Re: Word For Today: “Uglification”

Tim Rentsch <tr.17687@z991.linuxsc.com> writes:

Kaz Kylheku <433-929-6894@kylheku.com> writes:
>
[some editing of white space done]
>

On 2024-03-12, Lawrence D'Oliveiro <ldo@nz.invalid> wrote:
>

From /usr/include/<<arch>>/bits/select.h on my Debian system:
>
  #define __FD_ZERO(s)                                                  \
    do {                                                                \
      unsigned int __i;                                                 \
      fd_set *__arr = (s);                                              \

>
This assignment has value;  it checks that, loosely speaking,
s is an "assignment compatible" pointer with a fd_set *,
so that there is a diagnostic if the macro is applied to
an object of the wrong type.

>
More to the point, if the macro is applied to a value of the wrong
type.
>

      for (__i = 0; __i < sizeof (fd_set) / sizeof (__fd_mask); ++__i)  \
        __FDS_BITS (__arr)[__i] = 0;                                    \

>
Here, I would have done memset(__arr, 0, sizeof *__arr).

>
That assumes that it is the entire fd_set that needs to be zeroed,
which may not be right.  Note the call to the __FDS_BITS() macro.
>
Better:
>
  #define __FD_ZERO(s) (                                                  \
    (void) memset(                                                        \
      __FDS_BITS( (fd_set*){(s)} ), 0, sizeof __FDS_BITS( (fd_set*){0} )  \
    )                                                                     \
  )
>
This definition:  avoids introducing any new identifiers;  checks
that the argument s yields an assignment compatible pointer;  and
provides a macro that can be used as a void expression (unlike the
original macro definition, which can be used only as a statement).

For context, here's the entire file from my system (Ubuntu 24.0.4,
package libc6-dev:amd64 2.35-0ubuntu3.6).  I get the impression that the
author(s) decided not to use memset to avoid the required #include,
which might increase compilation times for code that indirectly includes
this header.  (I offer no opinion on whether that's a good tradeoff.)

Note that __FD_ZERO is very clearly *not* intended to be invoked by
arbitrary code.

```
/* Copyright (C) 1997-2022 Free Software Foundation, Inc.
   This file is part of the GNU C Library.

   The GNU C Library is free software; you can redistribute it and/or
   modify it under the terms of the GNU Lesser General Public
   License as published by the Free Software Foundation; either
   version 2.1 of the License, or (at your option) any later version.

   The GNU C Library 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
   Lesser General Public License for more details.

   You should have received a copy of the GNU Lesser General Public
   License along with the GNU C Library; if not, see
   <https://www.gnu.org/licenses/>.  */

#ifndef _SYS_SELECT_H
# error "Never use <bits/select.h> directly; include <sys/select.h> instead."
#endif


/* We don't use `memset' because this would require a prototype and
   the array isn't too big.  */
#define __FD_ZERO(s) \
  do {       \
    unsigned int __i;       \
    fd_set *__arr = (s);       \
    for (__i = 0; __i < sizeof (fd_set) / sizeof (__fd_mask); ++__i)       \
      __FDS_BITS (__arr)[__i] = 0;       \
  } while (0)
#define __FD_SET(d, s) \
  ((void) (__FDS_BITS (s)[__FD_ELT(d)] |= __FD_MASK(d)))
#define __FD_CLR(d, s) \
  ((void) (__FDS_BITS (s)[__FD_ELT(d)] &= ~__FD_MASK(d)))
#define __FD_ISSET(d, s) \
  ((__FDS_BITS (s)[__FD_ELT (d)] & __FD_MASK (d)) != 0)

```

--
Keith Thompson (The_Other_Keith) Keith.S.Thompson+u@gmail.com
Working, but not speaking, for Medtronic
void Void(void) { Void(); } /* The recursive call of the void */