Warnings in GCC 16.1.1

[ablaette]

Using a Docker container with Fedora 44 and GCC, I am able to reproduce a warning issued by CRAN:

macro.c: In function ‘macro_iterator_seek’:
macro.c:1099:32: warning: assignment discards ‘const’ qualifier from pointer target type [-Wdiscarded-qualifiers]
1099 | if (NULL != (first_paren = strchr(prefix, '[')))
| ^
gcc -c -o parser.tab.o -O2 -Wall -fPIC -DR_PACKAGE -DNDEBUG -DUSE_TERMCAP -DUSE_READLINE -DCOMPILE_DATE=""Thu Aug 13 22:45:10 UTC 2026"" -DCWB_VERSION="3.4.33" -I/usr/include/glib-2.0 -I/usr/lib64/glib-2.0/include -I/usr/include/sysprof-6 -pthread parser.tab.c
parser.tab.c:1649:1: warning: ‘yy_symbol_print’ defined but not used [-Wunused-function]
1649 | || defined __cplusplus || defined _MSC_VER)
| ^ ~~~~~~~~~~
parser.tab.c:832:27: warning: ‘yyrhs’ defined but not used [-Wunused-const-variable=]
832 |
| ^
parser.tab.c:791:28: warning: ‘yyprhs’ defined but not used [-Wunused-const-variable=]
791 | /* YYPRHS[YYN] -- Index of the first RHS symbol of rule number YYN in
| ^~~~~~

[ablaette]

This is what Claude tells me:

This is GCC catching a real type-safety issue, not just being pedantic.

Why it happens: GCC's built-in prototypes for functions like strchr, strstr, strrchr, strpbrk, and memchr are const-correct — if you pass a const char * in, GCC treats the return value as const char * too, matching C++ overload behavior. This is meant to catch exactly the situation you have: prefix is declared const char *, so strchr(prefix, '[') semantically returns a pointer into that const data, but you're assigning it to first_paren, which is a plain (non-const) char *. That would let you write through first_paren into memory the compiler was told is const — undefined behavior, or a segfault if prefix actually points into read-only memory (e.g. a string literal).

Fix depends on what you do with first_paren afterward:

If you only read through first_paren (e.g. just checking where the bracket is, not modifying the string) — make it const too:
c
const char *first_paren;
...
if (NULL != (first_paren = strchr(prefix, '[')))
If you need to write through first_paren (e.g. the common pattern of null-terminating the string at that point) — then prefix shouldn't really be const in the first place, or you need a modifiable copy:
c
char prefix_copy = strdup(prefix); / check for NULL */
...
char *first_paren = strchr(prefix_copy, '[');
if (first_paren != NULL) *first_paren = '\0';
...
free(prefix_copy);

Or, if prefix is genuinely mutable data that was just over-declared as const somewhere upstream, remove the const on prefix's declaration instead.

Which of these fits — are you just locating the bracket, or modifying the string at that position?