std::vector<int[2]> intializing trouble
| Author |
Message |
|
John Siedlick
#1 / 15
|
 std::vector<int[2]> intializing trouble I am trying to create a vector containing simple fixed sized base type arrays. I want to create an initial element in the vector that will be filled later. I do not have a concrete base object to construct from (or else I would simply call vector::push_back). I thought I could use vector::resize or vector::push_back
as follows, but no luck.
Can anyone explain why the second example does not work?
This works:
std::vector<int> test;
test.resize(1);
or
test.push_back(std::vector<int>::value_type());
This does not work.
std::vector<int[2]> test;
test.resize(1);
or
test.push_back(std::vector<int[2]>::value_type());
Thanks for any input.
John
|
| Tue, 23 Aug 2005 00:33:14 GMT |
|
 |
|
Igor Tandetni
#2 / 15
|
std::vector<int[2]> intializing trouble You can't do that - C arrays are not assignable. In other words, this doesn't work: int a[2];
int b[2];
b = a; // doesn't compile
int c[2] = a; // doesn't compile
Consider using vector<pair<int, int> >
--
With best wishes,
Igor Tandetnik
"For every complex problem, there is a solution that is simple, neat,
and wrong." H.L. Mencken
Quote: > I am trying to create a vector containing simple fixed > sized base type arrays. I want to create an initial > element in the vector that will be filled later. I do > not have a concrete base object to construct from (or > else I would simply call vector::push_back). > I thought I could use vector::resize or vector::push_back
> as follows, but no luck.
> Can anyone explain why the second example does not work?
> This works:
> std::vector<int> test;
> test.resize(1);
> or
> test.push_back(std::vector<int>::value_type());
> This does not work.
> std::vector<int[2]> test;
> test.resize(1);
> or
> test.push_back(std::vector<int[2]>::value_type());
> Thanks for any input.
> John
|
| Tue, 23 Aug 2005 01:22:16 GMT |
|
|
|
John
#3 / 15
|
std::vector<int[2]> intializing trouble vector<pair<int, int> > won't work as I have a C array that is 928 elements. I only used [2] as a simple example.
Quote: > You can't do that - C arrays are not assignable. In other words, this > doesn't work: > int a[2];
> int b[2];
> b = a; // doesn't compile
> int c[2] = a; // doesn't compile
> Consider using vector<pair<int, int> >
> --
> With best wishes,
> Igor Tandetnik
> "For every complex problem, there is a solution that is simple, neat,
> and wrong." H.L. Mencken
> > I am trying to create a vector containing simple fixed
> > sized base type arrays. I want to create an initial
> > element in the vector that will be filled later. I do
> > not have a concrete base object to construct from (or
> > else I would simply call vector::push_back).
> > I thought I could use vector::resize or vector::push_back
> > as follows, but no luck.
> > Can anyone explain why the second example does not work?
> > This works:
> > std::vector<int> test;
> > test.resize(1);
> > or
> > test.push_back(std::vector<int>::value_type());
> > This does not work.
> > std::vector<int[2]> test;
> > test.resize(1);
> > or
> > test.push_back(std::vector<int[2]>::value_type());
> > Thanks for any input.
> > John
|
| Tue, 23 Aug 2005 01:33:21 GMT |
|
|
|
Ken Alverso
#4 / 15
|
std::vector<int[2]> intializing trouble
Quote: > vector<pair<int, int> > won't work as I have a C array that is 928
elements.
> I only used [2] as a simple example.
vector<vector<int> > is probably your best choice then. Ken
|
| Tue, 23 Aug 2005 01:46:42 GMT |
|
|
|
c nicholso
#5 / 15
|
std::vector<int[2]> intializing trouble what's wrong with vector< pair< pair< pair< pair<pair< pair<pair< pair<pair< pair<pair<
pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
pair<pair< pair<pair< pair<pair< pair<?
maybe you could simplify it with recursive macros!
=]
charles nicholson
Quote:
> > vector<pair<int, int> > won't work as I have a C array that is 928 > elements. > > I only used [2] as a simple example. > vector<vector<int> > is probably your best choice then.
> Ken
|
| Tue, 23 Aug 2005 03:41:45 GMT |
|
|
|
Igor Tandetni
#6 / 15
|
std::vector<int[2]> intializing trouble How would you access a k-th element of such a tuple, with k determined at run-time? -- With best wishes, Igor Tandetnik "For every complex problem, there is a solution that is simple, neat,
and wrong." H.L. Mencken
Quote: > what's wrong with > vector< pair< pair< pair< pair<pair< pair<pair< pair<pair< pair<pair<
> pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
pair<pair<
> pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
pair<pair<
> pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
pair<pair<
> pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
pair<pair<
> pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
pair<pair<
> pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
pair<pair<
> pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
pair<pair<
> pair<pair< pair<pair< pair<pair< pair<?
> maybe you could simplify it with recursive macros!
> =]
> charles nicholson
> > > vector<pair<int, int> > won't work as I have a C array that is 928
> > elements.
> > > I only used [2] as a simple example.
> > vector<vector<int> > is probably your best choice then.
> > Ken
|
| Tue, 23 Aug 2005 03:50:48 GMT |
|
|
|
c nicholso
#7 / 15
|
std::vector<int[2]> intializing trouble Why, with more macros, of course! (it was a joke, btw) charles nicholson
Quote: > How would you access a k-th element of such a tuple, with k determined > at run-time? > -- > With best wishes, > Igor Tandetnik > "For every complex problem, there is a solution that is simple, neat,
> and wrong." H.L. Mencken
> > what's wrong with
> > vector< pair< pair< pair< pair<pair< pair<pair< pair<pair< pair<pair<
> > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> pair<pair<
> > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> pair<pair<
> > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> pair<pair<
> > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> pair<pair<
> > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> pair<pair<
> > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> pair<pair<
> > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> pair<pair<
> > pair<pair< pair<pair< pair<pair< pair<?
> > maybe you could simplify it with recursive macros!
> > =]
> > charles nicholson
> > > > vector<pair<int, int> > won't work as I have a C array that is 928
> > > elements.
> > > > I only used [2] as a simple example.
> > > vector<vector<int> > is probably your best choice then.
> > > Ken
|
| Tue, 23 Aug 2005 03:57:11 GMT |
|
|
|
Ken Alverso
#8 / 15
|
std::vector<int[2]> intializing trouble Probably something along the lines of (after changing the format from pair<pair<>,int> to pair<int,pair<> >): template <typename tuple>
int& access_element(tuple& t, int n) {
if (n == 0) return t.first; else return access_element(t.second,n-1);
Quote: }
template <> int& access_element(int& t, int n) { if (n != 0) throw std::exception("accessing beyond the end of the tuple"); return t; Quote: }
pair<int,pair<int,pair<int,int> > > tuple; for (int i=0;i<4;++i) access_element(tuple,i) = 2*i; for (int i=0;i<4;++i) std::cerr << access_element(tuple,i) << std::endl; Not that I recommend such a thing :)
Ken
Quote: > How would you access a k-th element of such a tuple, with k determined > at run-time? > -- > With best wishes, > Igor Tandetnik > "For every complex problem, there is a solution that is simple, neat,
> and wrong." H.L. Mencken
> > what's wrong with
> > vector< pair< pair< pair< pair<pair< pair<pair< pair<pair<
pair<pair<
> > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> pair<pair<
> > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> pair<pair<
> > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> pair<pair<
> > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> pair<pair<
> > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> pair<pair<
> > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> pair<pair<
> > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> pair<pair<
> > pair<pair< pair<pair< pair<pair< pair<?
> > maybe you could simplify it with recursive macros!
> > =]
> > charles nicholson
> > > > vector<pair<int, int> > won't work as I have a C array that is
928
> > > elements.
> > > > I only used [2] as a simple example.
> > > vector<vector<int> > is probably your best choice then.
> > > Ken
|
| Tue, 23 Aug 2005 04:04:24 GMT |
|
|
|
Igor Tandetni
#9 / 15
|
std::vector<int[2]> intializing trouble That would work, though it makes accessing n-th element an operation of O(n) complexity instead of O(1). Also, recursive functions typically don't get inlined, so it's going to be slow and bloated. -- With best wishes, Igor Tandetnik "For every complex problem, there is a solution that is simple, neat,
and wrong." H.L. Mencken
Quote: > Probably something along the lines of (after changing the format from > pair<pair<>,int> to pair<int,pair<> >): > template <typename tuple>
> int& access_element(tuple& t, int n) {
> if (n == 0) return t.first; else return
access_element(t.second,n-1); Quote: > } > template <>
> int& access_element(int& t, int n) {
> if (n != 0) throw std::exception("accessing beyond the end of the
> tuple");
> return t;
> }
> pair<int,pair<int,pair<int,int> > > tuple;
> for (int i=0;i<4;++i) access_element(tuple,i) = 2*i;
> for (int i=0;i<4;++i) std::cerr << access_element(tuple,i) <<
std::endl;
> Not that I recommend such a thing :)
> Ken
> > How would you access a k-th element of such a tuple, with k
determined
> > at run-time?
> > --
> > With best wishes,
> > Igor Tandetnik
> > "For every complex problem, there is a solution that is simple,
neat,
> > and wrong." H.L. Mencken
> > > what's wrong with
> > > vector< pair< pair< pair< pair<pair< pair<pair< pair<pair<
> pair<pair<
> > > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> > pair<pair<
> > > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> > pair<pair<
> > > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> > pair<pair<
> > > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> > pair<pair<
> > > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> > pair<pair<
> > > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> > pair<pair<
> > > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> > pair<pair<
> > > pair<pair< pair<pair< pair<pair< pair<?
> > > maybe you could simplify it with recursive macros!
> > > =]
> > > charles nicholson
> > > > > vector<pair<int, int> > won't work as I have a C array that is
> 928
> > > > elements.
> > > > > I only used [2] as a simple example.
> > > > vector<vector<int> > is probably your best choice then.
> > > > Ken
|
| Tue, 23 Aug 2005 04:11:37 GMT |
|
|
|
Ken Alverso
#10 / 15
|
std::vector<int[2]> intializing trouble Actually, it's not technically recursive, since each call is to a different instance of the function. It will create N functions, where N is the size of your tuple, but I wouldn't be surprised if the compiler was able to completely inline the access. (I wouldn't be surprised if it didn't, either.) Ken
Quote: > That would work, though it makes accessing n-th element an operation of > O(n) complexity instead of O(1). Also, recursive functions typically > don't get inlined, so it's going to be slow and bloated. > -- > With best wishes, > Igor Tandetnik > "For every complex problem, there is a solution that is simple, neat,
> and wrong." H.L. Mencken
> > Probably something along the lines of (after changing the format
from
> > pair<pair<>,int> to pair<int,pair<> >):
> > template <typename tuple>
> > int& access_element(tuple& t, int n) {
> > if (n == 0) return t.first; else return
> access_element(t.second,n-1);
> > }
> > template <>
> > int& access_element(int& t, int n) {
> > if (n != 0) throw std::exception("accessing beyond the end of the
> > tuple");
> > return t;
> > }
> > pair<int,pair<int,pair<int,int> > > tuple;
> > for (int i=0;i<4;++i) access_element(tuple,i) = 2*i;
> > for (int i=0;i<4;++i) std::cerr << access_element(tuple,i) <<
> std::endl;
> > Not that I recommend such a thing :)
> > Ken
> > > How would you access a k-th element of such a tuple, with k
> determined
> > > at run-time?
> > > --
> > > With best wishes,
> > > Igor Tandetnik
> > > "For every complex problem, there is a solution that is simple,
> neat,
> > > and wrong." H.L. Mencken
> > > > what's wrong with
> > > > vector< pair< pair< pair< pair<pair< pair<pair< pair<pair<
> > pair<pair<
> > > > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
pair<pair<
> > > pair<pair<
> > > > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
pair<pair<
> > > pair<pair<
> > > > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
pair<pair<
> > > pair<pair<
> > > > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
pair<pair<
> > > pair<pair<
> > > > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
pair<pair<
> > > pair<pair<
> > > > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
pair<pair<
> > > pair<pair<
> > > > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
pair<pair<
> > > pair<pair<
> > > > pair<pair< pair<pair< pair<pair< pair<?
> > > > maybe you could simplify it with recursive macros!
> > > > =]
> > > > charles nicholson
> > > > > > vector<pair<int, int> > won't work as I have a C array that
is
> > 928
> > > > > elements.
> > > > > > I only used [2] as a simple example.
> > > > > vector<vector<int> > is probably your best choice then.
> > > > > Ken
|
| Tue, 23 Aug 2005 04:16:51 GMT |
|
|
|
Ken Alverso
#11 / 15
|
std::vector<int[2]> intializing trouble Err...that can't work with a runtime value though. Right. Empirical testing shows that the element is accessed directly when the
index is a constant integer and the function is inlined.
Taking a dynamic index does take O(N) time, but the entire process is
inlined (at least at a size 6 tuple).
Ken
Quote: > Actually, it's not technically recursive, since each call is to a > different instance of the function. It will create N functions, where N > is the size of your tuple, but I wouldn't be surprised if the compiler > was able to completely inline the access. (I wouldn't be surprised if > it didn't, either.) > Ken
> > That would work, though it makes accessing n-th element an operation
> of
> > O(n) complexity instead of O(1). Also, recursive functions typically
> > don't get inlined, so it's going to be slow and bloated.
> > --
> > With best wishes,
> > Igor Tandetnik
> > "For every complex problem, there is a solution that is simple,
neat,
> > and wrong." H.L. Mencken
> > > Probably something along the lines of (after changing the format
> from
> > > pair<pair<>,int> to pair<int,pair<> >):
> > > template <typename tuple>
> > > int& access_element(tuple& t, int n) {
> > > if (n == 0) return t.first; else return
> > access_element(t.second,n-1);
> > > }
> > > template <>
> > > int& access_element(int& t, int n) {
> > > if (n != 0) throw std::exception("accessing beyond the end of
the
> > > tuple");
> > > return t;
> > > }
> > > pair<int,pair<int,pair<int,int> > > tuple;
> > > for (int i=0;i<4;++i) access_element(tuple,i) = 2*i;
> > > for (int i=0;i<4;++i) std::cerr << access_element(tuple,i) <<
> > std::endl;
> > > Not that I recommend such a thing :)
> > > Ken
> > > > How would you access a k-th element of such a tuple, with k
> > determined
> > > > at run-time?
> > > > --
> > > > With best wishes,
> > > > Igor Tandetnik
> > > > "For every complex problem, there is a solution that is simple,
> > neat,
> > > > and wrong." H.L. Mencken
> > > > > what's wrong with
> > > > > vector< pair< pair< pair< pair<pair< pair<pair< pair<pair<
> > > pair<pair<
> > > > > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> pair<pair<
> > > > pair<pair<
> > > > > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> pair<pair<
> > > > pair<pair<
> > > > > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> pair<pair<
> > > > pair<pair<
> > > > > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> pair<pair<
> > > > pair<pair<
> > > > > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> pair<pair<
> > > > pair<pair<
> > > > > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> pair<pair<
> > > > pair<pair<
> > > > > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> pair<pair<
> > > > pair<pair<
> > > > > pair<pair< pair<pair< pair<pair< pair<?
> > > > > maybe you could simplify it with recursive macros!
> > > > > =]
> > > > > charles nicholson
> > > > > > > vector<pair<int, int> > won't work as I have a C array
that
> is
> > > 928
> > > > > > elements.
> > > > > > > I only used [2] as a simple example.
> > > > > > vector<vector<int> > is probably your best choice then.
> > > > > > Ken
|
| Tue, 23 Aug 2005 04:39:14 GMT |
|
|
|
Ken Alverso
#12 / 15
|
std::vector<int[2]> intializing trouble
Quote: > That would work, though it makes accessing n-th element an operation
of
> O(n) complexity instead of O(1). Also, recursive functions typically
> don't get inlined, so it's going to be slow and bloated.
Of course, you can get O(lg(n)) performance out of a pair<pair<>,pair<> Quote: > structure.
template <typename tuple> inline int size(const tuple& t) { return size(t.first)+size(t.second); Quote: }
template <> inline int size(const int& t) { return 1; Quote: }
template <typename tuple> inline int& get(tuple& t, int i) { if (i < size(t.first)) return get(t.first,i); else return get(t.second,i-size(t.first)); Quote: }
template <> inline int& get(int& t, int i) { return t; Quote: }
pair<pair<pair<int,int>,pair<int,int> >,pair<pair<int,int>,pair<int,int> Quote: > > tuple;
for (int i=0;i<8;++i) get(tuple,i) = 2*i; for (int i=0;i<8;++i) std::cerr << get(tuple,i) << std::endl; Again, not that I recommend any of this hackery :)
Ok, back to productive work with me.
Ken
|
| Tue, 23 Aug 2005 04:55:48 GMT |
|
|
|
c nicholso
#13 / 15
|
std::vector<int[2]> intializing trouble
Quote: > Err...that can't work with a runtime value though. Right.
Right, until you make a lisp-ish cascading macro like Loki's type-list creators that will access the Nth element of any composed pair construct... too gross to think about.... charles nicholson
Quote: > Empirical testing shows that the element is accessed directly when the > index is a constant integer and the function is inlined. > Taking a dynamic index does take O(N) time, but the entire process is
> inlined (at least at a size 6 tuple).
> Ken
> > Actually, it's not technically recursive, since each call is to a
> > different instance of the function. It will create N functions, where
> N
> > is the size of your tuple, but I wouldn't be surprised if the compiler
> > was able to completely inline the access. (I wouldn't be surprised if
> > it didn't, either.)
> > Ken
> > > That would work, though it makes accessing n-th element an operation
> > of
> > > O(n) complexity instead of O(1). Also, recursive functions typically
> > > don't get inlined, so it's going to be slow and bloated.
> > > --
> > > With best wishes,
> > > Igor Tandetnik
> > > "For every complex problem, there is a solution that is simple,
> neat,
> > > and wrong." H.L. Mencken
> > > > Probably something along the lines of (after changing the format
> > from
> > > > pair<pair<>,int> to pair<int,pair<> >):
> > > > template <typename tuple>
> > > > int& access_element(tuple& t, int n) {
> > > > if (n == 0) return t.first; else return
> > > access_element(t.second,n-1);
> > > > }
> > > > template <>
> > > > int& access_element(int& t, int n) {
> > > > if (n != 0) throw std::exception("accessing beyond the end of
> the
> > > > tuple");
> > > > return t;
> > > > }
> > > > pair<int,pair<int,pair<int,int> > > tuple;
> > > > for (int i=0;i<4;++i) access_element(tuple,i) = 2*i;
> > > > for (int i=0;i<4;++i) std::cerr << access_element(tuple,i) <<
> > > std::endl;
> > > > Not that I recommend such a thing :)
> > > > Ken
> > > > > How would you access a k-th element of such a tuple, with k
> > > determined
> > > > > at run-time?
> > > > > --
> > > > > With best wishes,
> > > > > Igor Tandetnik
> > > > > "For every complex problem, there is a solution that is simple,
> > > neat,
> > > > > and wrong." H.L. Mencken
> > > > > > what's wrong with
> > > > > > vector< pair< pair< pair< pair<pair< pair<pair< pair<pair<
> > > > pair<pair<
> > > > > > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> > pair<pair<
> > > > > pair<pair<
> > > > > > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> > pair<pair<
> > > > > pair<pair<
> > > > > > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> > pair<pair<
> > > > > pair<pair<
> > > > > > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> > pair<pair<
> > > > > pair<pair<
> > > > > > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> > pair<pair<
> > > > > pair<pair<
> > > > > > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> > pair<pair<
> > > > > pair<pair<
> > > > > > pair<pair< pair<pair< pair<pair< pair<pair< pair<pair<
> > pair<pair<
> > > > > pair<pair<
> > > > > > pair<pair< pair<pair< pair<pair< pair<?
> > > > > > maybe you could simplify it with recursive macros!
> > > > > > =]
> > > > > > charles nicholson
> > > > > > > > vector<pair<int, int> > won't work as I have a C array
> that
> > is
> > > > 928
> > > > > > > elements.
> > > > > > > > I only used [2] as a simple example.
> > > > > > > vector<vector<int> > is probably your best choice then.
> > > > > > > Ken
|
| Tue, 23 Aug 2005 05:18:14 GMT |
|
|
|
Ken Alverso
#14 / 15
|
std::vector<int[2]> intializing trouble
Quote:
> > Err...that can't work with a runtime value though. Right. > Right, until you make a lisp-ish cascading macro like Loki's type-list
> creators that will access the Nth element of any composed pair
construct...
> too gross to think about....
Actually, that's more or less what I did. I meant it can't work in constant time with a runtime index. It works just fine with a runtime index in linear time. Or a compile
time index in constant time. I imagine Loki's access speed is similar.
Ken
|
| Tue, 23 Aug 2005 05:30:19 GMT |
|
|
|
Igor Tandetni
#15 / 15
|
std::vector<int[2]> intializing trouble
Quote: > It works just fine with a runtime index in linear time. Or a compile
> time index in constant time. I imagine Loki's access speed is
similar. Loki works on lists of types, not on lists of values. Type manipulation
is always done compile-time. So, it makes compile time linear or worse,
but no run-time code is actually generated.
--
With best wishes,
Igor Tandetnik
"For every complex problem, there is a solution that is simple, neat,
and wrong." H.L. Mencken
|
| Tue, 23 Aug 2005 05:48:37 GMT |
|
| |
|
