi,
in order to remove the dependency from iterator to conversion I suggest to move |boost/implicit_cast.hpp to MPL.
|Vicente
/conversion/ <http://www.pdimov.com/tmp/report-6d1f271/conversion.html>
|<boost/implicit_cast.hpp>|
* from |<boost/pending/iterator_tests.hpp>|
|//boost/implicit_cast.hpp| | |template <typename T>
inline T implicit_cast (typename mpl::identity<T>::type x) { return x; }
I would suggest moving it to Core and modifying to not depend on MPL (i.e. reimplementing mpl::identity).
participants (6)
-
Andrey Semashev -
Antony Polukhin -
Daniel James -
Peter Dimov -
Sebastian Redl -
Vicente J. Botet Escriba