tag:old.nabble.com,2006:forum-3430Nabble - The Spirit Parser Library2009-12-20T16:28:09ZSpirit is an object oriented recursive descent parser generator framework implemented using template meta-programming techniques. The Spirit Parser Library home is <a href="http://sourceforge.net/projects/spirit/" target="_top" rel="nofollow">here</a>.tag:old.nabble.com,2006:post-26868661Re: permutation with all elements required2009-12-20T16:28:09Z2009-12-20T16:28:09ZKenny Riddile<table cellspacing="0" cellpadding="0" border="0"><tr><td valign="top" style="font: inherit;">--- On <b>Fri, 12/18/09, Joel de Guzman <i><<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26868661&i=0" target="_top" rel="nofollow">joel@...</a>></i></b> wrote:<br><blockquote style="border-left: 2px solid rgb(16, 16, 255); margin-left: 5px; padding-left: 5px;"><br>From: Joel de Guzman <<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26868661&i=1" target="_top" rel="nofollow">joel@...</a>><br>Subject: Re: [Spirit-general] permutation with all elements required<br>To: <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26868661&i=2" target="_top" rel="nofollow">spirit-general@...</a><br>Date: Friday, December 18, 2009, 10:02 PM<br><br><div class="plainMail">Carl Barron wrote:<div class='shrinkable-quote'><br>> On Dec 18, 2009, at 8:25 PM, Joel de Guzman wrote:<br>> <br>>> Carl Barron wrote:<br>>>> On Dec 18, 2009, at 7:13 PM, Joel de Guzman wrote:<br>>>>> The attribute of the permutation operator given a rule<br>>>>><br>>>>> a ^ b ^ c<br>>>>><br>>>>> is<br>>>>><br>>>>>
vector<optional<A>, optional<B>, optional<C> ><br>>>>><br>>>>> After a successful match, you just check that all optionals<br>>>>> for A, B and C are initialized (see Boost Optional).<br>>>>><br>>>> vwctor< ...> or tuple<...> ?<br>>> Sorry, yeah, tuple. I'm just used to fusion::vector.<br>> <br>> Then phoenix and fusion make this fairly easy for an an arbitrary # of op ^()'s in sequence<br>> a1 ^ a2 ^ .... ^a_n, with one phoenix function for any n.<br>> <br>> use fusion::find_if to test for any optional containing 'nothing' and compare the result to<br>> fusion::end. in a phoenix function.<br>> <br>> Something like:<br>> <br>> class no_empties_impl<br>> {<br>>
struct optional_empty<br>> {<br>> template <class A><br>> bool operator () (const A &a) const<br>> {<br>> // return true if a is 'empty'.<br>> return !a;<br>> }<br>> };<br>> public:<br>> template <class A><br>> struct result {typedef bool type;};<br>> <br>> template <class A><br>> bool operator () (const A &a) const<br>> {<br>>
// look for an empty optional, if any return false else true.<br>> return boost::fusion::find_if<optional_empty>(a) == boost::fusion::end(a);<br>> }<br>> };<br>> <br>> Looks like an general implimentation of no_empties<br>> <br>> foo %= a1 ^... ^a_n [_pass = no_empties(qi::_0)]; // or is it _1?? the fusion seq,<br>> <br>> should provide result for any n. permutation of n parsers n at a time.</div><br>Use fusion::all instead of find_if: <a href="http://tinyurl.com/yzm94gc" target="_blank" rel="nofollow">http://tinyurl.com/yzm94gc</a><br><br>Regards,<br>-- <br>Joel de Guzman<br><a href="http://www.boostpro.com" target="_blank" rel="nofollow">http://www.boostpro.com</a><br><a href="http://spirit.sf.net" target="_blank" rel="nofollow">http://spirit.sf.net</a><br><a href="http://www.facebook.com/djowel" target="_blank" rel="nofollow">http://www.facebook.com/djowel</a><br><br>Meet me at BoostCon<br><a href="http://www.boostcon.com/home" target="_blank" rel="nofollow">http://www.boostcon.com/home</a><br><a href="http://www.facebook.com/boostcon" target="_blank" rel="nofollow">http://www.facebook.com/boostcon</a><br><br><br><br>------------------------------------------------------------------------------<br>This SF.Net email is sponsored by the Verizon Developer Community<br>Take advantage of Verizon's best-in-class app development support<br>A streamlined, 14 day to market process makes app distribution fast and easy<br>Join now and get one step closer to millions of Verizon customers<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_blank" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a> <br>_______________________________________________<br>Spirit-general mailing list<br><a ymailto="mailto:Spirit-general@lists.sourceforge.net" href="/mc/compose?to=Spirit-general@lists.sourceforge.net" target="_top" rel="nofollow">Spirit-general@...</a><br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_blank" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a></div><div class="plainMail"><br></div></blockquote>I'm getting a compile error when attempting something similar to this. Here's my code:<div><br></div><div><div>namespace detail</div><div>{</div><div><br></div><div> class RequireAllImpl</div><div> {</div><div> private:</div><div> struct IsOptionalPopulated</div><div> {</div><div> template< typename T ></div><div> bool operator()( const T& t ) const</div><div> {</div><div>
return t;</div><div> }</div><div> };</div><div> </div><div> public:</div><div> template< typename T ></div><div> struct result</div><div> {</div><div> typedef bool type;</div><div> };</div><div> </div><div> template< typename T ></div><div> bool operator()( const T& t ) const</div><div> {</div><div> return boost::fusion::all( t, IsOptionalPopulated() ); // no error if I return a constant
here</div><div> }</div><div> };</div><div><br></div><div>} // end namespace detail</div><div><br></div><div>template< typename T ></div><div>bool RequireAll( const T& t )</div><div>{</div><div> return detail::RequireAllImpl()( t );</div><div>}</div><div><br></div><div>I call this like so:</div><div><br></div><div>foo %= (a ^ b ^ c)[_pass = RequireAll(_0)];</div><div><br></div><div>and get this error from boost/fusion/support/detail/category_of.hpp:</div><div><br></div><div>error C2039: 'category' : is not a member of 'boost::phoenix::actor<Eval>'</div><div><br></div><div>Am I missing an include?</div><div><br></div><div><br></div></div></td></tr></table><br />------------------------------------------------------------------------------
<br>This SF.Net email is sponsored by the Verizon Developer Community
<br>Take advantage of Verizon's best-in-class app development support
<br>A streamlined, 14 day to market process makes app distribution fast and easy
<br>Join now and get one step closer to millions of Verizon customers
<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_top" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a> <br />_______________________________________________
<br>Spirit-general mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26868661&i=3" target="_top" rel="nofollow">Spirit-general@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a><br><p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26867902inputs of models with ranges of values, parsing from text file input streams2009-12-20T14:38:14Z2009-12-20T14:38:14Zhm567Hello,
My need is relatively simple and I am not sure I required the full power of spirit to achieve my target.
<br>
However let me pose the simple issue here:<br>
I have models that take inputs that are of boolean/arithmetic types and produce outputs.<br>
All models have a base class, and each model inputs is a class that derives from model_inputs abstract base class.<br>
The inputs for these models can be a set of fixed values (bool or arithmetic) or the same set but of ranges of values. The problem is simple in that if a model required 7 inputs, it requires all of them, but any number between 0 and 7 inputs can be a range of values.<br>
Typically, I wish to be able to write once<br>
<b>struct model1_inputs_fixed {<br>
double i1;<br>
double i2;<br>
int i3;<br>
static const size_t number_inputs = 3;<br>
};</b><br>
and have some mechanism to generate some other struct, that would embed members and flags.<br>
These flags tell whether the object holds a fixed value for some input, or a range of values.<br>
An example may be:<br>
<b>// fixed or range inputs<br>
struct model1_inputs : public model_inputs {<br>
bool isi1fixed; double i1f; std::pair<double, double> i1r;<br>
bool isi2fixed; double i2f; std::pair<double, double> i2r;<br>
bool isi3fixed; double i3f; std::pair<double, double> i3r;<br>
};</b><br>
There might be up to 100 different model_inputs classes, so memory layout and runtime performance are not essential in this part of the code, rather clarity and convenience are at a premium<br>
The next step is to be able to I/O these types from text files like:<br>
<br>
i1 = 5<br>
i2 = 14..16<br>
i3 = 0<br>
If an input is missing, repeated or invalid, the inputting function should throw an exception.<br>
The order of the inputs shouldn't matter.<br>
At this stage, I don't see that my models may require more evolved, but if there is a solution that would allow better to accept more complex inputs, it may be worth for me spending the time to learn it.<br>
<br>
Is spirit a reasonable tool for what I think is a simple task?<br>
<br>
I see other libraries that may provide the parsing part (program_options, property_tree) but not the compile-time generation of the model1_inputs and the compile_time parsing from text file of these types.<br>
I will follow the post at http://lists.boost.org/boost-users/2009/12/54770.php as a start, that seems to do what I want, except for the range functionality.<br> Is the ".." convention reasonable?<br>
<br>
Regards,<br><p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26861064Re: Lex & Qi question2009-12-19T21:01:00Z2009-12-19T21:01:00ZCARL BARRON-3<br>On Dec 19, 2009, at 11:12 PM, Jason McKesson wrote:
<br><div class='shrinkable-quote'><br>> On 12/19/2009 5:35 PM, Carl Barron wrote:
<br>>> On Dec 19, 2009, at 3:27 PM, Jason McKesson wrote:
<br>>>
<br>>>
<br>>>> I'm trying to write a simple expression parser as the first steps
<br>>>> towards something bigger. And I'd like to be able to use Lex to break
<br>>>> the input up into tokens. My problem is that combining the two seems to
<br>>>> be constraining on the utility of Qi.
<br>>>>
<br>>>> Here's what I mean. Let's say I have a lexer that looks something like this:
<br>>>>
<br>>>> template<typename Lexer>
<br>>>> struct math_tokens : lex::lexer<Lexer>
<br>>>> {
<br>>>> math_tokens()
<br>>>> {
<br>>>> this->self.add
<br>>>> ("\\+", ID_PLUS)
<br>>>> ("\d+f?", ID_FLOAT)
<br>>>> ("\d*.\d+f?", ID_FLOAT)
<br>>>> ("\d+", ID_INTEGER)
<br>>>> ;
<br>>>> }
<br>>>> };
<br>>>>
<br>>>> A parser for this would look something like this:
<br>>>>
<br>>>> template<typename Iterator>
<br>>>> struct math_grammar : qi::grammar<Iterator>
<br>>>> {
<br>>>> template<typename TokenDef>
<br>>>> math_grammar(TokenDef const& tok)
<br>>>> : math_grammar::base_type(start)
<br>>>> {
<br>>>> terminals = qi::token<ID_FLOAT> | qi::token<ID_INTEGER>;
<br>>>>
<br>>>> start = terminals>> qi::token<ID_PLUS> >> terminals;
<br>>>> }
<br>>>>
<br>>>> qi::rule<Iterator> start;
<br>>>> qi::rule<Iterator> terminals;
<br>>>> };
<br>>>>
<br>>>> I haven't tested this specific code, but you have the general idea of
<br>>>> the pattern I'm interested in.
<br>>>>
<br>>>> The problem is this. If I were to build this purely without lex, the
<br>>>> grammar would look something like this:
<br>>>>
<br>>>> template<typename Iterator>
<br>>>> struct math_qi_grammar : qi::grammar<Iterator>
<br>>>> {
<br>>>> math_qi_grammar()
<br>>>> : math_grammar::base_type(start)
<br>>>> {
<br>>>> terminals = integer_ | double_;
<br>>>>
<br>>>> start = terminals>> '+'>> terminals;
<br>>>> }
<br>>>>
<br>>>> qi::rule<Iterator> start;
<br>>>> qi::rule<Iterator> terminals;
<br>>>> };
<br>>>>
<br>>>> The difference is this. If I put a semantic action on 'integer_', the
<br>>>> action gets the synthesized attribute value as a parameter. This
<br>>>> attribute is an integer. If I put a semantic action on
<br>>>> 'qi::token<ID_INTEGER>', the synthesized attribute is a string. This is
<br>>>> not as useful as getting an integer, as I now must find a way to convert
<br>>>> this string into an integer.
<br>>>>
<br>>>> Is there some way to combine Lex and Qi that still provides useful
<br>>>> synthesized attribute behavior of Qi? Is there something I can wrap the
<br>>>> "qi::token<ID_INTEGER>" parser in that will provide a properly
<br>>>> synthesized attribute?
<br>>>>
<br>>>
<br>>>
<br>>> Spirit::lex is desigmed to be used with spirit::qi. a token_def<X> holds a boost::variant< boost::interator_range<SourceIteratorType>, X>. the attribute of a token_def<X> is X. The conversion from an iterator range to X occurs when the token is parsed successfully.
<br>>>
<br>>> Use naned token defs like
<br>>>
<br>>> math_tokens()
<br>>> {
<br>>> plus = '+'; // plus is a literal +.
<br>>> integer = ...; // integer
<br>>> real = ...; // real number;
<br>>> // now integer must be added to self before real so a real is a real that is not an integer.
<br>>> this->self = plus | integer | real;
<br>>> }
<br>>> lex::token_def<> plus;
<br>>> lex::token,def<int> integer;
<br>>> lex::token_def<double> real;
<br>>> };
<br>>>
<br>>> now in your qi::grammar you refer to the tokens by name they will be converted to their attribute
<br>>> values when qi first successes in matching the token.
<br>>>
<br>>> start = terminals>> tok.plus>> terminals; // EA tuple<variant,variant>
<br>>> terminals = tok.integer | tok.real; // EA variant<int,double>
<br>>>
<br>>> Does this make any sense?
<br>>>
<br>>
<br>> Yes it does. Thanks.
<br>>
<br>> Obviously, there's some code somewhere that converts the iterator range
<br>> into the token_def's type. Can you hook something into this to do
<br>> specialized processing for generating an arbitrary attribute?
</div> Yes via assign_to_attribute_from_iterator specialization in boost::spirit::traits.
<br><br> The parser api in token_def<X> calls a spirit function assign_to which uses the hook
<br> assign_to_attribute_from_iterators to do the actual assignment. See the advanced section
<br>on customization points for details on customization points [ often denoted CP on this list]
<br><br><br><br><br>------------------------------------------------------------------------------
<br>This SF.Net email is sponsored by the Verizon Developer Community
<br>Take advantage of Verizon's best-in-class app development support
<br>A streamlined, 14 day to market process makes app distribution fast and easy
<br>Join now and get one step closer to millions of Verizon customers
<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_top" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a>
<br>_______________________________________________
<br>Spirit-general mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26861064&i=0" target="_top" rel="nofollow">Spirit-general@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a><br><p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26860894Re: Lex & Qi question2009-12-19T20:12:51Z2009-12-19T20:12:51ZJason McKessonOn 12/19/2009 5:35 PM, Carl Barron wrote:
<div class='shrinkable-quote'><br>> On Dec 19, 2009, at 3:27 PM, Jason McKesson wrote:
<br>>
<br>>
<br>>> I'm trying to write a simple expression parser as the first steps
<br>>> towards something bigger. And I'd like to be able to use Lex to break
<br>>> the input up into tokens. My problem is that combining the two seems to
<br>>> be constraining on the utility of Qi.
<br>>>
<br>>> Here's what I mean. Let's say I have a lexer that looks something like this:
<br>>>
<br>>> template<typename Lexer>
<br>>> struct math_tokens : lex::lexer<Lexer>
<br>>> {
<br>>> math_tokens()
<br>>> {
<br>>> this->self.add
<br>>> ("\\+", ID_PLUS)
<br>>> ("\d+f?", ID_FLOAT)
<br>>> ("\d*.\d+f?", ID_FLOAT)
<br>>> ("\d+", ID_INTEGER)
<br>>> ;
<br>>> }
<br>>> };
<br>>>
<br>>> A parser for this would look something like this:
<br>>>
<br>>> template<typename Iterator>
<br>>> struct math_grammar : qi::grammar<Iterator>
<br>>> {
<br>>> template<typename TokenDef>
<br>>> math_grammar(TokenDef const& tok)
<br>>> : math_grammar::base_type(start)
<br>>> {
<br>>> terminals = qi::token<ID_FLOAT> | qi::token<ID_INTEGER>;
<br>>>
<br>>> start = terminals>> qi::token<ID_PLUS> >> terminals;
<br>>> }
<br>>>
<br>>> qi::rule<Iterator> start;
<br>>> qi::rule<Iterator> terminals;
<br>>> };
<br>>>
<br>>> I haven't tested this specific code, but you have the general idea of
<br>>> the pattern I'm interested in.
<br>>>
<br>>> The problem is this. If I were to build this purely without lex, the
<br>>> grammar would look something like this:
<br>>>
<br>>> template<typename Iterator>
<br>>> struct math_qi_grammar : qi::grammar<Iterator>
<br>>> {
<br>>> math_qi_grammar()
<br>>> : math_grammar::base_type(start)
<br>>> {
<br>>> terminals = integer_ | double_;
<br>>>
<br>>> start = terminals>> '+'>> terminals;
<br>>> }
<br>>>
<br>>> qi::rule<Iterator> start;
<br>>> qi::rule<Iterator> terminals;
<br>>> };
<br>>>
<br>>> The difference is this. If I put a semantic action on 'integer_', the
<br>>> action gets the synthesized attribute value as a parameter. This
<br>>> attribute is an integer. If I put a semantic action on
<br>>> 'qi::token<ID_INTEGER>', the synthesized attribute is a string. This is
<br>>> not as useful as getting an integer, as I now must find a way to convert
<br>>> this string into an integer.
<br>>>
<br>>> Is there some way to combine Lex and Qi that still provides useful
<br>>> synthesized attribute behavior of Qi? Is there something I can wrap the
<br>>> "qi::token<ID_INTEGER>" parser in that will provide a properly
<br>>> synthesized attribute?
<br>>>
<br>>
<br>>
<br>> Spirit::lex is desigmed to be used with spirit::qi. a token_def<X> holds a boost::variant< boost::interator_range<SourceIteratorType>, X>. the attribute of a token_def<X> is X. The conversion from an iterator range to X occurs when the token is parsed successfully.
<br>>
<br>> Use naned token defs like
<br>>
<br>> math_tokens()
<br>> {
<br>> plus = '+'; // plus is a literal +.
<br>> integer = ...; // integer
<br>> real = ...; // real number;
<br>> // now integer must be added to self before real so a real is a real that is not an integer.
<br>> this->self = plus | integer | real;
<br>> }
<br>> lex::token_def<> plus;
<br>> lex::token,def<int> integer;
<br>> lex::token_def<double> real;
<br>> };
<br>>
<br>> now in your qi::grammar you refer to the tokens by name they will be converted to their attribute
<br>> values when qi first successes in matching the token.
<br>>
<br>> start = terminals>> tok.plus>> terminals; // EA tuple<variant,variant>
<br>> terminals = tok.integer | tok.real; // EA variant<int,double>
<br>>
<br>> Does this make any sense?
<br>>
</div><br>Yes it does. Thanks.
<br><br>Obviously, there's some code somewhere that converts the iterator range
<br>into the token_def's type. Can you hook something into this to do
<br>specialized processing for generating an arbitrary attribute?
<br><br>------------------------------------------------------------------------------
<br>This SF.Net email is sponsored by the Verizon Developer Community
<br>Take advantage of Verizon's best-in-class app development support
<br>A streamlined, 14 day to market process makes app distribution fast and easy
<br>Join now and get one step closer to millions of Verizon customers
<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_top" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a>
<br>_______________________________________________
<br>Spirit-general mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26860894&i=0" target="_top" rel="nofollow">Spirit-general@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a><br><p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26860682Re: Lex & Qi question2009-12-19T17:35:08Z2009-12-19T17:35:08ZCARL BARRON-3<br>On Dec 19, 2009, at 3:27 PM, Jason McKesson wrote:
<br><div class='shrinkable-quote'><br>> I'm trying to write a simple expression parser as the first steps
<br>> towards something bigger. And I'd like to be able to use Lex to break
<br>> the input up into tokens. My problem is that combining the two seems to
<br>> be constraining on the utility of Qi.
<br>>
<br>> Here's what I mean. Let's say I have a lexer that looks something like this:
<br>>
<br>> template <typename Lexer>
<br>> struct math_tokens : lex::lexer<Lexer>
<br>> {
<br>> math_tokens()
<br>> {
<br>> this->self.add
<br>> ("\\+", ID_PLUS)
<br>> ("\d+f?", ID_FLOAT)
<br>> ("\d*.\d+f?", ID_FLOAT)
<br>> ("\d+", ID_INTEGER)
<br>> ;
<br>> }
<br>> };
<br>>
<br>> A parser for this would look something like this:
<br>>
<br>> template <typename Iterator>
<br>> struct math_grammar : qi::grammar<Iterator>
<br>> {
<br>> template <typename TokenDef>
<br>> math_grammar(TokenDef const& tok)
<br>> : math_grammar::base_type(start)
<br>> {
<br>> terminals = qi::token<ID_FLOAT> | qi::token<ID_INTEGER>;
<br>>
<br>> start = terminals>> qi::token<ID_PLUS> >> terminals;
<br>> }
<br>>
<br>> qi::rule<Iterator> start;
<br>> qi::rule<Iterator> terminals;
<br>> };
<br>>
<br>> I haven't tested this specific code, but you have the general idea of
<br>> the pattern I'm interested in.
<br>>
<br>> The problem is this. If I were to build this purely without lex, the
<br>> grammar would look something like this:
<br>>
<br>> template <typename Iterator>
<br>> struct math_qi_grammar : qi::grammar<Iterator>
<br>> {
<br>> math_qi_grammar()
<br>> : math_grammar::base_type(start)
<br>> {
<br>> terminals = integer_ | double_;
<br>>
<br>> start = terminals>> '+'>> terminals;
<br>> }
<br>>
<br>> qi::rule<Iterator> start;
<br>> qi::rule<Iterator> terminals;
<br>> };
<br>>
<br>> The difference is this. If I put a semantic action on 'integer_', the
<br>> action gets the synthesized attribute value as a parameter. This
<br>> attribute is an integer. If I put a semantic action on
<br>> 'qi::token<ID_INTEGER>', the synthesized attribute is a string. This is
<br>> not as useful as getting an integer, as I now must find a way to convert
<br>> this string into an integer.
<br>>
<br>> Is there some way to combine Lex and Qi that still provides useful
<br>> synthesized attribute behavior of Qi? Is there something I can wrap the
<br>> "qi::token<ID_INTEGER>" parser in that will provide a properly
<br>> synthesized attribute?
</div><br><br><br> Spirit::lex is desigmed to be used with spirit::qi. a token_def<X> holds a boost::variant< boost::interator_range<SourceIteratorType>, X>. the attribute of a token_def<X> is X. The conversion from an iterator range to X occurs when the token is parsed successfully.
<br><br> Use naned token defs like
<br><br> math_tokens()
<br> {
<br> plus = '+'; // plus is a literal +.
<br> integer = ...; // integer
<br> real = ...; // real number;
<br> // now integer must be added to self before real so a real is a real that is not an integer.
<br> this->self = plus | integer | real;
<br> }
<br> lex::token_def<> plus;
<br> lex::token,def<int> integer;
<br> lex::token_def<double> real;
<br> };
<br><br> now in your qi::grammar you refer to the tokens by name they will be converted to their attribute
<br>values when qi first successes in matching the token.
<br><br> start = terminals >> tok.plus >> terminals; // EA tuple<variant,variant>
<br> terminals = tok.integer | tok.real; // EA variant<int,double>
<br><br> Does this make any sense?
<br>------------------------------------------------------------------------------
<br>This SF.Net email is sponsored by the Verizon Developer Community
<br>Take advantage of Verizon's best-in-class app development support
<br>A streamlined, 14 day to market process makes app distribution fast and easy
<br>Join now and get one step closer to millions of Verizon customers
<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_top" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a>
<br>_______________________________________________
<br>Spirit-general mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26860682&i=0" target="_top" rel="nofollow">Spirit-general@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a><br><p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26858265Lex & Qi question2009-12-19T12:27:50Z2009-12-19T12:27:50ZJason McKessonI'm trying to write a simple expression parser as the first steps
<br>towards something bigger. And I'd like to be able to use Lex to break
<br>the input up into tokens. My problem is that combining the two seems to
<br>be constraining on the utility of Qi.
<br><br>Here's what I mean. Let's say I have a lexer that looks something like this:
<br><br>template <typename Lexer>
<br>struct math_tokens : lex::lexer<Lexer>
<br>{
<br> math_tokens()
<br> {
<br> this->self.add
<br> ("\\+", ID_PLUS)
<br> ("\d+f?", ID_FLOAT)
<br> ("\d*.\d+f?", ID_FLOAT)
<br> ("\d+", ID_INTEGER)
<br> ;
<br> }
<br>};
<br><br>A parser for this would look something like this:
<br><br>template <typename Iterator>
<br>struct math_grammar : qi::grammar<Iterator>
<br>{
<br> template <typename TokenDef>
<br> math_grammar(TokenDef const& tok)
<br> : math_grammar::base_type(start)
<br> {
<br> terminals = qi::token<ID_FLOAT> | qi::token<ID_INTEGER>;
<br><br> start = terminals>> qi::token<ID_PLUS> >> terminals;
<br> }
<br><br> qi::rule<Iterator> start;
<br> qi::rule<Iterator> terminals;
<br>};
<br><br>I haven't tested this specific code, but you have the general idea of
<br>the pattern I'm interested in.
<br><br>The problem is this. If I were to build this purely without lex, the
<br>grammar would look something like this:
<br><br>template <typename Iterator>
<br>struct math_qi_grammar : qi::grammar<Iterator>
<br>{
<br> math_qi_grammar()
<br> : math_grammar::base_type(start)
<br> {
<br> terminals = integer_ | double_;
<br><br> start = terminals>> '+'>> terminals;
<br> }
<br><br> qi::rule<Iterator> start;
<br> qi::rule<Iterator> terminals;
<br>};
<br><br>The difference is this. If I put a semantic action on 'integer_', the
<br>action gets the synthesized attribute value as a parameter. This
<br>attribute is an integer. If I put a semantic action on
<br>'qi::token<ID_INTEGER>', the synthesized attribute is a string. This is
<br>not as useful as getting an integer, as I now must find a way to convert
<br>this string into an integer.
<br><br>Is there some way to combine Lex and Qi that still provides useful
<br>synthesized attribute behavior of Qi? Is there something I can wrap the
<br>"qi::token<ID_INTEGER>" parser in that will provide a properly
<br>synthesized attribute?
<br><br>Now, I do understand that there is a difference between "double_" and
<br>the particular tokens that give ID_FLOAT that I wrote in the Lex
<br>example. Namely, the optional trailing 'f' character that identifies a
<br>sequence of digits as a floating-point value. So in my above example, I
<br>would have to manually convert the Lex tokens into some kind of float
<br>(possibly with a Karma converter?) in order to get the right value. But
<br>there should be a simple way to get an integer from a token that parses
<br>integers, yes?
<br><br><br><br><br>------------------------------------------------------------------------------
<br>This SF.Net email is sponsored by the Verizon Developer Community
<br>Take advantage of Verizon's best-in-class app development support
<br>A streamlined, 14 day to market process makes app distribution fast and easy
<br>Join now and get one step closer to millions of Verizon customers
<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_top" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a>
<br>_______________________________________________
<br>Spirit-general mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26858265&i=0" target="_top" rel="nofollow">Spirit-general@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a><br><p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26852618Re: permutation with all elements required2009-12-18T20:02:46Z2009-12-18T20:02:46ZJoel de Guzman-2Carl Barron wrote:
<div class='shrinkable-quote'><br>> On Dec 18, 2009, at 8:25 PM, Joel de Guzman wrote:
<br>>
<br>>> Carl Barron wrote:
<br>>>> On Dec 18, 2009, at 7:13 PM, Joel de Guzman wrote:
<br>>>>> The attribute of the permutation operator given a rule
<br>>>>>
<br>>>>> a ^ b ^ c
<br>>>>>
<br>>>>> is
<br>>>>>
<br>>>>> vector<optional<A>, optional<B>, optional<C> >
<br>>>>>
<br>>>>> After a successful match, you just check that all optionals
<br>>>>> for A, B and C are initialized (see Boost Optional).
<br>>>>>
<br>>>> vwctor< ...> or tuple<...> ?
<br>>> Sorry, yeah, tuple. I'm just used to fusion::vector.
<br>>
<br>> Then phoenix and fusion make this fairly easy for an an arbitrary # of op ^()'s in sequence
<br>> a1 ^ a2 ^ .... ^a_n, with one phoenix function for any n.
<br>>
<br>> use fusion::find_if to test for any optional containing 'nothing' and compare the result to
<br>> fusion::end. in a phoenix function.
<br>>
<br>> Something like:
<br>>
<br>> class no_empties_impl
<br>> {
<br>> struct optional_empty
<br>> {
<br>> template <class A>
<br>> bool operator () (const A &a) const
<br>> {
<br>> // return true if a is 'empty'.
<br>> return !a;
<br>> }
<br>> };
<br>> public:
<br>> template <class A>
<br>> struct result {typedef bool type;};
<br>>
<br>> template <class A>
<br>> bool operator () (const A &a) const
<br>> {
<br>> // look for an empty optional, if any return false else true.
<br>> return boost::fusion::find_if<optional_empty>(a) == boost::fusion::end(a);
<br>> }
<br>> };
<br>>
<br>> Looks like an general implimentation of no_empties
<br>>
<br>> foo %= a1 ^... ^a_n [_pass = no_empties(qi::_0)]; // or is it _1?? the fusion seq,
<br>>
<br>> should provide result for any n. permutation of n parsers n at a time.
</div><br>Use fusion::all instead of find_if: <a href="http://tinyurl.com/yzm94gc" target="_top" rel="nofollow">http://tinyurl.com/yzm94gc</a><br><br>Regards,
<br>--
<br>Joel de Guzman
<br><a href="http://www.boostpro.com" target="_top" rel="nofollow">http://www.boostpro.com</a><br><a href="http://spirit.sf.net" target="_top" rel="nofollow">http://spirit.sf.net</a><br><a href="http://www.facebook.com/djowel" target="_top" rel="nofollow">http://www.facebook.com/djowel</a><br><br>Meet me at BoostCon
<br><a href="http://www.boostcon.com/home" target="_top" rel="nofollow">http://www.boostcon.com/home</a><br><a href="http://www.facebook.com/boostcon" target="_top" rel="nofollow">http://www.facebook.com/boostcon</a><br><br><br><br>------------------------------------------------------------------------------
<br>This SF.Net email is sponsored by the Verizon Developer Community
<br>Take advantage of Verizon's best-in-class app development support
<br>A streamlined, 14 day to market process makes app distribution fast and easy
<br>Join now and get one step closer to millions of Verizon customers
<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_top" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a>
<br>_______________________________________________
<br>Spirit-general mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26852618&i=0" target="_top" rel="nofollow">Spirit-general@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a><br><p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26852062Re: permutation with all elements required2009-12-18T17:59:03Z2009-12-18T17:59:03ZCARL BARRON-3<br>On Dec 18, 2009, at 8:25 PM, Joel de Guzman wrote:
<br><div class='shrinkable-quote'><br>> Carl Barron wrote:
<br>>> On Dec 18, 2009, at 7:13 PM, Joel de Guzman wrote:
<br>>>> The attribute of the permutation operator given a rule
<br>>>>
<br>>>> a ^ b ^ c
<br>>>>
<br>>>> is
<br>>>>
<br>>>> vector<optional<A>, optional<B>, optional<C> >
<br>>>>
<br>>>> After a successful match, you just check that all optionals
<br>>>> for A, B and C are initialized (see Boost Optional).
<br>>>>
<br>>>
<br>>> vwctor< ...> or tuple<...> ?
<br>>
<br>> Sorry, yeah, tuple. I'm just used to fusion::vector.
</div><br> Then phoenix and fusion make this fairly easy for an an arbitrary # of op ^()'s in sequence
<br> a1 ^ a2 ^ .... ^a_n, with one phoenix function for any n.
<br><br> use fusion::find_if to test for any optional containing 'nothing' and compare the result to
<br> fusion::end. in a phoenix function.
<br><br> Something like:
<br>
<br> class no_empties_impl
<br> {
<br> struct optional_empty
<br> {
<br> template <class A>
<br> bool operator () (const A &a) const
<br> {
<br> // return true if a is 'empty'.
<br> return !a;
<br> }
<br> };
<br> public:
<br> template <class A>
<br> struct result {typedef bool type;};
<br>
<br> template <class A>
<br> bool operator () (const A &a) const
<br> {
<br> // look for an empty optional, if any return false else true.
<br> return boost::fusion::find_if<optional_empty>(a) == boost::fusion::end(a);
<br> }
<br> };
<br><br>Looks like an general implimentation of no_empties
<br><br> foo %= a1 ^... ^a_n [_pass = no_empties(qi::_0)]; // or is it _1?? the fusion seq,
<br><br>should provide result for any n. permutation of n parsers n at a time.
<br>
<br>
<br><br><br>------------------------------------------------------------------------------
<br>This SF.Net email is sponsored by the Verizon Developer Community
<br>Take advantage of Verizon's best-in-class app development support
<br>A streamlined, 14 day to market process makes app distribution fast and easy
<br>Join now and get one step closer to millions of Verizon customers
<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_top" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a>
<br>_______________________________________________
<br>Spirit-general mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26852062&i=0" target="_top" rel="nofollow">Spirit-general@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a><br><p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26851897Re: permutation with all elements required2009-12-18T17:25:00Z2009-12-18T17:25:00ZJoel de Guzman-2Carl Barron wrote:
<div class='shrinkable-quote'><br>> On Dec 18, 2009, at 7:13 PM, Joel de Guzman wrote:
<br>>> The attribute of the permutation operator given a rule
<br>>>
<br>>> a ^ b ^ c
<br>>>
<br>>> is
<br>>>
<br>>> vector<optional<A>, optional<B>, optional<C> >
<br>>>
<br>>> After a successful match, you just check that all optionals
<br>>> for A, B and C are initialized (see Boost Optional).
<br>>>
<br>>
<br>> vwctor< ...> or tuple<...> ?
</div><br>Sorry, yeah, tuple. I'm just used to fusion::vector.
<br><br>Regards,
<br>--
<br>Joel de Guzman
<br><a href="http://www.boostpro.com" target="_top" rel="nofollow">http://www.boostpro.com</a><br><a href="http://spirit.sf.net" target="_top" rel="nofollow">http://spirit.sf.net</a><br><a href="http://www.facebook.com/djowel" target="_top" rel="nofollow">http://www.facebook.com/djowel</a><br><br>Meet me at BoostCon
<br><a href="http://www.boostcon.com/home" target="_top" rel="nofollow">http://www.boostcon.com/home</a><br><a href="http://www.facebook.com/boostcon" target="_top" rel="nofollow">http://www.facebook.com/boostcon</a><br><br><br><br>------------------------------------------------------------------------------
<br>This SF.Net email is sponsored by the Verizon Developer Community
<br>Take advantage of Verizon's best-in-class app development support
<br>A streamlined, 14 day to market process makes app distribution fast and easy
<br>Join now and get one step closer to millions of Verizon customers
<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_top" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a>
<br>_______________________________________________
<br>Spirit-general mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26851897&i=0" target="_top" rel="nofollow">Spirit-general@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a><br><p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26851815Re: permutation with all elements required2009-12-18T17:12:15Z2009-12-18T17:12:15ZCARL BARRON-3<br>On Dec 18, 2009, at 7:13 PM, Joel de Guzman wrote:
<br><div class='shrinkable-quote'><br>> Kenneth Riddile wrote:
<br>>> How can I express a permutation in Spirit where each element is required
<br>>> to be present? ie:
<br>>>
<br>>> a ^ b ^ c would only match:
<br>>>
<br>>> a b c
<br>>> a c b
<br>>> b a c
<br>>> b c a
<br>>> c a b
<br>>> c b a
<br>>>
<br>>> and not:
<br>>>
<br>>> a b
<br>>> b a
<br>>> c a
<br>>> ...
<br>>
<br>> The attribute of the permutation operator given a rule
<br>>
<br>> a ^ b ^ c
<br>>
<br>> is
<br>>
<br>> vector<optional<A>, optional<B>, optional<C> >
<br>>
<br>> After a successful match, you just check that all optionals
<br>> for A, B and C are initialized (see Boost Optional).
<br>>
</div><br> vwctor< ...> or tuple<...> ?
<br><br>
<br><br>------------------------------------------------------------------------------
<br>This SF.Net email is sponsored by the Verizon Developer Community
<br>Take advantage of Verizon's best-in-class app development support
<br>A streamlined, 14 day to market process makes app distribution fast and easy
<br>Join now and get one step closer to millions of Verizon customers
<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_top" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a>
<br>_______________________________________________
<br>Spirit-general mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26851815&i=0" target="_top" rel="nofollow">Spirit-general@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a><br><p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26851727Re: permutation with all elements required2009-12-18T16:13:45Z2009-12-18T16:13:45ZJoel de Guzman-2Kenneth Riddile wrote:
<div class='shrinkable-quote'><br>> How can I express a permutation in Spirit where each element is required
<br>> to be present? ie:
<br>>
<br>> a ^ b ^ c would only match:
<br>>
<br>> a b c
<br>> a c b
<br>> b a c
<br>> b c a
<br>> c a b
<br>> c b a
<br>>
<br>> and not:
<br>>
<br>> a b
<br>> b a
<br>> c a
<br>> ...
</div><br>The attribute of the permutation operator given a rule
<br><br> a ^ b ^ c
<br><br>is
<br><br> vector<optional<A>, optional<B>, optional<C> >
<br><br>After a successful match, you just check that all optionals
<br>for A, B and C are initialized (see Boost Optional).
<br><br>Regards,
<br>--
<br>Joel de Guzman
<br><a href="http://www.boostpro.com" target="_top" rel="nofollow">http://www.boostpro.com</a><br><a href="http://spirit.sf.net" target="_top" rel="nofollow">http://spirit.sf.net</a><br><a href="http://www.facebook.com/djowel" target="_top" rel="nofollow">http://www.facebook.com/djowel</a><br><br>Meet me at BoostCon
<br><a href="http://www.boostcon.com/home" target="_top" rel="nofollow">http://www.boostcon.com/home</a><br><a href="http://www.facebook.com/boostcon" target="_top" rel="nofollow">http://www.facebook.com/boostcon</a><br><br><br><br>------------------------------------------------------------------------------
<br>This SF.Net email is sponsored by the Verizon Developer Community
<br>Take advantage of Verizon's best-in-class app development support
<br>A streamlined, 14 day to market process makes app distribution fast and easy
<br>Join now and get one step closer to millions of Verizon customers
<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_top" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a>
<br>_______________________________________________
<br>Spirit-general mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26851727&i=0" target="_top" rel="nofollow">Spirit-general@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a><br><p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26851027Re: permutation with all elements required2009-12-18T15:14:09Z2009-12-18T15:14:09Zbytecolor-2Kenneth Riddile <kfriddile <at> yahoo.com> writes:
<br><br>>
<br>> How can I express a permutation in Spirit where each element is required to be
<br>present? ie:
<div class='shrinkable-quote'><br>> a ^ b ^ c would only match:
<br>>
<br>> a b c
<br>> a c b
<br>> b a c
<br>> b c a
<br>> c a b
<br>> c b a
<br>>
<br>> and not:
<br>>
<br>> a b
<br>> b a
<br>> c a
<br>> ...
<br>>
</div><br>I think you have this backwards?
<br><br>I believe a ^ b ^ c will match:
<br>a, b, c, ab, ac, bc, cb, abc, cba, etc.
<br><br>IOW, every example you gave.
<br><br>If I'm understanding, you want to match a AND b AND c in any order.
<br>So something like:
<br>foo
<br> = char_("abc")[_a = _1] >> !lit(_a) >>
<br> char_("abc")[_b = _1] >> !(lit(_a) | lit(_b)) >>
<br> char_("abc")
<br> ;
<br><br>Probably much better ways to do this, but that should work
<br>(for simple characters, any way)
<br><br><br><br><br><br><br>------------------------------------------------------------------------------
<br>This SF.Net email is sponsored by the Verizon Developer Community
<br>Take advantage of Verizon's best-in-class app development support
<br>A streamlined, 14 day to market process makes app distribution fast and easy
<br>Join now and get one step closer to millions of Verizon customers
<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_top" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a>
<br>_______________________________________________
<br>Spirit-general mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26851027&i=0" target="_top" rel="nofollow">Spirit-general@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a><br><p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26846102permutation with all elements required2009-12-18T08:37:30Z2009-12-18T08:37:30ZKenny Riddile<table cellspacing="0" cellpadding="0" border="0"><tr><td valign="top" style="font: inherit;">How can I express a permutation in Spirit where each element is required to be present? ie:<div><br></div><div>a ^ b ^ c would only match:</div><div><br></div><div>a b c</div><div>a c b</div><div>b a c</div><div>b c a</div><div>c a b</div><div>c b a</div><div><br></div><div>and not:</div><div><br></div><div>a b</div><div>b a</div><div>c a</div><div>...</div></td></tr></table><br />------------------------------------------------------------------------------
<br>This SF.Net email is sponsored by the Verizon Developer Community
<br>Take advantage of Verizon's best-in-class app development support
<br>A streamlined, 14 day to market process makes app distribution fast and easy
<br>Join now and get one step closer to millions of Verizon customers
<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_top" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a> <br />_______________________________________________
<br>Spirit-general mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26846102&i=0" target="_top" rel="nofollow">Spirit-general@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a><br><p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26825968Re: Spirit 2.1 numeric parsing2009-12-17T02:24:06Z2009-12-17T02:24:06ZJoel de Guzman-2Glenn Macgregor wrote:
<br>>
<br>> This works fine as all numerics (2 & 3.6) are handled in the double
<br>> rule. I have a need to distinguish between a double and a long (and
<br>> potentially ulong as well). I have changed my grammar to:
<br><br>strict_real_policies is your friend. See <a href="http://tinyurl.com/ybftj8v" target="_top" rel="nofollow">http://tinyurl.com/ybftj8v</a>.
<br>Example:
<br><br> using boost::spirit::qi::strict_real_policies;
<br> real_parser<double, strict_real_policies<double> > strict_double;
<br><br>If you also need to distinguish ulong, then use ulong_ and explicitly
<br>parse the sign yourself to distinguish signedness.
<br><br>Regards,
<br>--
<br>Joel de Guzman
<br><a href="http://www.boostpro.com" target="_top" rel="nofollow">http://www.boostpro.com</a><br><a href="http://spirit.sf.net" target="_top" rel="nofollow">http://spirit.sf.net</a><br><a href="http://www.facebook.com/djowel" target="_top" rel="nofollow">http://www.facebook.com/djowel</a><br><br>Meet me at BoostCon
<br><a href="http://www.boostcon.com/home" target="_top" rel="nofollow">http://www.boostcon.com/home</a><br><a href="http://www.facebook.com/boostcon" target="_top" rel="nofollow">http://www.facebook.com/boostcon</a><br><br><br><br>------------------------------------------------------------------------------
<br>This SF.Net email is sponsored by the Verizon Developer Community
<br>Take advantage of Verizon's best-in-class app development support
<br>A streamlined, 14 day to market process makes app distribution fast and easy
<br>Join now and get one step closer to millions of Verizon customers
<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_top" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a>
<br>_______________________________________________
<br>Spirit-general mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26825968&i=0" target="_top" rel="nofollow">Spirit-general@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a><br><p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26825525Spirit 2.1 numeric parsing2009-12-17T01:43:44Z2009-12-17T01:43:44ZGlenn Macgregor-3<html xmlns:v="urn:schemas-microsoft-com:vml" xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:w="urn:schemas-microsoft-com:office:word" xmlns:x="urn:schemas-microsoft-com:office:excel" xmlns:p="urn:schemas-microsoft-com:office:powerpoint" xmlns:a="urn:schemas-microsoft-com:office:access" xmlns:dt="uuid:C2F41010-65B3-11d1-A29F-00AA00C14882" xmlns:s="uuid:BDC6E3F0-6DA3-11d1-A2A3-00AA00C14882" xmlns:rs="urn:schemas-microsoft-com:rowset" xmlns:Z="urn:schemas-microsoft-com:" xmlns:b="urn:schemas-microsoft-com:office:publisher" xmlns:ss="urn:schemas-microsoft-com:office:spreadsheet" xmlns:c="urn:schemas-microsoft-com:office:component:spreadsheet" xmlns:odc="urn:schemas-microsoft-com:office:odc" xmlns:oa="urn:schemas-microsoft-com:office:activation" xmlns:html="http://www.w3.org/TR/REC-html40" xmlns:q="http://schemas.xmlsoap.org/soap/envelope/" xmlns:rtc="http://microsoft.com/officenet/conferencing" xmlns:D="DAV:" xmlns:Repl="http://schemas.microsoft.com/repl/" xmlns:mt="http://schemas.microsoft.com/sharepoint/soap/meetings/" xmlns:x2="http://schemas.microsoft.com/office/excel/2003/xml" xmlns:ppda="http://www.passport.com/NameSpace.xsd" xmlns:ois="http://schemas.microsoft.com/sharepoint/soap/ois/" xmlns:dir="http://schemas.microsoft.com/sharepoint/soap/directory/" xmlns:ds="http://www.w3.org/2000/09/xmldsig#" xmlns:dsp="http://schemas.microsoft.com/sharepoint/dsp" xmlns:udc="http://schemas.microsoft.com/data/udc" xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:sub="http://schemas.microsoft.com/sharepoint/soap/2002/1/alerts/" xmlns:ec="http://www.w3.org/2001/04/xmlenc#" xmlns:sp="http://schemas.microsoft.com/sharepoint/" xmlns:sps="http://schemas.microsoft.com/sharepoint/soap/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:udcs="http://schemas.microsoft.com/data/udc/soap" xmlns:udcxf="http://schemas.microsoft.com/data/udc/xmlfile" xmlns:udcp2p="http://schemas.microsoft.com/data/udc/parttopart" xmlns:wf="http://schemas.microsoft.com/sharepoint/soap/workflow/" xmlns:dsss="http://schemas.microsoft.com/office/2006/digsig-setup" xmlns:dssi="http://schemas.microsoft.com/office/2006/digsig" xmlns:mdssi="http://schemas.openxmlformats.org/package/2006/digital-signature" xmlns:mver="http://schemas.openxmlformats.org/markup-compatibility/2006" xmlns:m="http://schemas.microsoft.com/office/2004/12/omml" xmlns:mrels="http://schemas.openxmlformats.org/package/2006/relationships" xmlns:spwp="http://microsoft.com/sharepoint/webpartpages" xmlns:ex12t="http://schemas.microsoft.com/exchange/services/2006/types" xmlns:ex12m="http://schemas.microsoft.com/exchange/services/2006/messages" xmlns:pptsl="http://schemas.microsoft.com/sharepoint/soap/SlideLibrary/" xmlns:spsl="http://microsoft.com/webservices/SharePointPortalServer/PublishedLinksService" xmlns:st="" xmlns="http://www.w3.org/TR/REC-html40">
<head>
<meta http-equiv=Content-Type content="text/html; charset=us-ascii">
<meta name=Generator content="Microsoft Word 12 (filtered medium)">
<!--[if gte mso 9]><xml>
<o:shapedefaults v:ext="edit" spidmax="1026" />
</xml><![endif]--><!--[if gte mso 9]><xml>
<o:shapelayout v:ext="edit">
<o:idmap v:ext="edit" data="1" />
</o:shapelayout></xml><![endif]-->
</head>
<body lang=EN-IE link=blue vlink=purple>
<div class=Section1>
<p class=MsoNormal>Hi All,<o:p></o:p></p>
<p class=MsoNormal><o:p> </o:p></p>
<p class=MsoNormal>Thanks for all the help thus far, it has been invaluable. I
now want to distinguish between ints, uint and floats in the parser, is that
possible?<o:p></o:p></p>
<p class=MsoNormal><o:p> </o:p></p>
<p class=MsoNormal>I currently have:<o:p></o:p></p>
<p class=MsoNormal><o:p> </o:p></p>
<p class=MsoNormal style='text-indent:36.0pt;text-autospace:none'><span style='font-size:10.0pt;font-family:"Courier New"'>mIdent<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'>
= raw[lexeme[alpha >> *(alnum | <span style='color:#A31515'>"_"</span>)]]<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'>
;<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'><o:p> </o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> mPrimary<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'>
= mIdent[_val = _1]<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'>
| double_[_val = _1]<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'>
| <span style='color:#A31515'>'('</span> >> mOrExpr[_val = _1] >> <span style='color:#A31515'>')'<o:p></o:p></span></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'>
| (<span style='color:#A31515'>'!'</span> >> mPrimary)[_val =
notOperator(_1)]<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'>
| (<span style='color:#A31515'>'-'</span> >> mPrimary)[_val =
negateOperator(_1)]<o:p></o:p></span></p>
<p class=MsoNormal><span style='font-size:10.0pt;font-family:"Courier New"'>
;</span><o:p></o:p></p>
<p class=MsoNormal><o:p> </o:p></p>
<p class=MsoNormal>….<o:p></o:p></p>
<p class=MsoNormal><o:p> </o:p></p>
<p class=MsoNormal>This works fine as all numerics (2 & 3.6) are handled in
the double rule. I have a need to distinguish between a double and a long (and
potentially ulong as well). I have changed my grammar to:<o:p></o:p></p>
<p class=MsoNormal><o:p> </o:p></p>
<p class=MsoNormal style='text-indent:36.0pt;text-autospace:none'><span style='font-size:10.0pt;font-family:"Courier New"'>mPrimary<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'>
= mIdent[_val = _1]<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'>
| double_[_val = _1]<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> |
int_[_val = _1]<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'>
| <span style='color:#A31515'>'('</span> >> mOrExpr[_val = _1] >> <span style='color:#A31515'>')'<o:p></o:p></span></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'>
| (<span style='color:#A31515'>'!'</span> >> mPrimary)[_val =
notOperator(_1)]<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'>
| (<span style='color:#A31515'>'-'</span> >> mPrimary)[_val =
negateOperator(_1)]<o:p></o:p></span></p>
<p class=MsoNormal><span style='font-size:10.0pt;font-family:"Courier New"'>
;<o:p></o:p></span></p>
<p class=MsoNormal><span style='font-size:10.0pt;font-family:"Courier New"'><o:p> </o:p></span></p>
<p class=MsoNormal><span style='font-size:10.0pt;font-family:"Courier New"'>Which
still read everything in the double rule. So I moved to:<o:p></o:p></span></p>
<p class=MsoNormal><o:p> </o:p></p>
<p class=MsoNormal style='text-indent:36.0pt;text-autospace:none'><span style='font-size:10.0pt;font-family:"Courier New"'>mPrimary<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'>
= mIdent[_val = _1]<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> |
int_[_val = _1]<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'>
| double_[_val = _1]<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'>
| <span style='color:#A31515'>'('</span> >> mOrExpr[_val = _1] >> <span style='color:#A31515'>')'<o:p></o:p></span></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'>
| (<span style='color:#A31515'>'!'</span> >> mPrimary)[_val =
notOperator(_1)]<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'>
| (<span style='color:#A31515'>'-'</span> >> mPrimary)[_val =
negateOperator(_1)]<o:p></o:p></span></p>
<p class=MsoNormal><span style='font-size:10.0pt;font-family:"Courier New"'>
;<o:p></o:p></span></p>
<p class=MsoNormal><span style='font-size:10.0pt;font-family:"Courier New"'><o:p> </o:p></span></p>
<p class=MsoNormal><span style='font-size:10.0pt;font-family:"Courier New"'>This
now read all numerics in the int rule so I lost all decimal places on reals. I
have read that I can use the longest_d but it seems to have been removed in the
Spirit 2.1. What, if any, is the prefered way to distinguish between double,
long and unsigned long values in a grammar?<o:p></o:p></span></p>
<p class=MsoNormal><span style='font-size:10.0pt;font-family:"Courier New"'><o:p> </o:p></span></p>
<p class=MsoNormal><span style='font-size:10.0pt;font-family:"Courier New"'>Thanks<o:p></o:p></span></p>
<p class=MsoNormal><span style='font-size:10.0pt;font-family:"Courier New"'><o:p> </o:p></span></p>
<p class=MsoNormal><span style='font-size:10.0pt;font-family:"Courier New"'> Glenn<o:p></o:p></span></p>
<p class=MsoNormal><o:p> </o:p></p>
<p class=MsoNormal><o:p> </o:p></p>
</div>
</body>
</html>
<br />------------------------------------------------------------------------------
<br>This SF.Net email is sponsored by the Verizon Developer Community
<br>Take advantage of Verizon's best-in-class app development support
<br>A streamlined, 14 day to market process makes app distribution fast and easy
<br>Join now and get one step closer to millions of Verizon customers
<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_top" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a> <br />_______________________________________________
<br>Spirit-general mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26825525&i=0" target="_top" rel="nofollow">Spirit-general@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a><br><p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26824698Re: phrase_parse() or parse()?2009-12-17T00:24:35Z2009-12-17T00:24:35ZOvermindDL1On Thu, Dec 17, 2009 at 12:44 AM, Tan, Tom (Shanghai) <<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26824698&i=0" target="_top" rel="nofollow">TTan@...</a>> wrote:
<div class='shrinkable-quote'><br>> Hi, this might be a naive question, but I have a question that cannot be
<br>> answered even after reading the doc:
<br>>
<br>> Except for the apparent different signatures,
<br>>
<br>> - what are the differences of parse() and phrase_parse() in terms
<br>> of implementation, performance etc ?
<br>>
<br>> - When to use which?
<br>>
<br>> - Is one a shortcut to a special case of another? And if so, which
<br>> one?
<br>>
<br>>
<br>>
<br>> I am especially puzzled by term “phrase level” compared to “character level”
<br>> as appearing here: “The first form parse works on the character level. The
<br>> second phrase_parse works on the phrase level and requires skip parser “ ,
<br>> isn’t comparing starting from “the character level”? How is phrase defined
<br>> here?
</div><br><br>phrase -> for parsing a grammar with no explicit skipper (but it may
<br>still have a skipper internally by using the skip[] directive).
<br>parse_phrase -> for parsing a grammar with an explicit skipper
<br><br>Example:
<br>rule<Iterator, blank_type> rule_skip;
<br>rule<Iterator> rule_noskip1;
<br><br>These work:
<br>parse_phrase(b,e, rule_skip, blank);
<br>parse(b,e, rule_noskip)
<br>parse_phrase(b,e, rule_noskip, unused);
<br>parse(b,e, skip(blank)[rule_skip]);
<br>parse_phrase(b,e, skip(blank)[rule_skip], unused);
<br><br>(I hope I got the syntax right, I am tired...)
<br><br>These do not work (fail to compile):
<br>parse_phrase(b,e, rule_noskip, blank);
<br>parse(b,e, rule_skip)
<br><br>------------------------------------------------------------------------------
<br>This SF.Net email is sponsored by the Verizon Developer Community
<br>Take advantage of Verizon's best-in-class app development support
<br>A streamlined, 14 day to market process makes app distribution fast and easy
<br>Join now and get one step closer to millions of Verizon customers
<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_top" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a>
<br>_______________________________________________
<br>Spirit-general mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26824698&i=1" target="_top" rel="nofollow">Spirit-general@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a><br><p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26824597phrase_parse() or parse()?2009-12-16T23:44:08Z2009-12-16T23:44:08ZTan, Tom (Shanghai)<html xmlns:v="urn:schemas-microsoft-com:vml" xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:w="urn:schemas-microsoft-com:office:word" xmlns:x="urn:schemas-microsoft-com:office:excel" xmlns:p="urn:schemas-microsoft-com:office:powerpoint" xmlns:a="urn:schemas-microsoft-com:office:access" xmlns:dt="uuid:C2F41010-65B3-11d1-A29F-00AA00C14882" xmlns:s="uuid:BDC6E3F0-6DA3-11d1-A2A3-00AA00C14882" xmlns:rs="urn:schemas-microsoft-com:rowset" xmlns:Z="urn:schemas-microsoft-com:" xmlns:b="urn:schemas-microsoft-com:office:publisher" xmlns:ss="urn:schemas-microsoft-com:office:spreadsheet" xmlns:c="urn:schemas-microsoft-com:office:component:spreadsheet" xmlns:odc="urn:schemas-microsoft-com:office:odc" xmlns:oa="urn:schemas-microsoft-com:office:activation" xmlns:html="http://www.w3.org/TR/REC-html40" xmlns:q="http://schemas.xmlsoap.org/soap/envelope/" xmlns:rtc="http://microsoft.com/officenet/conferencing" xmlns:D="DAV:" xmlns:Repl="http://schemas.microsoft.com/repl/" xmlns:mt="http://schemas.microsoft.com/sharepoint/soap/meetings/" xmlns:x2="http://schemas.microsoft.com/office/excel/2003/xml" xmlns:ppda="http://www.passport.com/NameSpace.xsd" xmlns:ois="http://schemas.microsoft.com/sharepoint/soap/ois/" xmlns:dir="http://schemas.microsoft.com/sharepoint/soap/directory/" xmlns:ds="http://www.w3.org/2000/09/xmldsig#" xmlns:dsp="http://schemas.microsoft.com/sharepoint/dsp" xmlns:udc="http://schemas.microsoft.com/data/udc" xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:sub="http://schemas.microsoft.com/sharepoint/soap/2002/1/alerts/" xmlns:ec="http://www.w3.org/2001/04/xmlenc#" xmlns:sp="http://schemas.microsoft.com/sharepoint/" xmlns:sps="http://schemas.microsoft.com/sharepoint/soap/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:udcs="http://schemas.microsoft.com/data/udc/soap" xmlns:udcxf="http://schemas.microsoft.com/data/udc/xmlfile" xmlns:udcp2p="http://schemas.microsoft.com/data/udc/parttopart" xmlns:wf="http://schemas.microsoft.com/sharepoint/soap/workflow/" xmlns:dsss="http://schemas.microsoft.com/office/2006/digsig-setup" xmlns:dssi="http://schemas.microsoft.com/office/2006/digsig" xmlns:mdssi="http://schemas.openxmlformats.org/package/2006/digital-signature" xmlns:mver="http://schemas.openxmlformats.org/markup-compatibility/2006" xmlns:m="http://schemas.microsoft.com/office/2004/12/omml" xmlns:mrels="http://schemas.openxmlformats.org/package/2006/relationships" xmlns:spwp="http://microsoft.com/sharepoint/webpartpages" xmlns:ex12t="http://schemas.microsoft.com/exchange/services/2006/types" xmlns:ex12m="http://schemas.microsoft.com/exchange/services/2006/messages" xmlns:pptsl="http://schemas.microsoft.com/sharepoint/soap/SlideLibrary/" xmlns:spsl="http://microsoft.com/webservices/SharePointPortalServer/PublishedLinksService" xmlns:st="" xmlns="http://www.w3.org/TR/REC-html40">
<head>
<meta http-equiv=Content-Type content="text/html; charset=us-ascii">
<meta name=Generator content="Microsoft Word 12 (filtered medium)">
<!--[if gte mso 9]><xml>
<o:shapedefaults v:ext="edit" spidmax="1026" />
</xml><![endif]--><!--[if gte mso 9]><xml>
<o:shapelayout v:ext="edit">
<o:idmap v:ext="edit" data="1" />
</o:shapelayout></xml><![endif]-->
</head>
<body lang=EN-US link=blue vlink=purple>
<div class=Section1>
<p class=MsoNormal>Hi, this might be a naive question, but I have a question
that cannot be answered even after reading the doc:<o:p></o:p></p>
<p class=MsoNormal>Except for the apparent different signatures, <o:p></o:p></p>
<p class=MsoListParagraph style='text-indent:-18.0pt;mso-list:l0 level1 lfo1'><![if !supportLists]><span style='mso-list:Ignore'>-<span style='font:7.0pt "Times New Roman"'>
</span></span><![endif]>what are the differences of parse() and phrase_parse()
in terms of implementation, performance etc ? <o:p></o:p></p>
<p class=MsoListParagraph style='text-indent:-18.0pt;mso-list:l0 level1 lfo1'><![if !supportLists]><span style='mso-list:Ignore'>-<span style='font:7.0pt "Times New Roman"'>
</span></span><![endif]>When to use which? <o:p></o:p></p>
<p class=MsoListParagraph style='text-indent:-18.0pt;mso-list:l0 level1 lfo1'><![if !supportLists]><span style='mso-list:Ignore'>-<span style='font:7.0pt "Times New Roman"'>
</span></span><![endif]>Is one a shortcut to a special case of another? And if
so, which one?<o:p></o:p></p>
<p class=MsoListParagraph style='margin-left:0cm'><o:p> </o:p></p>
<p class=MsoListParagraph style='margin-left:0cm'>I am especially puzzled by
term “phrase level” compared to “character level” as
appearing here: “<span class=apple-style-span><span style='font-size:
7.0pt;font-family:"Arial","sans-serif";color:black'>The first form</span></span><span class=apple-converted-space><span style='font-size:7.0pt;font-family:"Arial","sans-serif";
color:black'> </span></span><span class=identifier><span style='font-size:
10.0pt;font-family:"Courier New";color:black'>parse</span></span><span class=apple-converted-space><span style='font-size:7.0pt;font-family:"Arial","sans-serif";
color:black'> </span></span><span class=apple-style-span><span style='font-size:7.0pt;font-family:"Arial","sans-serif";color:black'>works on
the character level. The second</span></span><span class=apple-converted-space><span style='font-size:7.0pt;font-family:"Arial","sans-serif";color:black'> </span></span><span class=identifier><span style='font-size:10.0pt;font-family:"Courier New";
color:black'>phrase_parse</span></span><span class=apple-converted-space><span style='font-size:7.0pt;font-family:"Arial","sans-serif";color:black'> </span></span><span class=apple-style-span><span style='font-size:7.0pt;font-family:"Arial","sans-serif";
color:black'>works on the phrase level and requires skip parser</span></span> “
, isn’t comparing starting from “<span class=apple-style-span><span style='font-size:7.0pt;font-family:"Arial","sans-serif";color:black'>the
character level</span></span>”? How is phrase defined here?<o:p></o:p></p>
<p class=MsoListParagraph style='margin-left:0cm'><o:p> </o:p></p>
<p class=MsoListParagraph style='margin-left:0cm'>Thanks,<o:p></o:p></p>
<p class=MsoListParagraph style='margin-left:0cm'>-Tom<o:p></o:p></p>
</div>
</body>
</html>
<br />------------------------------------------------------------------------------
<br>This SF.Net email is sponsored by the Verizon Developer Community
<br>Take advantage of Verizon's best-in-class app development support
<br>A streamlined, 14 day to market process makes app distribution fast and easy
<br>Join now and get one step closer to millions of Verizon customers
<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_top" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a> <br />_______________________________________________
<br>Spirit-general mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26824597&i=0" target="_top" rel="nofollow">Spirit-general@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a><br><p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26824001Re: match |foobar| not |foo bar |2009-12-16T22:53:29Z2009-12-16T22:53:29ZMichael Caisse-2Carl Barron wrote:
<div class='shrinkable-quote'><br>> On Dec 16, 2009, at 11:22 PM, OvermindDL1 wrote:
<br>>
<br>>
<br>>> On Wed, Dec 16, 2009 at 6:42 PM, Carl Barron <<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26824001&i=0" target="_top" rel="nofollow">cbarron413@...</a>> wrote:
<br>>>
<br>>>> result is the 'short answere' and no explicit actions.
<br>>>>
<br>>> Well he is wanting to convert spaces to underscores, so he would need
<br>>> actions in that case, or make his own terminal. I just do not know if
<br>>> he wants one underscore per space, or if he wants one underscore per
<br>>> set of spaces. His code signifies the latter.
<br>>>
<br>> Don't need actions or more primitives. to do the latter. [or the former with a minor change]
<br>>
</div><br>That was clever Carl.
<br><br>--
<br><br>----------------------------------
<br>Michael Caisse
<br>Object Modeling Designs
<br>www.objectmodelingdesigns.com
<br><br><br><br>------------------------------------------------------------------------------
<br>This SF.Net email is sponsored by the Verizon Developer Community
<br>Take advantage of Verizon's best-in-class app development support
<br>A streamlined, 14 day to market process makes app distribution fast and easy
<br>Join now and get one step closer to millions of Verizon customers
<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_top" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a>
<br>_______________________________________________
<br>Spirit-general mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26824001&i=1" target="_top" rel="nofollow">Spirit-general@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a><br><p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26823562Re: match |foobar| not |foo bar |2009-12-16T21:47:34Z2009-12-16T21:47:34Zbytecolor-2OvermindDL1 <overminddl1 <at> gmail.com> writes:
<br><br>> Well he is wanting to convert spaces to underscores, so he would need
<br>> actions in that case, or make his own terminal. I just do not know if
<br>> he wants one underscore per space, or if he wants one underscore per
<br>> set of spaces. His code signifies the latter.
<br><br>I did actually want to squeeze the whitespace. Sorry if I wasn't clear
<br>on that. I was more concerned with the trailing ws in the original
<br>post.
<br><br>lexeme[&alpha >> (+alnum[_val += _1] %
<br> (+blank >> &alnum)[_val += '_'])]
<br>|a| -> a
<br>|a b | -> a_b
<br>| a b | -> a_b
<br>| a b 9| -> a_b_9
<br><br><br><br>------------------------------------------------------------------------------
<br>This SF.Net email is sponsored by the Verizon Developer Community
<br>Take advantage of Verizon's best-in-class app development support
<br>A streamlined, 14 day to market process makes app distribution fast and easy
<br>Join now and get one step closer to millions of Verizon customers
<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_top" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a>
<br>_______________________________________________
<br>Spirit-general mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26823562&i=0" target="_top" rel="nofollow">Spirit-general@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a><br><p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26823349Re: match |foobar| not |foo bar |2009-12-16T21:10:40Z2009-12-16T21:10:40ZOvermindDL1On Wed, Dec 16, 2009 at 9:51 PM, Carl Barron <<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26823349&i=0" target="_top" rel="nofollow">cbarron413@...</a>> wrote:
<div class='shrinkable-quote'><br>>
<br>> On Dec 16, 2009, at 11:22 PM, OvermindDL1 wrote:
<br>>
<br>>> On Wed, Dec 16, 2009 at 6:42 PM, Carl Barron <<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26823349&i=1" target="_top" rel="nofollow">cbarron413@...</a>> wrote:
<br>>>> result is the 'short answere' and no explicit actions.
<br>>>
<br>>> Well he is wanting to convert spaces to underscores, so he would need
<br>>> actions in that case, or make his own terminal. I just do not know if
<br>>> he wants one underscore per space, or if he wants one underscore per
<br>>> set of spaces. His code signifies the latter.
<br>> Don't need actions or more primitives. to do the latter. [or the former with a minor change]
</div><br>Yeah, quite true...
<br>To replace any number of spaces by an underscore:
<br><br>// No need for raw or lexeme with this rule type
<br>// This will replace any number of blanks with a single _
<br>rule<Iterator, std::string()> r = alpha >> *(
<br>((omit[+blank]>>attr('_'))|eps) >> alnum));
<br><br>Probably not the most pretty it could be, and you do not need all of
<br>those parenthesis (but I am parenthesis-happy), but it should work.
<br><br>------------------------------------------------------------------------------
<br>This SF.Net email is sponsored by the Verizon Developer Community
<br>Take advantage of Verizon's best-in-class app development support
<br>A streamlined, 14 day to market process makes app distribution fast and easy
<br>Join now and get one step closer to millions of Verizon customers
<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_top" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a>
<br>_______________________________________________
<br>Spirit-general mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26823349&i=2" target="_top" rel="nofollow">Spirit-general@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a><br><p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26823174Re: match |foobar| not |foo bar |2009-12-16T20:51:56Z2009-12-16T20:51:56ZCARL BARRON-3<br>On Dec 16, 2009, at 11:22 PM, OvermindDL1 wrote:
<br><br>> On Wed, Dec 16, 2009 at 6:42 PM, Carl Barron <<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26823174&i=0" target="_top" rel="nofollow">cbarron413@...</a>> wrote:
<br>>> result is the 'short answere' and no explicit actions.
<br>>
<br>> Well he is wanting to convert spaces to underscores, so he would need
<br>> actions in that case, or make his own terminal. I just do not know if
<br>> he wants one underscore per space, or if he wants one underscore per
<br>> set of spaces. His code signifies the latter.
<br> Don't need actions or more primitives. to do the latter. [or the former with a minor change]
<br><br /> <br />------------------------------------------------------------------------------
<br>This SF.Net email is sponsored by the Verizon Developer Community
<br>Take advantage of Verizon's best-in-class app development support
<br>A streamlined, 14 day to market process makes app distribution fast and easy
<br>Join now and get one step closer to millions of Verizon customers
<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_top" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a> <br />_______________________________________________
<br>Spirit-general mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26823174&i=1" target="_top" rel="nofollow">Spirit-general@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a><br><div class="small"><br/><img src="http://old.nabble.com/images/icon_attachment.gif" > <strong>main.cpp</strong> (776 bytes) <a href="http://old.nabble.com/attachment/26823174/0/main.cpp" target="_top">Download Attachment</a></div><p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26822945Re: match |foobar| not |foo bar |2009-12-16T20:22:50Z2009-12-16T20:22:50ZOvermindDL1On Wed, Dec 16, 2009 at 6:42 PM, Carl Barron <<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26822945&i=0" target="_top" rel="nofollow">cbarron413@...</a>> wrote:
<br>> result is the 'short answere' and no explicit actions.
<br><br>Well he is wanting to convert spaces to underscores, so he would need
<br>actions in that case, or make his own terminal. I just do not know if
<br>he wants one underscore per space, or if he wants one underscore per
<br>set of spaces. His code signifies the latter.
<br><br>------------------------------------------------------------------------------
<br>This SF.Net email is sponsored by the Verizon Developer Community
<br>Take advantage of Verizon's best-in-class app development support
<br>A streamlined, 14 day to market process makes app distribution fast and easy
<br>Join now and get one step closer to millions of Verizon customers
<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_top" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a>
<br>_______________________________________________
<br>Spirit-general mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26822945&i=1" target="_top" rel="nofollow">Spirit-general@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a><br><p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26822119Re: match |foobar| not |foo bar |2009-12-16T17:44:29Z2009-12-16T17:44:29Zbytecolor-2Daniel James <daniel_james <at> fmail.co.uk> writes:
<br><br>> Sorry, it's actually due to backtracking, it calls the action after
<br>> parsing the trailing blanks, but before realising that they're not
<br>> followed by an 'alnum' and backtracking. So the solution is to check
<br>> before the action:
<br>>
<br>> lexeme[&alpha
<br>> >> (+alnum[_val += _1] % (+blank >> &alnum)[_val += "_"])]
<br>>
<br><br>I think we have a winner. I believe the word `elegant' is appropriate
<br>here. That's a very cool bit of code.
<br><br>> No, that's wrong (see above). Even if it was right, you can't call
<br>> members of phoenix placeholders like that (since they can't overload
<br>> '.'), you have to either use bind, or by using '->*'.
<br><br>I need to read up on actors.
<br><br>I think Spirit is similar to my brain in that I'm only utilizing a
<br>small percentage of its potential.
<br><br>Thanks for the lesson Daniel.
<br><br>You know I was thinking (I do that a lot) there must be literally
<br>hundreds of unique parsers like this on the list. If they could
<br>somehow be categorized and stored... Actually finding what you are
<br>looking for in the repo might prove to be a challenge though.
<br>Maybe tags.
<br><br><br><br><br><br><br>------------------------------------------------------------------------------
<br>This SF.Net email is sponsored by the Verizon Developer Community
<br>Take advantage of Verizon's best-in-class app development support
<br>A streamlined, 14 day to market process makes app distribution fast and easy
<br>Join now and get one step closer to millions of Verizon customers
<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_top" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a>
<br>_______________________________________________
<br>Spirit-general mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26822119&i=0" target="_top" rel="nofollow">Spirit-general@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a><br><p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26822849Re: match |foobar| not |foo bar |2009-12-16T17:42:33Z2009-12-16T17:42:33ZCARL BARRON-3<br>On Dec 16, 2009, at 8:23 AM, bytecolor wrote:
<br><div class='shrinkable-quote'><br>> I'd like to parse:
<br>> an identifier with spaces in it
<br>>
<br>> So I use:
<br>> lexeme[alpha >> *(alpha | blank)]
<br>>
<br>> but it matches
<br>> |an identifier with spaces in it | <- trailing whitespace
<br>>
<br>> instead of
<br>> |an identifier with spaces in it| <- no trailing whitespace
<br>>
<br>> I'm already incorporating semantic actions to convert the blanks to
<br>> underscores, so it would be nothing to chop the trailing whitespace
<br>> at the same time. But I was wondering if there was a way to do this.
<br>>
<br>> In a Python regular expression, it would be:
<br>> [A-Za-z](?:[A-Za-z0-9 \v\t]*[A-Za-z0-9])?
<br>>
<br>> Which I thought would translate to:
<br>> alpha >> -(*(alnum | blank) >> alnum)
<br>>
<br>> With [_val += _1] all around it compiles but does not match. I think
<br>> *(alnum | blank) is greedy, not leaving anything for the last alnum,
<br>> unlike the Python regex. Even though I'm using the greedy op there
<br>> instead of *?.
<br>>
<br>> Just a curiosity, really.
</div><br> Short answer var = raw[lexeme[ alpha >> *(*blank >> alnum) ]];
<br><br> a ( b | c)* b in EBNF [qi notation but not desired a >> *( b|c ) >> b]
<br>The problem is qi would match all the b's inside the op *() so we need to be more specific. We want
<br>what ever is matched in op *() to end with a b. *c >> b does the trick.
<br> var = alpha >> *( *blank >> alnum));
<br><br> lexeme to not apply the skipper, and raw to get it as an iterator_range, which is convertable to
<br>std::string.
<br><br>result is the 'short answere' and no explicit actions.
<br><br><br>------------------------------------------------------------------------------
<br>This SF.Net email is sponsored by the Verizon Developer Community
<br>Take advantage of Verizon's best-in-class app development support
<br>A streamlined, 14 day to market process makes app distribution fast and easy
<br>Join now and get one step closer to millions of Verizon customers
<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_top" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a>
<br>_______________________________________________
<br>Spirit-general mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26822849&i=0" target="_top" rel="nofollow">Spirit-general@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a><br><p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26821420Re: match |foobar| not |foo bar |2009-12-16T17:00:30Z2009-12-16T17:00:30ZDaniel James2009/12/16 bytecolor <<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26821420&i=0" target="_top" rel="nofollow">bytecolor@...</a>>:
<br>> Daniel James <daniel_james <at> fmail.co.uk> writes:
<br>>
<br>>> lexeme[α
<br><br>Funny, something thinks that '&alpha' is the html character entity 'α'.
<br><br>>> lexeme[&alpha
<br>>> >> (+alnum[_val += _1] % (+blank)[_val += "_"])]
<br>>
<br>> I had been looking at the peek operator but couldn't quite figure out
<br>> how to incorporate it. Make sense now. I actually had *blank after
<br>> the initial alpha. I tried a few dozen variants but a trailing '_'
<br>> still gets appended.
<br><br>Sorry, it's actually due to backtracking, it calls the action after
<br>parsing the trailing blanks, but before realising that they're not
<br>followed by an 'alnum' and backtracking. So the solution is to check
<br>before the action:
<br><br> lexeme[&alpha
<br> >> (+alnum[_val += _1] % (+blank >> &alnum)[_val += "_"])]
<br><br>This kind of problem turns up when putting actions inside expressions,
<br>it doesn't happen with synthesized attributes.
<br><br>> I *think* (+blank) is triggering the action even if it does not match
<br>> on the final pass. I tried using a few variations of:> (+blank)[if_(_1.size())[_val += '_']]
<br>>
<br>> but I get 'no size() member error'.
<br><br>No, that's wrong (see above). Even if it was right, you can't call
<br>members of phoenix placeholders like that (since they can't overload
<br>'.'), you have to either use bind, or by using '->*'.
<br><br>> I can't tell what (+blank) is
<br>> synthesizing. I though it would be a vector<char>, but from the
<br>> error diag it looks like the parens are creating a
<br>> vector1<vector<char> >. Not sure.
<br><br>It's 'std::vector<char>'. What you're seeing is the way spirit packs
<br>the attributes when it calls the actions. The phoenix placeholders
<br>that spirit provides unpacks these attributes. If you don't use those
<br>placeholders you have to do it yourself manually (not advisable).
<br><br>Daniel
<br><br>------------------------------------------------------------------------------
<br>This SF.Net email is sponsored by the Verizon Developer Community
<br>Take advantage of Verizon's best-in-class app development support
<br>A streamlined, 14 day to market process makes app distribution fast and easy
<br>Join now and get one step closer to millions of Verizon customers
<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_top" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a>
<br>_______________________________________________
<br>Spirit-general mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26821420&i=1" target="_top" rel="nofollow">Spirit-general@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a><br><p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26820143Re: match |foobar| not |foo bar |2009-12-16T15:05:25Z2009-12-16T15:05:25Zbytecolor-2Daniel James <daniel_james <at> fmail.co.uk> writes:
<br><br><div class='shrinkable-quote'><br>> Do you want it to match 'a b'? How about:
<br>>
<br>> lexeme[α
<br>> >> (+alnum[_val += _1] % blank[_val += "_"])]
<br>>
<br>> The 'α' matches a character but doesn't consume it. And if you
<br>> want to allow more than one consecutive whitespace character:
<br>>
<br>> lexeme[α
<br>> >> (+alnum[_val += _1] % (+blank)[_val += "_"])]
<br>>
<br>> Daniel
</div><br>I had been looking at the peek operator but couldn't quite figure out
<br>how to incorporate it. Make sense now. I actually had *blank after
<br>the initial alpha. I tried a few dozen variants but a trailing '_'
<br>still gets appended.
<br><br>I *think* (+blank) is triggering the action even if it does not match
<br>on the final pass. I tried using a few variations of:
<br><br>(+blank)[if_(_1.size())[_val += '_']]
<br><br>but I get 'no size() member error'. I can't tell what (+blank) is
<br>synthesizing. I though it would be a vector<char>, but from the
<br>error diag it looks like the parens are creating a
<br>vector1<vector<char> >. Not sure.
<br><br><br>------------------------------------------------------------------------------
<br>This SF.Net email is sponsored by the Verizon Developer Community
<br>Take advantage of Verizon's best-in-class app development support
<br>A streamlined, 14 day to market process makes app distribution fast and easy
<br>Join now and get one step closer to millions of Verizon customers
<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_top" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a>
<br>_______________________________________________
<br>Spirit-general mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26820143&i=0" target="_top" rel="nofollow">Spirit-general@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a><br><p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26819852Re: hanging grammar??2009-12-16T14:46:09Z2009-12-16T14:46:09ZdWiGhT-TXOk... figured out that the "eoi" was the issue...
<br><br>changing these lines makes it work as expected:
<br> lines2Convert %= ((oneLine | twoLine | bypassLine)
<br> >> ( eol[_val += '\n'] );
<br> bypassLine %= *(char_ - eol);
<br><br><br>... sometimes you just have to search before posting... "; )
<br><br>dWiGhT
<br><br><blockquote class="quote light-black dark-border-color"><div class="quote light-border-color">
<div class="quote-author" style="font-weight: bold;">dWiGhT-TX wrote:</div>
<div class="quote-message shrinkable-quote">First a disclaimer: I'm a Spirit virgin and just starting to play around with
<br>v2.1... so please keep that in mind. (thanks)
<br><br>Ok... playing around with Spirit v2.1 and put this together to take a string as
<br>input and output the converted string as a little test.
<br><br>Here is the part of the grammar:
<br>...
<br> convertLines = eps[_val = ""] >> *lines2Convert[_val += _1];
<br> lines2Convert %= ((oneLine | twoLine ) >> (eol[_val += '\n'] | eoi));
<br> oneLine = lit("one")[_val = "ONE"];
<br> twoLine = lit("two")[_val = "TWO"];
<br>...
<br><br>All the rules are defined as having the std::string attribute. I use
<br>qi::parse() to parse so that I don't use a skip parser.
<br><br>This works correctly for parsing something like this, "one\ntwo\ntwo\none"
<br>returning "ONE\nTWO\nTWO\nONE".
<br><br>So now if I want it to handle lines that don't start with "one" or "two" by just
<br>passing them though I add the rule "bypassLine"
<br>...
<br> lines2Convert %= ((oneLine | twoLine | bypassLine)
<br> >> (eol[_val += '\n'] | eoi));
<br> bypassLine %= *(char_ - (eol | eoi));
<br>...
<br><br>This will hang when parsing this "one\ntwo\nNotALine\ntwo\none\n" (actually it
<br>hangs with anything passed to it)...
<br><br>couple of things:
<br>1) I can't determine what is going on, it looks like it should work
<br>2) Is there any good way to debug something like this?
<br><br><b>Thanks in advance!!!</b><br><br>dWiGhT
</div>
</div></blockquote>
<p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26819267hanging grammar??2009-12-16T14:03:48Z2009-12-16T14:03:48ZdWiGhT-TXFirst a disclaimer: I'm a Spirit virgin and just starting to play around with
<br>v2.1... so please keep that in mind. (thanks)
<br><br>Ok... playing around with Spirit v2.1 and put this together to take a string as
<br>input and output the converted string as a little test.
<br><br>Here is the part of the grammar:
<br>...
<br> convertLines = eps[_val = ""] >> *lines2Convert[_val += _1];
<br> lines2Convert %= ((oneLine | twoLine ) >> (eol[_val += '\n'] | eoi));
<br> oneLine = lit("one")[_val = "ONE"];
<br> twoLine = lit("two")[_val = "TWO"];
<br>...
<br><br>All the rules are defined as having the std::string attribute. I use
<br>qi::parse() to parse so that I don't use a skip parser.
<br><br>This works correctly for parsing something like this, "one\ntwo\ntwo\none"
<br>returning "ONE\nTWO\nTWO\nONE".
<br><br>So now if I want it to handle lines that don't start with "one" or "two" by just
<br>passing them though I add the rule "bypassLine"
<br>...
<br> lines2Convert %= ((oneLine | twoLine | bypassLine)
<br> >> (eol[_val += '\n'] | eoi));
<br> bypassLine %= *(char_ - (eol | eoi));
<br>...
<br><br>This will hang when parsing this "one\ntwo\nNotALine\ntwo\none\n" (actually it
<br>hangs with anything passed to it)...
<br><br>couple of things:
<br>1) I can't determine what is going on, it looks like it should work
<br>2) Is there any good way to debug something like this?
<br><br><b>Thanks in advance!!!</b><br><br>dWiGhT
<br><p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26814819Re: match |foobar| not |foo bar |2009-12-16T09:12:39Z2009-12-16T09:12:39ZDaniel James2009/12/16 bytecolor <<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26814819&i=0" target="_top" rel="nofollow">bytecolor@...</a>>:
<br>>
<br>> Ah, thanks OvermindDL1. I had to finesse it a bit:
<br>>
<br>> lexeme[alpha[_val += _1]
<br>> >> -(+alnum[_val += _1] % blank[_val += "_"])]
<br>>
<br>> No more chopper.
<br><br>Do you want it to match 'a b'? How about:
<br><br>lexeme[&alpha
<br> >> (+alnum[_val += _1] % blank[_val += "_"])]
<br><br>The '&alpha' matches a character but doesn't consume it. And if you
<br>want to allow more than one consecutive whitespace character:
<br><br>lexeme[&alpha
<br> >> (+alnum[_val += _1] % (+blank)[_val += "_"])]
<br><br>Daniel
<br><br>------------------------------------------------------------------------------
<br>This SF.Net email is sponsored by the Verizon Developer Community
<br>Take advantage of Verizon's best-in-class app development support
<br>A streamlined, 14 day to market process makes app distribution fast and easy
<br>Join now and get one step closer to millions of Verizon customers
<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_top" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a>
<br>_______________________________________________
<br>Spirit-general mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26814819&i=1" target="_top" rel="nofollow">Spirit-general@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a><br><p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26814365calc_ast: ast analysis2009-12-16T08:47:57Z2009-12-16T08:47:57Zhaunsone of the advertised benefits for spirit2 is the fact that the ast comes for free (albeit not very well documented). as some of you know i have been looking into the calc2_ast example recently.
<br><br><a href="https://svn.boost.org/trac/boost/browser/trunk/libs/spirit/example/qi/calc2_ast.cpp" target="_top" rel="nofollow">https://svn.boost.org/trac/boost/browser/trunk/libs/spirit/example/qi/calc2_ast.cpp</a>
<br><br>i am interested in tagging the tree with level numbers. e.g., topmost (level 0) is an (binary) operator, which has a left and right and side expression underneath (levels 1, each). is there an elegant way to get to those levels? specifically,
<br>- should i attempt the level tagging while parsing, i.e., in the grammar (i currently do not see a way how a parsing rule can know about is parent, i.e., the rule it was called from)
<br>- or should this ideally be done in a post processing step (i.e., walking the tree akin to ast_print and trying to keep an index somehow)
<br>- for both alternatives i do not see how to pass down an index elegantly as it is all based on rigid boost interfaces (apply_visitor)
<br><br>i'd like to do the right thing from the beginning, so any insights/suggestions (maybe with reference to the example) are very much appreciated.
<br><br><br>------------------------------------------------------------------------------
<br>This SF.Net email is sponsored by the Verizon Developer Community
<br>Take advantage of Verizon's best-in-class app development support
<br>A streamlined, 14 day to market process makes app distribution fast and easy
<br>Join now and get one step closer to millions of Verizon customers
<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_top" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a>
<br>_______________________________________________
<br>Spirit-general mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26814365&i=0" target="_top" rel="nofollow">Spirit-general@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a><br><p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26813448Re: ast for expressions: phoenix functions2009-12-16T07:52:43Z2009-12-16T07:52:43Zhaunsi didn't try this one but declaring TWO locals worked. thanks
<br>everyone who contributed!i feel much more comfortable with the
<br> framework now, only after a few days.
<br><br><br><br><br><br>------------------------------------------------------------------------------
<br>This SF.Net email is sponsored by the Verizon Developer Community
<br>Take advantage of Verizon's best-in-class app development support
<br>A streamlined, 14 day to market process makes app distribution fast and easy
<br>Join now and get one step closer to millions of Verizon customers
<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_top" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a>
<br>_______________________________________________
<br>Spirit-general mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26813448&i=0" target="_top" rel="nofollow">Spirit-general@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a><br><p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26812674Re: match |foobar| not |foo bar |2009-12-16T07:13:47Z2009-12-16T07:13:47Zbytecolor-2OvermindDL1 <overminddl1 <at> gmail.com> writes:
<br><div class='shrinkable-quote'><br>> Remember, Spirit is a PEG parser, in PEG, *EVERYTHING* is greedy. You
<br>> might try this:
<br>>
<br>> raw[alpha >> (alnum%blank)]
<br>>
<br>> That should return a string (raw returns a string that is matches by
<br>> the containing) that starts with an alpha, and is following by alnum
<br>> and blanks, but does not end with a blank.
<br>>
<br>> % means do the left separated by the right, meaning the right can be
<br>> contained, but will not appear at the end.
</div><br>Ah, thanks OvermindDL1. I had to finesse it a bit:
<br><br>lexeme[alpha[_val += _1]
<br> >> -(+alnum[_val += _1] % blank[_val += "_"])]
<br><br>No more chopper.
<br><br><br>------------------------------------------------------------------------------
<br>This SF.Net email is sponsored by the Verizon Developer Community
<br>Take advantage of Verizon's best-in-class app development support
<br>A streamlined, 14 day to market process makes app distribution fast and easy
<br>Join now and get one step closer to millions of Verizon customers
<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_top" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a>
<br>_______________________________________________
<br>Spirit-general mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26812674&i=0" target="_top" rel="nofollow">Spirit-general@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a><br><p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26812546Re: Parsing simple sequence of two double values2009-12-16T07:06:55Z2009-12-16T07:06:55ZbenCFDI'm sorry, it was my fault. It didn't recompile correctly. It's working fine now!
<br><br>Thanks a lot!
<br><br><br><blockquote class="quote light-black dark-border-color"><div class="quote light-border-color">
<div class="quote-author" style="font-weight: bold;">Hartmut Kaiser wrote:</div>
<div class="quote-message shrinkable-quote">> I'm currently doing my first steps with Spirit 2.1. I'm trying to parse
<br>> a
<br>> very simple sequence of two double values. My parsing and main methods
<br>> look
<br>> like this:
<br>>
<br>> template <typename Iterator>
<br>> bool Parse(Iterator first, Iterator last)
<br>> {
<br>> using ascii::space;
<br>> using qi::phrase_parse;
<br>> using qi::double_;
<br>>
<br>> bool r = phrase_parse(
<br>> first,
<br>> last,
<br>> double_ >> double_,
<br>> space
<br>> );
<br>>
<br>> if(first != last)
<br>> return false; // method returns here
<br>> return r;
<br>> }
<br>>
<br>> int main()
<br>> {
<br>> cout << "Input: ";
<br>> string str;
<br>> cin >> str;
<br>>
<br>> if(Parse(str.begin(), str.end()))
<br>> cout << "Parsing successful" << endl;
<br>> else
<br>> cout << "Parsing failed" << endl;
<br>> getchar();
<br>>
<br>> return 0;
<br>> }
<br>>
<br>> However if I input a sequence like "123 456" and start parsing, the
<br>> message
<br>> returns false, because of the inequality check of the iterators. A more
<br>> complex rule like
<br>>
<br>> double_ >> *(' ' >> double_)
<br>>
<br>> works quite well, even for the above mentioned example of "123 456".
<br>> What's
<br>> the problem with my code?
<br>>
<br>> It is probably a silly beginner's problem, I'm sorry. But I've been
<br>> trying
<br>> to sort it out for two days now ...
<br><br>Your code leaves a newline character at the end of the input which won't be
<br>matched by your grammar. So the input will not consumed completely. Use
<br>ascii::blank as your skipper and it will work.
<br><br>Regards Hartmut
<br><br>---------------
<br>Meet me at BoostCon
<br>www.boostcon.com
<br><br><br><br><br>------------------------------------------------------------------------------
<br>This SF.Net email is sponsored by the Verizon Developer Community
<br>Take advantage of Verizon's best-in-class app development support
<br>A streamlined, 14 day to market process makes app distribution fast and easy
<br>Join now and get one step closer to millions of Verizon customers
<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_top" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a>
<br>_______________________________________________
<br>Spirit-general mailing list
<br>Spirit-general@lists.sourceforge.net
<br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a></div>
</div></blockquote>
<p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26812483Re: Parsing simple sequence of two double values2009-12-16T07:03:22Z2009-12-16T07:03:22ZbenCFDHartmut, thank you very much for your quick response.
<br>I really seem to be too stupid for this framework. I've replaced my "space" skipping parser with "blank" but this doesn't solve the problem. I'm still getting the same result ...
<br><br>using ascii::blank;
<br>using qi::phrase_parse;
<br>using qi::double_;
<br><br>bool r = phrase_parse(
<br> first,
<br> last,
<br> double_ >> double_,
<br> blank
<br>);
<br><br><blockquote class="quote light-black dark-border-color"><div class="quote light-border-color">
<div class="quote-author" style="font-weight: bold;">Hartmut Kaiser wrote:</div>
<div class="quote-message shrinkable-quote">> I'm currently doing my first steps with Spirit 2.1. I'm trying to parse
<br>> a
<br>> very simple sequence of two double values. My parsing and main methods
<br>> look
<br>> like this:
<br>>
<br>> template <typename Iterator>
<br>> bool Parse(Iterator first, Iterator last)
<br>> {
<br>> using ascii::space;
<br>> using qi::phrase_parse;
<br>> using qi::double_;
<br>>
<br>> bool r = phrase_parse(
<br>> first,
<br>> last,
<br>> double_ >> double_,
<br>> space
<br>> );
<br>>
<br>> if(first != last)
<br>> return false; // method returns here
<br>> return r;
<br>> }
<br>>
<br>> int main()
<br>> {
<br>> cout << "Input: ";
<br>> string str;
<br>> cin >> str;
<br>>
<br>> if(Parse(str.begin(), str.end()))
<br>> cout << "Parsing successful" << endl;
<br>> else
<br>> cout << "Parsing failed" << endl;
<br>> getchar();
<br>>
<br>> return 0;
<br>> }
<br>>
<br>> However if I input a sequence like "123 456" and start parsing, the
<br>> message
<br>> returns false, because of the inequality check of the iterators. A more
<br>> complex rule like
<br>>
<br>> double_ >> *(' ' >> double_)
<br>>
<br>> works quite well, even for the above mentioned example of "123 456".
<br>> What's
<br>> the problem with my code?
<br>>
<br>> It is probably a silly beginner's problem, I'm sorry. But I've been
<br>> trying
<br>> to sort it out for two days now ...
<br><br>Your code leaves a newline character at the end of the input which won't be
<br>matched by your grammar. So the input will not consumed completely. Use
<br>ascii::blank as your skipper and it will work.
<br><br>Regards Hartmut
<br><br>---------------
<br>Meet me at BoostCon
<br>www.boostcon.com
<br><br><br><br><br>------------------------------------------------------------------------------
<br>This SF.Net email is sponsored by the Verizon Developer Community
<br>Take advantage of Verizon's best-in-class app development support
<br>A streamlined, 14 day to market process makes app distribution fast and easy
<br>Join now and get one step closer to millions of Verizon customers
<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_top" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a>
<br>_______________________________________________
<br>Spirit-general mailing list
<br>Spirit-general@lists.sourceforge.net
<br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a></div>
</div></blockquote>
<p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>tag:old.nabble.com,2006:post-26812492Re: Qi and boost::arrays2009-12-16T07:03:01Z2009-12-16T07:03:01ZHartmut Kaiser> > It's now fixed in SVN. The fix has the nice side effect of not
<br>> requiring
<br>> > the
<br>> > container to be default constructible anymore, BTW.
<br>>
<br>> Thanks, Hartmut. I'd like to give this a try, but I haven't been using
<br>> Spirit from SVN. Is it as simple as replacing Spirit in my Boost folder
<br>> with a checkout from the repository?
<br><br>Yes, this should do the trick. Just put the checked out spirit folder in
<br>front of your boost root onto the include search path of your compiler (see
<br>here for some hints: <a href="http://tinyurl.com/y9gdtnn" target="_top" rel="nofollow">http://tinyurl.com/y9gdtnn</a>).
<br><br>Regards Hartmut
<br><br>---------------
<br>Meet me at BoostCon
<br>www.boostcon.com
<br><br><br><br><br>------------------------------------------------------------------------------
<br>This SF.Net email is sponsored by the Verizon Developer Community
<br>Take advantage of Verizon's best-in-class app development support
<br>A streamlined, 14 day to market process makes app distribution fast and easy
<br>Join now and get one step closer to millions of Verizon customers
<br><a href="http://p.sf.net/sfu/verizon-dev2dev" target="_top" rel="nofollow">http://p.sf.net/sfu/verizon-dev2dev</a>
<br>_______________________________________________
<br>Spirit-general mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26812492&i=0" target="_top" rel="nofollow">Spirit-general@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/spirit-general" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/spirit-general</a><br><p>From forum: <a href="http://old.nabble.com/spirit-general-f3432.html" embed="fixTarget[3432]" target="_top" >spirit-general</a></p>