[boost] [mpl] Hard upper limit for range_c
I was testing for_each and range_c and I've discovered that there's a hard limit in the range_c template. I'm attaching an example (copied from the for_each docs http://www.boost.org/doc/libs/1_55_0/libs/mpl/doc/refmanual/for-each.html) to illustrate it. This program compiles fine for me (g++ 4.9.0), but if I set NUM to a number over 890, the compilation fails. I receive an error about template instantiation depth, so I guess I could tune it in the compiler to make it work. But, anyway, why is this template recursively defined? Is it on purpose? I get that maybe range_c was intended just for small ranges, but, is it the only way it can work? Also, it only allows for statically defined range limits. Also, whenever it compiles, the bigger the number you put, the longer it takes to compile. Thanks. -- Javier Dehesa
On 05/13/14 11:23, Javier Dehesa wrote:> I was testing for_each and range_c and I've discovered that there's a
hard limit in the range_c template.
I'm attaching an example (copied from the for_each docs
http://www.boost.org/doc/libs/1_55_0/libs/mpl/doc/refmanual/for-each.html)
to illustrate it. This program compiles fine for me (g++ 4.9.0), but if I set NUM to a number over 890, the compilation fails. I receive an error about template instantiation depth, so I guess I could tune it in the compiler to make it work. But, anyway, why is this template recursively defined? Is it on purpose?
On purpose: http://www.boost.org/doc/libs/1_31_0/libs/mpl/doc/paper/mpl_paper.html The reason for only using recursion is that an imperative like loop: std::vector<int> int_list; for(unsigned i=0; i<890; ++i) int_list.push_back(i); the value of i and int_list are modified during each iteration of the loop. TMP is a pure *functional* language, meaning there is no changing of the value of a variable. Once it's calculated, the value cannot be changed. HTH. -regards, Larry
participants (2)
-
Javier Dehesa -
Larry Evans