How to combine skipping and non-skipping (lexeme) rules? - c++

my parser is nearly working :)
(still amazed by Spirit feature set (and compiletimes) and the very welcoming community here on stack overflow)
small sample for online try:
http://coliru.stacked-crooked.com/a/1c1bf88909dce7e3
so i've learned to use more lexeme-rules and try to prevent no_skip -
my rules are smaller and better to read as a result but now i stuck with
combining lexeme-rules and skipping-rules what seems to be not possible (compiletime error with warning about not castable to Skipper)
my problem is the comma seperated list in subscriptions
which does not skip spaces around expressions
parses:
"a.b[a,b]"
fails:
"a.b[ a , b ]"
these are my rules:
qi::rule<std::string::const_iterator, std::string()> identifier_chain;
qi::rule<std::string::const_iterator, std::string()>
expression_list = identifier_chain >> *(qi::char_(',') >> identifier_chain);
qi::rule < std::string::const_iterator, std::string() >
subscription = qi::char_('[') >> expression_list >> qi::char_(']');
qi::rule<std::string::const_iterator, std::string()>
identifier = qi::ascii::alpha >> *(qi::ascii::alnum | '_');
identifier_chain = identifier >> *(('.' >> identifier) | subscription);
as you can see all rules are "lexeme" and i think the subscription rule should be a ascii::space_type skipper but that does not compile
should i add space eaters in the front and back of identifier_chains in the expression_list?
feels like writing an regex :(
expression_list = *qi::blank >> identifier_chain >> *(*qi::blank >> qi::char_(',') >> *qi::blank >> identifier_chain >> *qi::blank);
it works but i've read that this will get me to an much bigger parser in the end (handling all the space skipping by myself)
thx for any advice
btw: any idea why i can't compile if surrounding the '.' in the indentifier_chain with qi::char_('.')
identifier_chain = identifier >> *(('.' >> identifier) | subscription);
UPDATE:
i've updated my expression list as suggested by sehe
qi::rule<std::string::const_iterator, spirit::ascii::blank_type, std::string()>
expression_list = identifier_chain >> *(qi::char_(',') >> identifier_chain);
qi::rule < std::string::const_iterator, std::string() >
subscription = qi::char_('[') >> qi::skip(qi::blank)[expression_list] >> qi::char_(']');
but still get compile error due to non castable Skipper: http://coliru.stacked-crooked.com/a/adcf665742b055dd
i also tried changed the identifer_chain to
identifier_chain = identifier >> *(('.' >> identifier) | qi::skip(qi::blank)[subscription]);
but i still can't compile the example

The answer I linked to earlier describes all the combinations (if I remember correctly): Boost spirit skipper issues
In short:
any rule that declares a skipper (so rule<It, Skipper[, Attr()]> or rule<It, Attr(), Skipper>) MUST be invoked with a compatible skipper (an expression that can be assigned to the type of Skipper).
any rule that does NOT declare a skipper (so of the form rule<It[, Attr()]>) will implicitly behave like a lexeme, meaning no input characters are skipped.
That's it. The slightly subtler ramifications are that given two rules:
rule<It, blank_type> a;
rule<It> b; // b is implicitly lexeme
You can invoke b from a:
a = "test" >> b;
But when you wish to invoke a from b you will find that you have to provide the skipper:
b = "oops" >> a; // DOES NOT COMPILE
b = "okay" >> qi::skip(qi::blank) [ a ];
That's almost all there is to it. There are a few more directives around skippers and lexemes in Qi, see again the answer linked above.
Side Question:
should i add space eaters in the front and back of identifier_chains in the expression_list?
If you look closely at the answer example here Parse a '.' chained identifier list, with qi::lexeme and prevent space skipping, you can see that it already does pre- and post skipping correctly, because I used phrase_parse:
" a.b " OK: ( "a" "b" )
----
"a . b" Failed
Remaining unparsed: "a . b"
----
You COULD also wrap the whole thing in an "outer" rule:
rule<std::string::const_iterator> main_rule =
qi::skip(qi::blank) [ identifier_chain ];
That's just the same but allows users to call parse without specifying the skipper.

Related

Spirit Qi: Error when replacing sequence with expectation operator

I have a grammar that works perfectly fine and contains the following lines.
element = container | list | pair;
container = name >> '(' >> -(arg % ',') >> ')' >> '{' >> +element > '}';
// trying to put an expectation operator here --------^
list = name >> '(' > (value % ',') > ')' > ';';
pair = name >> ':' > value > ';';
To have meaningful error messages, I want to make sure that container does not backtrack as soon as it hits '{'. But for some reason, if I replace the sequence operator with an expectation operator right after the '{', I get a huge compiler error. Any ideas what the problem might be?
element is a boost::variant; container, list and pair are own structs with BOOST_FUSION_ADAPT_STRUCT applied. Please have a look here for the full source code: https://github.com/fklemme/liberty_tool/blob/master/src/liberty_grammar.hpp#L24
Yes. Because the precedences of operator>> and operator> aren't equal, the resulting synthesized attribute type is different.
In fact, it is no longer automatically compatible with the intended exposed attribute type.
In this case the problem can be quickly neutralized with some disambiguating parentheses around the sub-expression:
container = name >> '(' >> -(arg % ',') >> ')' >> ('{' > +element > '}');

Boost Spirit - Trimming spaces between last character and separator

Boost Spirit newcomer here.
I have a string in the form of "Key:Value\r\nKey2:Value2\r\n" that I'm trying to parse. In that specific form, it's trivial to parse with Boost Spirit. However, in order to be more robust, I also need to handle cases such as this one:
" My Key : Value \r\n My2ndKey : Long<4 spaces>Value \r\n"
In this case, I need to trim leading and trailing spaces before and after the key/value separators so that I get the following map:
"My Key", "Value"
"My2ndKey", "Long<4 spaces>Value"
I played with qi::hold to achieve this but I get compile errors because of unsupported boost::multi_pass iterator with the embedded parser I was trying to use. There has to be a simple way to achieve this.
I read the following articles (and many others on the subject):
http://boost-spirit.com/home/articles/qi-example/parsing-a-list-of-key-value-pairs-using-spirit-qi/
http://boost-spirit.com/home/2010/02/24/parsing-skippers-and-skipping-parsers/
Boost spirit parsing string with leading and trailing whitespace
I am looking for a solution to my problem, which doesn't seem to be entirely covered by those articles. I would also like to better understand how this is achieved. As a small bonus question, I keep seeing the '%=' operator, is this useful in my case? MyRule %= MyRule ... is used for recursive parsing?
The code below parses my strings properly except that it doesn't remove the spaces between the last non-space character and the separator. :( The skipper used is qi::blank_type (space without EOL).
Thanks!
template <typename Iterator, typename Skipper>
struct KeyValueParser : qi::grammar<Iterator, std::map<std::string, std::string>(), Skipper> {
KeyValueParser() : KeyValueParser::base_type(ItemRule) {
ItemRule = PairRule >> *(qi::lit(END_OF_CMD) >> PairRule);
PairRule = KeyRule >> PAIR_SEP >> ValueRule;
KeyRule = +(qi::char_ - qi::lit(PAIR_SEP));
ValueRule = +(qi::char_ - qi::lit(END_OF_CMD));
}
qi::rule<Iterator, std::map<std::string, std::string>(), Skipper> ItemRule;
qi::rule<Iterator, std::pair<std::string, std::string>(), Skipper> PairRule;
qi::rule<Iterator, std::string()> KeyRule;
qi::rule<Iterator, std::string()> ValueRule;
};
You need to use KeyRule = qi::raw[ +(qi::char_ - qi::lit(PAIR_SEP)) ];
In order to see why, let's try to study several ways to parse the string "a b :".
First let's keep in mind how the following parsers/directives work:
lexeme[subject]: This directive matches subject while disabling the skipper.
raw[subject]: Discards subject's attribute and returns an iterator pair that points to the matched characters in the input stream.
+subject: The plus parser tries to match 1 or more times its subject.
a-b: The difference parser first tries to parse b and if b succeeds, a-b fails. When b fails, it matches a.
char_: matches any char. It's a PrimitiveParser.
lit(':'): matches ':' but ignores its attribute. It's a PrimitiveParser.
lexeme[ +(char_ - lit(':')) ]: by removing the skipper from your rule you have an implicit lexeme. Since there is no skipper it goes like this:
'a' -> ':' fails, char_ matches 'a', the current synthesized attribute is "a"
' ' -> ':' fails, char_ matches ' ', the current synthesized attribute is "a "
'b' -> ':' fails, char_ matches 'b', the current synthesized attribute is "a b"
' ' -> ':' fails, char_ matches ' ', the current synthesized attribute is "a b "
':' -> ':' succeeds, the final synthesized attribute is "a b "
+(char_ - lit(':')): Since it has a skipper every PrimitiveParser will pre-skip before being tried:
'a' -> ':' fails, char_ matches 'a', the current synthesized attribute is "a"
' ' -> this is skipped before ':' is tried
'b' -> ':' fails, char_ matches 'b', the current synthesized attribute is "ab"
' ' -> this is skipped before ':' is tried
':' -> ':' succeeds, the final synthesized attribute is "ab"
raw[ +(char_ - lit(':') ) ]: The subject is exactly the same as 2.. The raw directive ignores "ab" and returns an iterator pair that goes from 'a' to 'b'. Since the attribute of your rule is std::string, a string is constructed from that iterator pair, resulting in "a b" which is what you want.

Trying to understand Boost Qi parsing into structs

I've got an embarrassingly simple problem that I can't seem to wrap my head around. I'm reading the boost documentation on how to parse into structs. The sample code provided for that chapter is straightforward - or so I thought. I would like to make a super simple change.
I want to split the start-rule:
start %=
lit("employee")
>> '{'
>> int_ >> ','
>> quoted_string >> ','
>> quoted_string >> ','
>> double_
>> '}'
;
...into two (or later more) rules, like this:
params %=
>> int_ >> ','
>> quoted_string >> ','
>> quoted_string >> ','
>> double_;
start %=
lit("employee")
>> '{'
>> params
>> '}'
;
No matter what I've tried I couldn't get it to parse values correctly into the employee struct. Even when I got a running program that recognized the input, the attributes didn't get written to the struct. It seems parsing only works correctly if everything is specified in the "top-level" rule. Surely, I'm mistaken?! I'll definitely need a more structured approach for the parser I actually need to implement.
Also I'm unclear what the correct type of the params rule should be. I'm thinking qi::rule<Iterator, fusion::vector<int, std::string, std::string, double>, ascii::space_type>, but my compiler didn't seem to like that very much...
I should mention that I'm working with Boost v1.46.1
In this situation, you could really just make params expose an employee attribute directly:
Live On Coliru
qi::rule<Iterator, employee(), ascii::space_type> params;

Parsing a string (with spaces) but ignoring the spaces at the end of the (Spirit)

I have an input string I'm trying to parse. It might look like either of the two:
sys(error1, 2.3%)
sys(error2 , 2.4%)
sys(this error , 3%)
Note the space sometimes before the comma. In my grammer (boost spirit library) I'd like to capture "error1", "error2", and "this error" respectively.
Here is the original grammar I had to capture this - which absorbed the space at the end of the name:
name_string %= lexeme[+(char_ - ',' - '"')];
name_string.name("Systematic Error Name");
start = (lit("sys")|lit("usys")) > '('
> name_string[boost::phoenix::bind(&ErrorValue::SetName, _val, _1)] > ','
> errParser[boost::phoenix::bind(&ErrorValue::CopyErrorAndRelative, _val, _1)]
> ')';
My attempt to fix this was first:
name_string %= lexeme[*(char_ - ',' - '"') > (char_ - ',' - '"' - ' ')];
however that completely failed. Looks like it failes to parse anything with a space in the middle.
I'm fairly new with Spirit - so perhaps I'm missing something simple. Looks like lexeme turns off skipping on the leading edge - I need something that does it on the leading and trailing edge.
Thanks in advance for any help!
Thanks to psur below, I was able to put together an answer. It isn't perfect (see below), but I thought I would update the post for everyone to see it in context and nicely formatted:
qi::rule<Iterator, std::string(), ascii::space_type> name_word;
qi::rule<Iterator, std::string(), ascii::space_type> name_string;
ErrorValueParser<Iterator> errParser;
name_word %= +(qi::char_("_a-zA-Z0-9+"));
//name_string %= lexeme[name_word >> *(qi::hold[+(qi::char_(' ')) >> name_word])];
name_string %= lexeme[+(qi::char_("-_a-zA-Z0-9+")) >> *(qi::hold[+(qi::char_(' ')) >> +(qi::char_("-_a-zA-Z0-9+"))])];
start = (
lit("sys")[bind(&ErrorValue::MakeCorrelated, _val)]
|lit("usys")[bind(&ErrorValue::MakeUncorrelated, _val)]
)
>> '('
>> name_string[bind(&ErrorValue::SetName, _val, _1)] >> *qi::lit(' ')
>> ','
>> errParser[bind(&ErrorValue::CopyErrorAndRelative, _val, _1)]
>> ')';
This works! They key to this is the name_string, and in it the qi::hold, a operator I was not familiar with before this. It is almost like a sub-rule: everything inside qi::hold[...] must successfully parse for it to go. So, above, it will only allow a space after a word if there is another word following. The result is that if a sequence of words end in a space(s), those last spaces will not be parsed! They can be absorbed by the *qi::lit(' ') that follows (see the start rule).
There are two things I'd like to figure out how to improve here:
It would be nice to put the actual string parsing into name_word. The problem is the declaration of name_word - it fails when it is put in the appropriate spot in the definition of name_string.
It would be even better if name_string could include the parsing of the trailing spaces, though its return value did not. I think I know how to do that...
When/if I figure these out I will update this post. Thanks for the help!
Below rules should work for you:
name_word %= +(qi::char_("_a-zA-Z0-9"));
start %= qi::lit("sys(")
>> qi::lexeme[ name_word >> *(qi::hold[ +(qi::char_(' ')) >> name_word ]) ]
>> *qi::lit(' ')
>> qi::lit(',')
// ...
name_word parse only one word in name; I assumed that it contains only letter, digits and underscore.
In start rule qi::hold is important. It will parse space only if next is name_word. In other case parser will rollback and move to *qi::lit(' ') and then to comma.

Parsing string, with Boost Spirit 2, to fill data in user defined struct

I'm using Boost.Spirit which was distributed with Boost-1.42.0 with VS2005. My problem is like this.
I've this string which was delimted with commas. The first 3 fields of it are strings and rest are numbers. like this.
String1,String2,String3,12.0,12.1,13.0,13.1,12.4
My rule is like this
qi::rule<string::iterator, qi::skip_type> stringrule = *(char_ - ',')
qi::rule<string::iterator, qi::skip_type> myrule= repeat(3)[*(char_ - ',') >> ','] >> (double_ % ',') ;
I'm trying to store the data in a structure like this.
struct MyStruct
{
vector<string> stringVector ;
vector<double> doubleVector ;
} ;
MyStruct var ;
I've wrapped it in BOOST_FUSION_ADAPT_STRUCTURE to use it with spirit.
BOOST_FUSION_ADAPT_STRUCT (MyStruct, (vector<string>, stringVector) (vector<double>, doubleVector))
My parse function parses the line and returns true and after
qi::phrase_parse (iterBegin, iterEnd, myrule, boost::spirit::ascii::space, var) ;
I'm expecting var.stringVector and var.doubleVector are properly filled. but it is not the case.
What is going wrong ?
The code sample is located here
Thanks in advance,
Surya
qi::skip_type is not something you could use a skipper. qi::skip_type is the type of the placeholder qi::skip, which is applicable for the skip[] directive only (to enable skipping inside a lexeme[] or to change skipper in use) and which is not a parser component matching any input on its own. You need to specify your specific skipper type instead (in your case that's boost::spirit::ascii:space_type).
Moreover, in order for your rules to return the parsed attribute, you need to specify the type of the expected attribute while defining your rule. That leaves you with:
qi::rule<string::iterator, std::string(), ascii:space_type>
stringrule = *(char_ - ',');
qi::rule<string::iterator, MyStruct(), ascii:space_type>
myrule = repeat(3)[*(char_ - ',') >> ','] >> (double_ % ',');
which should do exactly what you expect.