Re: Small grammar, long compilation time

On 10/27/2017 08:24 AM, Søren Enevoldsen wrote:
> On 2017-10-27 14:41, Larry Evans wrote:
>> On 10/26/2017 04:33 PM, Larry Evans wrote:
>>> On 10/26/2017 09:03 AM, Søren Enevoldsen wrote:
>>>> The following program takes about 10s to compile. However, the rule 
>>>> currently used in parenProcess, 'nullProcess' should actually be 
>>>> 'process'. However, when I make the change, compilation does not 
>>>> finish. Instead CPU keeps spinning hot, and memory usage gradually 
>>>> increases. Since it is a rather small grammar, I am guessing I am 
>>>> doing something wrong. How do I make the grammar work then I correct 
>>>> it by changing parenProcess to '(' > process > ')' ? (Note I cannot 
>>>> reduce the number of rules between process and parenProcess for a 
>>>> smaller example, since then the compile time immediately decreases).
>>>>
>>>> - Søren
>>>>
>>>> ====================================================================
>>>>
>>>> #include <boost/config/warning_disable.hpp>
>>>> #include <boost/spirit/home/x3.hpp>
>>>> #include <boost/spirit/home/x3/support/ast/variant.hpp>
>>>> #include <boost/spirit/include/qi_char_class.hpp>
>>>> #include <boost/fusion/include/adapt_struct.hpp>
>>> [snip]
>>>> x3::rule<struct processRule, AstAnyProcess> process{"process"};
>>>>
>>>> ///  Here be compile time dragons.
>>>> auto const parenProcess = '(' > nullProcess > ')';
>>>>
>>>> auto const primitive = rule<AstAnyProcess>("primitive")
>>>>      = parenProcess
>>>>      | nullProcess
>>>>      | constant;
>>> [snip]
>>> With this "convenience" change:
>>>
>>> #define USE_PROCESS_IN_PAREN
>>> auto const parenProcess
>>>    = '('
>>> #ifdef USE_PROCESS_IN_PAREN
>>>    ///  Here be compile time dragons.
>>>    > process
>>> #else
>>>    > nullProcess
>>> #endif
>>>    > ')'
>>>    ;
>>>
>>> and when compiled, it did take more than 10s (way more)
>>> and I did see my memory use go up and 100% of some of
>>> my cpu's.  However, when the get_rhs branch of a spirit
>>> fork here:
>>>
>>> https://github.com/cppljevans/spirit/tree/get_rhs
>>>
>>> was -I included in front of the standard boost,
>>> it compiled in a flash.
>>>
>>> This problem has been reported before; however, I've
>>> been advised that it needs more documentation before
>>> submitting it as a pull request.  I'm working slowly
>>> on that.
>>>
>>> HTH.
>>>
>>> -regards,
>>> Larry
>>
>> If you limit the instantiation depth to 200, you can
>> get some idea of why it fails as shown in the attached.
>> In brief, the context is being repeatedly augmented
>> by the make_unique_context leading to the excessive
>> instantiations.  The get_rhs branch avoids that
>> repeated call to make_unique_context.
>>
>> -regards,
>> Larry
>>
>>
>> ------------------------------------------------------------------------------ 
>>
>> Check out the vibrant tech community on one of the world's most
>> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
>>
>>
>> _______________________________________________
>> Spirit-general mailing list
>> [email protected]
>> https://lists.sourceforge.net/lists/listinfo/spirit-general
> 
> Thanks for the explanation and pointers Larry. But I am not sure what is 
> happening for me.
> 
> I am using GCC 7.2.0 and compile using:
> g++ wccs_parser.cpp -Wall -pedantic -O2 -march=x86-64 -std=c++17 -m64 
> -I. -flto=4 -static-libgcc -static-libstdc++ -ljemalloc -o wccs_parser
> 
> A. Compiling without the equivalent of "USE_PROCESS_IN_PAREN" takes ~14s 
> and ~930MB (maximum resident).
> 
> 1. I clone your repo into ~/spirit
> 2. I add ~/spirit to very front of my CPLUS_INCLUDE_PATH . The full 
> boost package (1.65.1) is very last. In fact these are the only two 
> paths there.
> 3. I compile again and it takes ~15s and ~925MB.
> 
> I am quite sure it now uses your ~/spirit because now I get a pragma 
> warning from the original boost (which I did not get before):
> 
> /home/soren/spirit/include/boost/spirit/home/x3/nonterminal/rule.hpp:23:19: 
> note: #pragma message: not(BOOST_SPIRIT_ATTR_XFORM_IN_RULE)
>     #pragma message "not(BOOST_SPIRIT_ATTR_XFORM_IN_RULE)"
>                     ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> /home/soren/spirit/include/boost/spirit/home/x3/nonterminal/rule.hpp:28:19: 
> note: #pragma message: not(BOOST_SPIRIT_GET_RHS_CRTP)
>     #pragma message "not(BOOST_SPIRIT_GET_RHS_CRTP)"

Ah!  I forgot to add that you need to #define those 2 macros to
avoid the long compile times.  Otherwise, the branch behaves
just like the existing spirit.

> 
> And I can still not compile with "USE_PROCESS_IN_PAREN".
> 
> - Søren
> 
> 
> 
> 
> 
> ------------------------------------------------------------------------------
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
> 
> 
> 
> _______________________________________________
> Spirit-general mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/spirit-general
> 



------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Spirit-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/spirit-general