[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Visual Studio - win_bison creates tab.h file exceeding compiler limi

From: Akim Demaille
Subject: Re: Visual Studio - win_bison creates tab.h file exceeding compiler limits
Date: Mon, 9 Nov 2020 19:09:08 +0100


> Le 9 nov. 2020 à 13:08, Jot Dot <jotdot@shaw.ca> a écrit :
> bison generates the tab.h file using YY_ASSERT macros in many places. 
> bison is generating lines longer than the microsoft compiler allows, which 
> causes a C2026 error. (Line exceeds 16380 bytes). 

Bummer.  My bad, I had not anticipated this.

> I am starting to write a SQL parser.

Sounds fun ;)

> bison is generating asserts to check for valid tokens but a couple lines are 
> over 21700 bytes long. 
> Since the compiler truncates the line,


> any YY_ASSERT() macro fails because the line is simply snipped. Thus it does 
> not even encounter the trailing ')'. Naturally the compiler spits out 
> gibberish at this point.

Always fun to workaround limitations.

> Although I've known about lex/yacc and flex/bison for years, this is my first 
> real attempt to use it. This means I could be doing something "non-standard". 

Nope, AFAICT, you're on the right tracks.  Wrong compiler though.

> But my work so far was working with the standard C generation method. It is 
> only now, converting it to C++, did I encounter this. 
> It's easy enough to build, manually comment out the offending line, and then 
> build again, but this is getting to be time consuming and highly annoying. 
> Any fix or ideas would be greatly appreciated.

I'll have to find something better than this long list, but it's going to take 
a bit of time.

Meanwhile, I recommend that you locate the file data/skeletons/variant.hh, and 

        YY_ASSERT (]m4_join([ || ], m4_map_sep([_b4_type_clause], [, ], 


        YY_ASSERT (]m4_join([ || ], m4_map_sep([_b4_type_clause], [, ], 


That should eliminate the problem locally until we have something standard.

Could you please confirm it works as expected?  (don't use %define 


reply via email to

[Prev in Thread] Current Thread [Next in Thread]