Re: glr2.cc compile errors under Windows

[Jot Dot]

>> "You need to compile with the /Zc:__cplusplus switch to see the updated value
>> of the __cplusplus macro. We tried updating the macro by default and 
>> discovered
>> that a lot of code doesn’t compile correctly when we change the value of
>> __cplusplus"

> No worries.  Should I understand that now it works as expected?

Ouch. Last night I dissected my code to introduce a major rewrite.
For you, I went to a backup and checked. I get the following error only with 
glr2.cc :

1>D:\data\c\gen - Copy\runtime\scanner.h(35,55): error C2039: 'semantic_type': 
is not a member of 'gen::Parser'
1>D:\data\c\gen - Copy\runtime\parser.h(122): message : see declaration of 
'gen::Parser'
... other irrelevant "Now I'm confused" errors

In my scanner.h file I have:
#undef YY_DECL
#define YY_DECL int gen::Scanner::get_next_token(union 
gen::Parser::semantic_type& yylval, location& yylloc, Driver* pDriver)

This worked for me with everything but glr2. (ie: glr.cc, lalr.cc)
Like usual, I'm probably not doing something quite right here.
Unless you have an idea, I'll have to check into this later and get back to you.

---------------

@Hans  - thanks for your input. I just wanted to say this to clarify the 
problem:

To anyone using Microsoft compilers:

In MSVC you can get to the properties page for your project and select the C++ 
version
ie: /std:c++17
You'd assume this switch affects the value of __cplusplus but it does not!
You have to manually add another switch to tell the compiler to update the 
__cplusplus value
that is reported: This is the /Zc:__cplusplus switch.
Otherwise, without the /Zc switch, the compiler will always report __cplusplus 
as 199711L

Thanks everyone