gmane.comp.parsers.spirit.general

Re: BOOST_FUSION_DEFINE_STRUCT at class/namespacescope

Joel de Guzman — 2012-05-18T03:52:51

Thanks, Nate. I'll apply review and patch ASAP.

Regards,

Re: BOOST_FUSION_DEFINE_STRUCT at class/namespace scope

Nathan Ridge — 2012-05-18T00:12:11


Attached is an updated patch that includes the TPL version.

Regards,
Nate
       ------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/_______________________________________________
Spirit-general mailing list
Spirit-general< at >lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/spirit-general

Re: [boost] [fusion] Trouble using BOOST_FUSION_ADAPT_ADT with boost::fusion::operator<

Nathan Ridge — 2012-05-17T23:09:22


Tests are attached. 


In the process of writing the tests, I discovered that only sequence
elements defined in a different namespace than the sequence are affected
(e.g. std::string in the OP's example). Here's why:


Suppose the sequence is of type N::Seq and its i'th element of type 
N::Elem, i.e. they are in the same namespace. When trying to resolve

operator<(adt_attribute_proxy<N::Seq, i>, adt_attribute_proxy<N::Seq, i>)

N is an associated namespace of adt_attribute_proxy<N::Seq, i>, and so ADL 
finds N::operator<(N::Elem, N::Elem) (assuming there is one). Since
adt_attribute_proxy<N::Seq, i> is convertible to N::Elem, the conversions
take place and it's a match.


On the other hand, suppose the sequence is of type N::Seq and its i'th
element of type M::Elem, i.e. they are in different namespaces. When
trying to resolve

operator<(adt_attribute_proxy<N::Seq, i>, adt_attribute_proxy<N::Seq, i>)


M is *not* an associated namespace of adt_attribute_proxy<N::Seq, i>
(even though adt_attribute_proxy<N:

Re: BOOST_FUSION_DEFINE_STRUCT at class/namespace scope

Nathan Ridge — 2012-05-17T07:10:20


Attached is a patch, including tests and documentation.

I haven't written a TPL version yet; I will write one and submit it in
a future patch.

Regards,
Nate
       ------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/_______________________________________________
Spirit-general mailing list
Spirit-general< at >lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/spirit-general

utree_cast compile issue

Nausca — 2012-05-17T06:30:25

Hi,
in utree detail,
there is a utree_cast template class which throw an exception in not 
convertible cast,
then return an object to the caller.
        template <typename From>
        To dispatch(From const&, boost::mpl::false_) const
        {
            throw std::bad_cast();
            return To();
        }

in some version og gcc, when:
TO=boost::spirit::any_ptr&
the compiler gives error: invalid value-initialization of reference types


I modify the code to define a new template class to handle the any_ptr& bad 
cast condition:

    template <typename To>
    struct utree_castobj {
    To dispatch() const {return To();}  };
    template <typename To>
    struct utree_castobj<To&> {
    To myTo;
    To& dispatch() {return myTo;}       };

and redefine the dispatch function of utree_cast:
        template <typename From>
        To dispatch(From const&, boost::mpl::false_) const
        {
            throw std::bad_cast();
            utree_castobj<To> mydispatch;
            return mydispatch.dispa

Re: [boost] [fusion] Trouble using BOOST_FUSION_ADAPT_ADT with boost::fusion::operator<

Nathan Ridge — 2012-05-17T01:32:50


Will do.

Regards,
Nate
       
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/

FW: [boost] [fusion] Trouble using BOOST_FUSION_ADAPT_ADT with boost::fusion::operator<

Nathan Ridge — 2012-05-16T20:02:10



----------------------------------------
       ------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/_______________________________________________
Spirit-general mailing list
Spirit-general< at >lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/spirit-general

Re: with BOOST :: SPIRIT, how to parse a math expression into syntax tree?

Johannes Lorenz — 2012-05-16T15:05:57

Hey,

I recently wrote this code. Supports variables x,y,v, +-*/() and kronecker 
delta. Please note that it is GPL3'd though. Works with boost 1_46 and 
probably newer. Code is appended.

A parsing example:

        client::expression_ast ast;
        if(! build_tree_from_equation(equation, &ast))
         puts("Could not parse equation string - Terminating\n");

        client::ast_print solver(x,y,grid[internal]); // pass x,y,v
        return solver(ast);

Hope it helps!

Greetings,
Johannes

Re: BOOST_FUSION_DEFINE_STRUCT at class/namespacescope

Joel de Guzman — 2012-05-16T10:52:51

I'm definitely interested. Nice one, Nathan! Are there docs yet?
Sample code would be lovely.

Regards,

Re: with BOOST :: SPIRIT, how to parse a math expression into syntax tree?

Seth Heeren — 2012-05-16T07:04:31

With all due respect, this is not a 'free programmer service'.

The samples in the repo are a good starting point. Look at the various 
ones named calc* ?

That is not a real question. There isn't a tree. You just manipulate the 
root like you would any other node.
Perhaps you need to define what the problem is. What manipulation are 
you trying to do, that you can't figure out how to?

Cheers,
Seth
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/_______________________________________________
Spirit-general mailing list
Spirit-general< at >lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/spirit-general

BOOST_FUSION_DEFINE_STRUCT at class/namespace scope

Nathan Ridge — 2012-05-16T03:06:26

Hi,

I have written a version of BOOST_FUSION_DEFINE_STRUCT that can be
used at namespace or class scope (whereas BOOST_FUSION_DEFINE_STRUCT
itself can only be used at global scope). It works by having the
generated class derive from sequence_facade, much like 'struct 
triple' in libs/fusion/example/extension/triple.cpp (except that in
my implementation, the iterator class is nested inside the generated 
class).

I don't think my version is suitable as a replacement for
BOOST_FUSION_DEFINE_STRUCT, because in C++03 it makes the generated
struct non-POD (since it has a base class), and the POD-ness of
the generated struct may be important for some people. However, I
think it would serve as a useful alternative to people who don't
need POD-ness (or who use C++11, where having a base class does not
inhibit POD-ness) but want to be able to define Fusion sequences
that are nested inside another class.

Is there any interest in adding this alternative version to Fusion?
If so, I am happy to Boostify my implementat

Re: with BOOST :: SPIRIT, how to parse a math expression into syntax tree?

yuyan taiji — 2012-05-16T02:09:27

yes ,of course,all   math expression

On 15 May 2012 16:41, Kostya Kanishev <kkanishev< at >gmail.com> wrote:

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/_______________________________________________
Spirit-general mailing list
Spirit-general< at >lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/spirit-general

Re: Problem creating a std::complex from two parsed floats

Seth Heeren — 2012-05-15T15:48:09

Good idea.

It's not about ignorance, it is about clear questions, mostly :)

Two things:

  * refer to http://www.boost.org/community/error_handling.html "How to 
design my exception classes" on how to properly design your custom 
exception.

  * refer to [Auto-rule assignment semantics 
<http://www.boost.org/doc/libs/1_49_0/libs/spirit/doc/html/spirit/qi/reference/nonterminal/rule.html#spirit.qi.reference.nonterminal.rule.expression_semantics>] 
(operator%= instead of operator=)

With these two _tiny_ tweaks the things compiles. I haven't looked any 
further for this moment.

If you have any other questions, feel free to ask

PS.: Also warmly recommended: 
http://stackoverflow.com/questions/tagged/boost-spirit

Seth
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile se

Problem creating a std::complex from twoparsed floats

Stephen Torri — 2012-05-15T14:42:26

I am using Boost 1.48 on Win 7 using Visual Studio 2010 C++ Express.

The problem is
1>------ Build started: Project: Debug_Phoenix_Ref, Configuration:
Debug Win32 ------
1> main.cpp
1>c:\boost\boost_1_48_0\boost\spirit\home\qi\detail\assign_to.hpp(123):
error C2440: 'static_cast' : cannot convert from 'const attr_type' to
'std::complex<float>'

Now it seems that the only way for me to move forward is to post a
message about my problem so that someone can post a simple reply
exposing my ignorance. What I am trying to do is parse a sequence of
complex numbers written in a "float,float\n" format in a text file.
One complex number per line. I don't see the problem with what I have
written but given a lengthy delay since the last time I have used
Boost Spirit anything is possible.

File is attached.

Regards,
Stephen
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape ha

Re: with BOOST :: SPIRIT, how to parse a math expression into syntax tree?

Kostya Kanishev — 2012-05-15T08:41:10

Do you need also to parse correctly this?
2*3+1*(3+5)
2012/5/15 yuyan taiji <taijiyuyan< at >gmail.com>

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/_______________________________________________
Spirit-general mailing list
Spirit-general< at >lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/spirit-general

with BOOST :: SPIRIT,how to parse a math expression into syntax tree?

yuyan taiji — 2012-05-15T04:49:48

with BOOST :: SPIRIT, how to parse a math expression into syntax tree?
eg.
1+2*3*(3+5)==>
+
1 *
  2 *
    3 +
      3 5

in spirit,how to manipulate the parent node of the tree ?
tks.
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/_______________________________________________
Spirit-general mailing list
Spirit-general< at >lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/spirit-general

(no subject)

Stephen Torri — 2012-05-14T19:34:18

I am trying to understand how I can adapt Rob Stewart's "Dispatching
on Expectation Point Failures". I have the preliminary functionality
coded up in the attached file.


Problem space: Trying to parse in a table of complex variables. The
file begins with one line stating the number of columns, number of
rows and if it is complex data or not. After that the data is read in
column oriented format. Each column value is written in a double value
followed by comma followed by a double value and ended with a newline
character. The idea is that in the end a boost multi_array containing
all the complex values converted from pairs of floats.


(Good input)

If I change the input to read:



const char* input_text = "cols = 1.0, rows = 2, complex =
1\n0.000000,0.000000\n0.000070,-0.000108";



I am expected a 2x1 matrix with the values { (0.0,0.0i), (0.00007,0.000108i) }.



(Bad input)

Now if for some reason the input string was:

const char* input_text = "col = 1.0, rows = 2, complex =
1\n0.000000,0.000000\n0.0000

Re: Typo in docs for BOOST_FUSION_ADAPT_TPL_ADT

Joel de Guzman — 2012-05-14T00:41:55

Patch applied. Thanks, Mario!

Regards,

Re: [fusion] [range] range::erase vs. fusion::erase

Joel de Guzman — 2012-05-14T00:38:42


Patches applied. Thanks, Nate!

Regards,

Typo in docs for BOOST_FUSION_ADAPT_TPL_ADT

Mario Lang — 2012-05-13T21:48:11

Hi.

I found the following minor typo in the synposis of BOOST_FUSION_ADAPT_TPL_ADT.
Maybe someone can/wants to apply this?

(relative to libs/fusion)
Index: doc/adapted.qbk
===================================================================
--- doc/adapted.qbk(revision 78461)
+++ doc/adapted.qbk(working copy)
< at >< at > -658,7 +658,7 < at >< at >
 
 [heading Synopsis]
 
-    BOOST_FUSION_ADAPT_ADT(
+    BOOST_FUSION_ADAPT_TPL_ADT(
         (template_param0)(template_param1)...,
         (type_name) (specialization_param0)(specialization_param1)...,
         (attribute_type0, attribute_const_type0, get_expr0, set_expr0)

Re: [fusion] [range] range::erase vs. fusion::erase

Nathan Ridge — 2012-05-13T03:27:06


Attached. (I know it's been a while since we discussed this, but better late
than never right?)

Other affected functions besides erase were copy, replace, replace_if,
reverse, count, count_if, for_each, insert, push_back, push_front, and
accumulate. I fixed all of them.

I have verified that all Fusion tests pass with these changes.

Thanks,
Nate
       ------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/_______________________________________________
Spirit-general mailing list
Spirit-general< at >lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/spirit-general