This reverts this change:

  Author:     Andreas Schwab <schwab@linux-m68k.org>
  AuthorDate: Sun Feb 8 12:34:02 2026 +0100

    Extend emacs server protocol for empty arguments

    An empty argument is represented by &0.  On the receiving side, &0 is
    replaced by nothing.

    * lisp/server.el (server-unquote-arg): Replace "&0" by nothing.
    (server-quote-arg): Produce "&0" for an empty string.
    * lib-src/emacsclient.c (quote_argument): Produce "&0" for an
    empty string.
    (unquote_argument): Replace "&0" by nothing.   (Bug#80356)

The bug in question was already fixed by this change:

  Author:     Sean Whitton <spwhitton@spwhitton.name>
  AuthorDate: Fri Nov 7 12:33:21 2025 +0000

    Don't discard empty string arguments from emacsclient

    * lisp/server.el (server--process-filter-1): Don't discard empty
    string arguments from emacsclient.
    (server-eval-args-left):
    * doc/emacs/misc.texi (emacsclient Options):
    * etc/NEWS: Document the change.

An empty argument is represented by &0.  On the receiving side, &0 is
replaced by nothing.

* lisp/server.el (server-unquote-arg): Replace "&0" by nothing.
(server-quote-arg): Produce "&0" for an empty string.
* lib-src/emacsclient.c (quote_argument): Produce "&0" for an
empty string.
(unquote_argument): Replace "&0" by nothing.   (Bug#80356)

* lib-src/make-docfile.c (main, add_global, compare_globals):
Prefer streq to strcmp where either will do.

* lib-src/make-docfile.c (scan_c_stream): Allow nested parens.
E.g., ‘(Lisp_Object XXX (force))’ for the identifier ‘force’,
where XXX is a macro.  This is for future changes.

* lib-src/make-docfile.c (write_c_args):
Don’t mishandle the identifiers ‘v’, ‘vo’, and ‘voi’.

* lib-src/make-docfile.c (write_c_args): Don’t diagnose “()”.
In C23 it’s equivalent to (void), and it works fine (albeit with
less type-checking) in older C.  Omit first arg, which is no
longer needed.  Caller changed.

This bit me when I ran ‘make’ with typos in the Emacs source.
* lib-src/make-docfile.c (struct rcsoc_state.buf_lim):
New member.
(read_c_string_or_comment): Initialize it.
(put_char, scan_c_stream): Check for buffer overflow.
(scan_c_stream): Output a diagnostic rather than aborting
when the buffer overflows.

Problem for mpz_get_d_rounded reported by Helmut Eller in:
https://lists.gnu.org/r/emacs-devel/2025-11/msg00795.html
* lib-src/make-docfile.c (DEFUN_pure): New constant.
(write_globals, scan_c_stream): Support "attributes: pure".
* src/bignum.h (mpz_get_d_rounded):
* src/data.c (Fsymbolp, Fmodule_function_p, Fintegerp, Fnumberp):
* src/lisp.h (bignum_to_double, bignum_to_intmax)
(bignum_to_uintmax, bignum_bufsize):
Now pure, not const, since they depend on current state.
For example, Fsymbolp now inspects symbols_with_pos_enabled,
and the bignum functions inspect bignum contents in memory.
* src/data.c (Feq):
* src/xfaces.c (Fface_attribute_relative_p):
No longer const, since they might abort when debugging.
* src/pdumper.h (pdumper_object_p, pdumper_cold_object_p)
(pdumper_find_object_type, pdumper_object_p_precise):
These are not const functions.  But there is no need to declare
them to be pure, either, as they’re inline so the compiler can
figure it out.

* admin/merge-gnulib (GNULIB_MODULES): Add stringeq.
With current Gnulib it is already present as in indirect dependency;
listing it here because Emacs now depends on it directly.
* lib-src/ebrowse.c, lib-src/etags.c:
(streq): Remove, as Gnulib defines this now.
* lib/fseterr.c, lib/fseterr.h, lib/issymlink.c, lib/issymlink.h:
* lib/issymlinkat.c, lib/stdio-consolesafe.c, lib/string.c:
* m4/fseterr.m4, m4/gettext_h.m4, m4/stringeq.m4:
New files from Gnulib.
* src/conf_post.h (tzfree) [__ANDROID_API__ >= 35]: Remove.