tr1

[Noah Roberts]

Is there a way to get just tr1 from boost? I would like to get a tr1
implementation independant of other stuff...the absolute minimum
required.

 

[P.J. Plauger]

Is there a way to get just tr1 from boost? I would like to get a tr1
implementation independant of other stuff...the absolute minimum
required.

That's nice, but TR1 adds nontrivial stuff to several existing C++
headers, such as <functional>, <memory>, and <utility>. And I'm
guessing that by TR1 you really mean the Boost-y stuff, not all the
additions required by special math and C99 compatibility. If you
want all that, you've got tons of additions to <cmath> and <math.h>,
plus sprinklings in other headers besides. So what you really need
is TR1 *integrated* with the Standard C++ library, not *isolated*
from it. In that case, we're the only game in town. See our Compleat
Library at our web site.

P.J. Plauger
Dinkumware, Ltd.
http://www.dinkumware.com

 

[Howard Hinnant]

Is there a way to get just tr1 from boost? I would like to get a tr1
implementation independant of other stuff...the absolute minimum
required.

I'm not positive, but I suspect that the only full implementation of
std::tr1 that's available is from Dinkumware
(http://www.dinkumware.com/tr1.aspx). You can get many parts of tr1
from boost, or perhaps from gcc or CodeWarrior (depending on your
platform needs). The special math functions are especially challenging
and Dinkumware has put considerable effort into their R&D. If you're
looking for a freebie, you're probably going to have to specify which
parts of tr1 you want (and on what platform/compiler).

-Howard

 

[Noah Roberts]

I'm not positive, but I suspect that the only full implementation of
std::tr1 that's available is from Dinkumware
(http://www.dinkumware.com/tr1.aspx). You can get many parts of tr1
from boost, or perhaps from gcc or CodeWarrior (depending on your
platform needs). The special math functions are especially challenging
and Dinkumware has put considerable effort into their R&D.

I don't want the special math functions; they aren't part of the draft
standard at this point. C99 is also not a big need as I try to use the
C++ part of C++. So really what I want is the added functionality
offered by boost's additions in a nice box. We may look at
dinkumware's version as we already use their library but at this point
I don't need it.

So, if anyone knows the answer to my question that would be great.

 

[Tom Widmer]

Is there a way to get just tr1 from boost? I would like to get a tr1
implementation independant of other stuff...the absolute minimum
required.

The boost tr1 stuff depends on a lot of the other boost stuff AFAIK.
However, you could either get the current version of boost from their
CVS system, or wait for the next boost release. That has a specific tr1
library:
http://www.boost.org/regression-logs/cs-win32_metacomm/doc/html/boost_tr1.html

Once you've got that, you could probably delete much of the rest of
boost that doesn't have anything to do with tr1.

Tom

 

[Noah Roberts]

The boost tr1 stuff depends on a lot of the other boost stuff AFAIK.
However, you could either get the current version of boost from their
CVS system, or wait for the next boost release. That has a specific tr1
library:
http://www.boost.org/regression-logs/cs-win32_metacomm/doc/html/boost_tr1.html

Once you've got that, you could probably delete much of the rest of
boost that doesn't have anything to do with tr1.

Well, from what I can tell the tr1 library in boost just imports the
boost stuff into std::tr1. That's fine, I would like to be able to
just get the minimum required for the tr1 dependencies though. Looks
like one has to weed it out.

 

[Tom Widmer]

Well, from what I can tell the tr1 library in boost just imports the
boost stuff into std::tr1. That's fine, I would like to be able to
just get the minimum required for the tr1 dependencies though. Looks
like one has to weed it out.

I recall there was discussion of some kind of tool to generate a
dependency graph for the whole of boost. I don't know whether it
actually exists or not. Otherwise, you could write one!

Tom