Re: [PATCH 04/17] gdb: move parser output post-processing to a script

[Simon Marchi <simon.marchi@polymtl.ca>]

On 2026-09-04 20:38, Kevin Buettner wrote:
> On Fri,  4 Sep 2026 12:56:36 -0400
> simon.marchi@polymtl.ca wrote:
> 
>> diff --git a/gdb/post-process-parser-output.sh b/gdb/post-process-parser-output.sh
>> new file mode 100755
>> index 000000000000..b246e63ddff1
>> --- /dev/null
>> +++ b/gdb/post-process-parser-output.sh
>> @@ -0,0 +1,67 @@
>> +#!/bin/sh
>> +
>> +# Copyright (C) 2026 Free Software Foundation, Inc.
>> +#
>> +# This file is part of GDB.
>> +#
>> +# This program is free software; you can redistribute it and/or modify
>> +# it under the terms of the GNU General Public License as published by
>> +# the Free Software Foundation; either version 3 of the License, or
>> +# (at your option) any later version.
>> +#
>> +# This program is distributed in the hope that it will be useful,
>> +# but WITHOUT ANY WARRANTY; without even the implied warranty of
>> +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
>> +# GNU General Public License for more details.
>> +#
>> +# You should have received a copy of the GNU General Public License
>> +# along with this program.  If not, see <http://www.gnu.org/licenses/>.
>> +
>> +# Post-process a file produced by flex or bison/byacc, reading it from
>> +# standard input and writing the result to standard output.
>> +#
>> +# Usage:
>> +#
>> +#    ./post-process-parser-output.sh GENERATOR NAME < INPUT > OUTPUT
>> +#
>> +# Where GENERATOR is "bison" (regardless of whether we're using the actual
>> +# bison or byacc) or "flex", and NAME is the name of the parser (e.g.
>> +# "cp-name-parser"), used to make the global symbols names bison produces
>> +# unique.
>> +
>> +set -e
>> +
>> +generator="$1"
>> +
>> +# Convert e.g. "cp-name-parser" to "cp_name_parser".
>> +name=$(echo "$2" | sed -e 's/-/_/g')
>> +
>> +common='
>> +/extern.*malloc/d
>> +/extern.*realloc/d
>> +/extern.*free/d
>> +/include.*malloc.h/d
>> +s/\([^x]\)malloc/\1xmalloc/g
>> +s/\([^x]\)realloc/\1xrealloc/g
>> +s/\([ \t;,(]\)free\([ \t]*[&(),]\)/\1xfree\2/g
>> +s/\([ \t;,(]\)free$/\1xfree/g
>> +'
>> +
>> +case "$generator" in
>> +    bison)
>> +	sed -e "$common" \
>> +	    -e '/^#line.*y.tab.c/d' \
>> +	    -e 's/YY_NULL/YY_NULLPTR/g' \
> 
> Another (pre-existing) nit found by AI code review:
> 
>   post-process-parser-output.sh:53: s/YY_NULL/YY_NULLPTR/g also
>   matches inside YY_NULLPTR, producing YY_NULLPTRPTR; harmless
>   (guarded by #ifndef), faithfully carried over from the old Makefile
>   sed.  Opportunity to fix with word boundaries.

Thanks, that made me look into why we do this substitution at all, and
it was become at some point before 2013, bison in C++11 mode would
output:

#define YY_NULL nullptr

while flex would output

#define YY_NULL 0

To avoid the clash, we would  rename the bison one.  It's been fixed in
bison 3.0.2, where it was renamed YY_NULLPTR.  I think it's safe to
assume that any GDB dev today will not use a 13 year old bison, so I
will just remove it.

Simon