help-bison
[Top][All Lists]
Advanced

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

Re: Enable experimental report features?


From: Hans Aberg
Subject: Re: Enable experimental report features?
Date: Mon, 6 May 2002 11:22:21 +0200

At 10:02 +0200 2002/05/06, Akim Demaille wrote:
>A pity that we don't have .output browser: Hans would have said that
>he has a check box to display or not to display the full item set :)

Not really: Apart from that I decided it not worth the effort to develop
check GUI boxes for options (even though it is not that difficult), in my
IDE, it is not possible to make that file specific (when using several .y
compiles in the same project). So I added a special .ym file containing the
startup arguments.

There are however Bison plugins available that make use of check-boxes for
startup options. (But if you move towards the GUI question, you will have
to think carefully to not impose restrictions on the usage.)

>Now, there remains the question of the lookaheads.
>
>For instance, still on G++, here is without, and then with.
>
>state 44
   ...

>state 44
>
>    .hush_warning  ->  .  [error, IDENTIFIER, TYPENAME, SELFNAME,
>PFUNCNAME, SCS
>PEC, TYPESPEC, CV_QUALIFIER, ENUM, TYPEOF, SIGOF, ATTRIBUTE, AGGR,
>OPERATOR, TYP
>ENAME_KEYWORD, TEMPLATE, SCOPE, PTYPENAME, NSNAME, ';', '&', '*', '~',
>'(']   (r
>ule 8)

Again, I have never used that feature, and would therefore not miss it (as
I learned to live without it). But again, it might be useful for
pedagogical reasons on say the Bison Help list. So it should perhaps be an
option that is normally turned off. One can then ask those that need it
when constructing a grammar to turn it on.

But should not the feature be displayed (or something)?
      $default [error, IDENTIFIER, TYPENAME, SELFNAME, PFUNCNAME, SCSPEC,
TYPESPEC, CV_QUALIFIER, ENUM, TYPEOF, SIGOF, ATTRIBUTE, AGGR, OPERATOR,
TYPENAME_KEYWORD, TEMPLATE, SCOPE, PTYPENAME, NSNAME, ';', '&', '*', '~',
'(']
         reduce using rule 8 (.hush_warning)
-- The default must always just at most one rule or a shift plus a possible
error.

  Hans Aberg





reply via email to

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