On Sun, Jun 26, 2011 at 8:42 AM, Paul Mensonides <pmenso57@comcast.net> wrote:

As Hartmut indicated, Chaos was developed after (and was loosely based on) Boost.Preprocessor.  However, unlike Boost.Preprocessor, Chaos targets standard C and C++ (not particular compilers) and is therefore not limited by broken preprocessors (e.g. VC++).

OK.

...

PP_STRING_CAT( b o o s t ) // expand to `boost` PP_STRING_AT(3, b o o s t ) // expand to `s` etc...

If yes, why were these pp-strings not ported from CHAOS to Boost.Preprocessor?

Any comment on "pp comma-separated strings" support (of lack of thereof) in Boost.Preprocessor and/or CHAOS? Thanks! --Lorenzo

Paul Mensonides wrote:

On Sun, 26 Jun 2011 11:00:16 -0400, Lorenzo Caminiti wrote:

...

Any comment on "pp comma-separated strings" support (of lack of thereof) in Boost.Preprocessor and/or CHAOS?

Boost.Preprocessor represents the lowest common denominator in what can be made portable across most C and C++ compilers. The main roadblock to more significant improvement (nowadays) is VC++'s horrible preprocessor.

By "comma-separated strings" I assume you mean something like: a, b, c? Boost.Preprocessor doesn't support them, of course. Chaos doesn't support them in that form (generally speaking, data structures containing open commas = bad design).

I am sorry, I meant to say "*space* separated strings" (not commas). These are tokens like `a b c` where a character can be a-z, A-Z, 0-9, or _ and all characters must be separated from each other by one or more spaces. Concatenation can be used to parse these "space separated strings", for example (leaving aside reentrancy issues, etc): #include <boost/preprocessor.hpp> #define PP_STRING_CHAR_PP_STRING_NULL_NEXT() /* nothing */ #define PP_STRING_CHAR_PP_STRING_NULL PP_STRING_CHAR_PP_STRING_NULL_NEXT( #define PP_STRING_NEXT_c(str) BOOST_PP_CAT(PP_STRING_CHAR_, str) ) #define PP_STRING_CHAR_c (c) PP_STRING_NEXT_c( #define PP_STRING_NEXT_b(str) BOOST_PP_CAT(PP_STRING_CHAR_, str) ) #define PP_STRING_CHAR_b (b) PP_STRING_NEXT_b( #define PP_STRING_NEXT_a(str) BOOST_PP_CAT(PP_STRING_CHAR_, str) ) #define PP_STRING_CHAR_a (a) PP_STRING_NEXT_a( #define PP_STRING_TO_SEQ(str) \ BOOST_PP_CAT(PP_STRING_CHAR_, str PP_STRING_NULL) ) PP_STRING_TO_SEQ( a b c ) // expand to (a) (b) (c) Does CHAOS support these spaced-strings? Where these spaced-strings considered for Boost.Preprocessor? NOTE: This is more a curiosity than anything else because I don't actually have a real use case for these spaced-strings. I researched them because there was a request for a macro like: STRING_TO_TYPE( MyType ) // expand to: typedef mpl::string<'M','y','T','y','p','e'> MyType_id_t; http://boost.2283326.n4.nabble.com/mpl-string-to-type-conversion-td2668494.h... As far as I know, such a macro cannot be implemented but both the following macros can be done (and I'd prefer somewhat (1) over (2)): SP_STRING_TO_TYPE( M y T y p e ) // (1) SEQ_STRING_TO_TYPE( (M)(y)(T)(y)(p)(e) ) // (2) Another example I was looking at is the possibility of removing the parenthesis around arbitrary function names in the signature below: CONTRACT_FUNCTION( void (myfunc) ( int x, (int&) y ) ) // (3) The following could be done (using pp spaced-strings): CONTRACT_FUNCTION( void m y f u n c ( int x, (int&) y ) ) // (4) But at the end I find (4) rather ugly and I chose to only support (3) instead (so no use for spaced-strings here). Thanks a lot! --Lorenzo -- View this message in context: http://boost.2283326.n4.nabble.com/preprocessor-pp-strings-tp3613014p3627742... Sent from the Boost - Dev mailing list archive at Nabble.com.

On Mon, Jun 27, 2011 at 5:34 PM, Mathias Gaunard < mathias.gaunard@ens-lyon.org> wrote:

FYI, You can do STRING_TO_TYPE("MyType") or STRING_TO_TYPE(MyType) in C++0x by expanding it to "MyType"[0], "MyType"[1] etc.

IIRC you can do it even more directly with C++0x user-defined literals, though I don't think any compilers support them yet: "MyType"mpl_string -- -Matt Calabrese

On 27/06/11 22:40, Matt Calabrese wrote:

On Mon, Jun 27, 2011 at 5:34 PM, Mathias Gaunard < mathias.gaunard@ens-lyon.org> wrote:

...

FYI, You can do STRING_TO_TYPE("MyType") or STRING_TO_TYPE(MyType) in C++0x by expanding it to "MyType"[0], "MyType"[1] etc.

IIRC you can do it even more directly with C++0x user-defined literals, though I don't think any compilers support them yet:

"MyType"mpl_string

By my reading of the standard that doesn't work with user defined string literals, only integer and floating point. User defined string literals can only be accessed at runtime. John Bytheway

Paul Mensonides wrote:

On Mon, 27 Jun 2011 06:39:00 -0700, lcaminiti wrote:

...

Paul Mensonides wrote:

...

On Sun, 26 Jun 2011 11:00:16 -0400, Lorenzo Caminiti wrote:

...

Any comment on "pp comma-separated strings" support (of lack of thereof) in Boost.Preprocessor and/or CHAOS?

Boost.Preprocessor represents the lowest common denominator in what can be made portable across most C and C++ compilers. The main roadblock to more significant improvement (nowadays) is VC++'s horrible preprocessor.

By "comma-separated strings" I assume you mean something like: a, b, c? Boost.Preprocessor doesn't support them, of course. Chaos doesn't support them in that form (generally speaking, data structures containing open commas = bad design).

I am sorry, I meant to say "*space* separated strings" (not commas). These are tokens like `a b c` where a character can be a-z, A-Z, 0-9, or _ and all characters must be separated from each other by one or more spaces. Concatenation can be used to parse these "space separated strings", for example (leaving aside reentrancy issues, etc):

Chaos supports such data structures. E.g.

#include <chaos/preprocessor/algorithm/reverse.h> #include <chaos/preprocessor/generics/strip.h> #include <chaos/preprocessor/string/core.h>

CHAOS_PP_REVERSE( (CHAOS_PP_STRING) a b c d e f ) // (CHAOS_PP_STRING) f e d c b a

CHAOS_PP_STRING_TO_UPPER( CHAOS_PP_STRIP( CHAOS_PP_REVERSE( (CHAOS_PP_STRING) a b c d e f ) ) ) // F E D C B A

Yep, this is pretty cool :)) especially the `(pp-data-type) pp-data` syntax. Great job with Chaos!! What will it take to port the (splendid) Chaos library into Boost? "Just" fixing VC++ pp or more? Btw, are Chaos docs online in HTML form somewhere? --Lorenzo -- View this message in context: http://boost.2283326.n4.nabble.com/preprocessor-pp-strings-tp3613014p3630324... Sent from the Boost - Dev mailing list archive at Nabble.com.