Adding upstream version 1.21.4.upstream/1.21.4

Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>

1 files changed, 1108 insertions, 0 deletions

diff --git a/lib/spawni.c b/lib/spawni.c
new file mode 100644
index 0000000..cc9511f
--- /dev/null
+++ b/lib/spawni.c
@@ -0,0 +1,1108 @@
+/* Guts of POSIX spawn interface.  Generic POSIX.1 version.
+   Copyright (C) 2000-2006, 2008-2023 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   This file 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.
+
+   This file 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 this program.  If not, see <https://www.gnu.org/licenses/>.  */
+
+#include <config.h>
+
+/* Specification.  */
+#include <spawn.h>
+#include "spawn_int.h"
+
+#include <alloca.h>
+#include <errno.h>
+
+#include <fcntl.h>
+#ifndef O_LARGEFILE
+# define O_LARGEFILE 0
+#endif
+
+#if _LIBC || HAVE_PATHS_H
+# include <paths.h>
+#else
+# define _PATH_BSHELL BOURNE_SHELL
+#endif
+
+#include <signal.h>
+#include <stdlib.h>
+#include <string.h>
+#include <unistd.h>
+
+#if _LIBC
+# include <not-cancel.h>
+#else
+# define close_not_cancel close
+# define open_not_cancel open
+#endif
+
+#if _LIBC
+# include <local-setxid.h>
+#else
+# if !HAVE_SETEUID
+#  define seteuid(id) setresuid (-1, id, -1)
+# endif
+# if !HAVE_SETEGID
+#  define setegid(id) setresgid (-1, id, -1)
+# endif
+# define local_seteuid(id) seteuid (id)
+# define local_setegid(id) setegid (id)
+#endif
+
+#if _LIBC
+# define alloca __alloca
+# define execve __execve
+# define dup2 __dup2
+# define fork __fork
+# define getgid __getgid
+# define getuid __getuid
+# define sched_setparam __sched_setparam
+# define sched_setscheduler __sched_setscheduler
+# define setpgid __setpgid
+# define sigaction __sigaction
+# define sigismember __sigismember
+# define sigprocmask __sigprocmask
+# define strchrnul __strchrnul
+# define vfork __vfork
+#endif
+
+
+/* The Unix standard contains a long explanation of the way to signal
+   an error after the fork() was successful.  Since no new wait status
+   was wanted there is no way to signal an error using one of the
+   available methods.  The committee chose to signal an error by a
+   normal program exit with the exit code 127.  */
+#define SPAWN_ERROR     127
+
+
+#if defined _WIN32 && ! defined __CYGWIN__
+/* Native Windows API.  */
+
+/* Define to 1 to enable DuplicateHandle optimization.
+   Define to 0 to disable this optimization.  */
+# ifndef SPAWN_INTERNAL_OPTIMIZE_DUPLICATEHANDLE
+#  define SPAWN_INTERNAL_OPTIMIZE_DUPLICATEHANDLE 1
+# endif
+
+/* Get declarations of the native Windows API functions.  */
+# define WIN32_LEAN_AND_MEAN
+# include <windows.h>
+
+# include <stdio.h>
+
+# include "filename.h"
+# include "concat-filename.h"
+# include "findprog.h"
+# include "malloca.h"
+# include "windows-spawn.h"
+
+/* Don't assume that UNICODE is not defined.  */
+# undef CreateFile
+# define CreateFile CreateFileA
+# undef STARTUPINFO
+# define STARTUPINFO STARTUPINFOA
+# undef CreateProcess
+# define CreateProcess CreateProcessA
+
+/* Grows inh_handles->count so that it becomes > newfd.
+   Returns 0 upon success.  In case of failure, -1 is returned, with errno set.
+ */
+static int
+grow_inheritable_handles (struct inheritable_handles *inh_handles, int newfd)
+{
+  if (inh_handles->allocated <= newfd)
+    {
+      size_t new_allocated = 2 * inh_handles->allocated + 1;
+      if (new_allocated <= newfd)
+        new_allocated = newfd + 1;
+      struct IHANDLE *new_ih =
+        (struct IHANDLE *)
+        realloc (inh_handles->ih, new_allocated * sizeof (struct IHANDLE));
+      if (new_ih == NULL)
+        {
+          errno = ENOMEM;
+          return -1;
+        }
+      inh_handles->allocated = new_allocated;
+      inh_handles->ih = new_ih;
+    }
+
+  struct IHANDLE *ih = inh_handles->ih;
+
+  for (; inh_handles->count <= newfd; inh_handles->count++)
+    ih[inh_handles->count].handle = INVALID_HANDLE_VALUE;
+
+  return 0;
+}
+
+# if SPAWN_INTERNAL_OPTIMIZE_DUPLICATEHANDLE
+
+/* Assuming inh_handles->ih[newfd].handle != INVALID_HANDLE_VALUE
+   and      (inh_handles->ih[newfd].flags & DELAYED_DUP2_NEWFD) != 0,
+   actually performs the delayed dup2 (oldfd, newfd).
+   Returns 0 upon success.  In case of failure, -1 is returned, with errno set.
+ */
+static int
+do_delayed_dup2 (int newfd, struct inheritable_handles *inh_handles,
+                 HANDLE curr_process)
+{
+  int oldfd = inh_handles->ih[newfd].linked_fd;
+  /* Check invariants.  */
+  if (!((inh_handles->ih[oldfd].flags & DELAYED_DUP2_OLDFD) != 0
+        && newfd == inh_handles->ih[oldfd].linked_fd
+        && inh_handles->ih[newfd].handle == inh_handles->ih[oldfd].handle))
+    abort ();
+  /* Duplicate the handle now.  */
+  if (!DuplicateHandle (curr_process, inh_handles->ih[oldfd].handle,
+                        curr_process, &inh_handles->ih[newfd].handle,
+                        0, TRUE, DUPLICATE_SAME_ACCESS))
+    {
+      errno = EBADF; /* arbitrary */
+      return -1;
+    }
+  inh_handles->ih[oldfd].flags &= ~DELAYED_DUP2_OLDFD;
+  inh_handles->ih[newfd].flags =
+    (unsigned char) inh_handles->ih[oldfd].flags | KEEP_OPEN_IN_CHILD;
+  return 0;
+}
+
+/* Performs the remaining delayed dup2 (oldfd, newfd).
+   Returns 0 upon success.  In case of failure, -1 is returned, with errno set.
+ */
+static int
+do_remaining_delayed_dup2 (struct inheritable_handles *inh_handles,
+                           HANDLE curr_process)
+{
+  size_t handles_count = inh_handles->count;
+  int newfd;
+
+  for (newfd = 0; newfd < handles_count; newfd++)
+    if (inh_handles->ih[newfd].handle != INVALID_HANDLE_VALUE
+        && (inh_handles->ih[newfd].flags & DELAYED_DUP2_NEWFD) != 0)
+      if (do_delayed_dup2 (newfd, inh_handles, curr_process) < 0)
+        return -1;
+  return 0;
+}
+
+# endif
+
+/* Closes the handles in inh_handles that are not meant to be preserved in the
+   child process, and reduces inh_handles->count to the minimum needed.  */
+static void
+shrink_inheritable_handles (struct inheritable_handles *inh_handles)
+{
+  struct IHANDLE *ih = inh_handles->ih;
+  size_t handles_count = inh_handles->count;
+  unsigned int fd;
+
+  for (fd = 0; fd < handles_count; fd++)
+    {
+      HANDLE handle = ih[fd].handle;
+
+      if (handle != INVALID_HANDLE_VALUE
+          && (ih[fd].flags & KEEP_OPEN_IN_CHILD) == 0)
+        {
+          if (!(ih[fd].flags & KEEP_OPEN_IN_PARENT))
+            CloseHandle (handle);
+          ih[fd].handle = INVALID_HANDLE_VALUE;
+        }
+    }
+
+  while (handles_count > 3
+         && ih[handles_count - 1].handle == INVALID_HANDLE_VALUE)
+    handles_count--;
+
+  inh_handles->count = handles_count;
+}
+
+/* Closes all handles in inh_handles.  */
+static void
+close_inheritable_handles (struct inheritable_handles *inh_handles)
+{
+  struct IHANDLE *ih = inh_handles->ih;
+  size_t handles_count = inh_handles->count;
+  unsigned int fd;
+
+  for (fd = 0; fd < handles_count; fd++)
+    {
+      HANDLE handle = ih[fd].handle;
+
+      if (handle != INVALID_HANDLE_VALUE
+          && !(ih[fd].flags & DELAYED_DUP2_NEWFD)
+          && !(ih[fd].flags & KEEP_OPEN_IN_PARENT))
+        CloseHandle (handle);
+    }
+}
+
+/* Tests whether a memory region, starting at P and N bytes long, contains only
+   zeroes.  */
+static bool
+memiszero (const void *p, size_t n)
+{
+  const char *cp = p;
+  for (; n > 0; cp++, n--)
+    if (*cp != 0)
+      return 0;
+  return 1;
+}
+
+/* Tests whether *S contains no signals.  */
+static bool
+sigisempty (const sigset_t *s)
+{
+  return memiszero (s, sizeof (sigset_t));
+}
+
+/* Executes a 'close' action.
+   Returns 0 upon success.  In case of failure, -1 is returned, with errno set.
+ */
+static int
+do_close (struct inheritable_handles *inh_handles, int fd, bool ignore_EBADF)
+{
+  if (!(fd >= 0 && fd < inh_handles->count
+        && inh_handles->ih[fd].handle != INVALID_HANDLE_VALUE))
+    {
+      if (ignore_EBADF)
+        return 0;
+      else
+        {
+          errno = EBADF;
+          return -1;
+        }
+    }
+
+# if SPAWN_INTERNAL_OPTIMIZE_DUPLICATEHANDLE
+  if ((inh_handles->ih[fd].flags & DELAYED_DUP2_NEWFD) != 0)
+    {
+      int dup2_oldfd = inh_handles->ih[fd].linked_fd;
+      /* Check invariants.  */
+      if (!((inh_handles->ih[dup2_oldfd].flags & DELAYED_DUP2_OLDFD) != 0
+            && fd == inh_handles->ih[dup2_oldfd].linked_fd
+            && inh_handles->ih[fd].handle == inh_handles->ih[dup2_oldfd].handle))
+        abort ();
+      /* Annihilate a delayed dup2 (..., fd) call.  */
+      inh_handles->ih[dup2_oldfd].flags &= ~DELAYED_DUP2_OLDFD;
+    }
+  else if ((inh_handles->ih[fd].flags & DELAYED_DUP2_OLDFD) != 0)
+    {
+      int dup2_newfd = inh_handles->ih[fd].linked_fd;
+      /* Check invariants.  */
+      if (!((inh_handles->ih[dup2_newfd].flags & DELAYED_DUP2_NEWFD) != 0
+            && fd == inh_handles->ih[dup2_newfd].linked_fd
+            && inh_handles->ih[fd].handle == inh_handles->ih[dup2_newfd].handle))
+        abort ();
+      /* Optimize a delayed dup2 (fd, ...) call.  */
+      inh_handles->ih[dup2_newfd].flags =
+        (inh_handles->ih[fd].flags & ~DELAYED_DUP2_OLDFD) | KEEP_OPEN_IN_CHILD;
+    }
+  else
+# endif
+    {
+      if (!(inh_handles->ih[fd].flags & KEEP_OPEN_IN_PARENT)
+          && !CloseHandle (inh_handles->ih[fd].handle))
+        {
+          inh_handles->ih[fd].handle = INVALID_HANDLE_VALUE;
+          errno = EIO;
+          return -1;
+        }
+    }
+  inh_handles->ih[fd].handle = INVALID_HANDLE_VALUE;
+
+  return 0;
+}
+
+/* Opens an inheritable HANDLE to a file.
+   Upon failure, returns INVALID_HANDLE_VALUE with errno set.  */
+static HANDLE
+open_handle (const char *name, int flags, mode_t mode)
+{
+  /* To ease portability.  Like in open.c.  */
+  if (strcmp (name, "/dev/null") == 0)
+    name = "NUL";
+
+  /* POSIX <https://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap04.html#tag_04_13>
+     specifies: "More than two leading <slash> characters shall be treated as
+     a single <slash> character."  */
+  if (ISSLASH (name[0]) && ISSLASH (name[1]) && ISSLASH (name[2]))
+    {
+      name += 2;
+      while (ISSLASH (name[1]))
+        name++;
+    }
+
+  size_t len = strlen (name);
+  size_t drive_prefix_len = (HAS_DEVICE (name) ? 2 : 0);
+
+  /* Remove trailing slashes (except the very first one, at position
+     drive_prefix_len), but remember their presence.  */
+  size_t rlen;
+  bool check_dir = false;
+
+  rlen = len;
+  while (rlen > drive_prefix_len && ISSLASH (name[rlen-1]))
+    {
+      check_dir = true;
+      if (rlen == drive_prefix_len + 1)
+        break;
+      rlen--;
+    }
+
+  /* Handle '' and 'C:'.  */
+  if (!check_dir && rlen == drive_prefix_len)
+    {
+      errno = ENOENT;
+      return INVALID_HANDLE_VALUE;
+    }
+
+  /* Handle '\\'.  */
+  if (rlen == 1 && ISSLASH (name[0]) && len >= 2)
+    {
+      errno = ENOENT;
+      return INVALID_HANDLE_VALUE;
+    }
+
+  const char *rname;
+  char *malloca_rname;
+  if (rlen == len)
+    {
+      rname = name;
+      malloca_rname = NULL;
+    }
+  else
+    {
+      malloca_rname = malloca (rlen + 1);
+      if (malloca_rname == NULL)
+        {
+          errno = ENOMEM;
+          return INVALID_HANDLE_VALUE;
+        }
+      memcpy (malloca_rname, name, rlen);
+      malloca_rname[rlen] = '\0';
+      rname = malloca_rname;
+    }
+
+  /* For the meaning of the flags, see
+     <https://docs.microsoft.com/en-us/cpp/c-runtime-library/reference/open-wopen>  */
+  /* Open a handle to the file.
+     CreateFile
+     <https://docs.microsoft.com/en-us/windows/desktop/api/fileapi/nf-fileapi-createfilea>
+     <https://docs.microsoft.com/en-us/windows/desktop/FileIO/creating-and-opening-files>  */
+  SECURITY_ATTRIBUTES sec_attr;
+  sec_attr.nLength = sizeof (SECURITY_ATTRIBUTES);
+  sec_attr.lpSecurityDescriptor = NULL;
+  sec_attr.bInheritHandle = TRUE;
+  HANDLE handle =
+    CreateFile (rname,
+                ((flags & (O_WRONLY | O_RDWR)) != 0
+                 ? GENERIC_READ | GENERIC_WRITE
+                 : GENERIC_READ),
+                FILE_SHARE_READ | FILE_SHARE_WRITE | FILE_SHARE_DELETE,
+                &sec_attr,
+                ((flags & O_CREAT) != 0
+                 ? ((flags & O_EXCL) != 0
+                    ? CREATE_NEW
+                    : ((flags & O_TRUNC) != 0 ? CREATE_ALWAYS : OPEN_ALWAYS))
+                 : ((flags & O_TRUNC) != 0
+                    ? TRUNCATE_EXISTING
+                    : OPEN_EXISTING)),
+                /* FILE_FLAG_BACKUP_SEMANTICS is useful for opening directories,
+                   which is out-of-scope here.  */
+                /* FILE_FLAG_POSIX_SEMANTICS (treat file names that differ only
+                   in case as different) makes sense only when applied to *all*
+                   filesystem operations.  */
+                /* FILE_FLAG_BACKUP_SEMANTICS | FILE_FLAG_POSIX_SEMANTICS */
+                FILE_ATTRIBUTE_NORMAL
+                | ((flags & O_TEMPORARY) != 0 ? FILE_FLAG_DELETE_ON_CLOSE : 0)
+                | ((flags & O_SEQUENTIAL ) != 0 ? FILE_FLAG_SEQUENTIAL_SCAN : 0)
+                | ((flags & O_RANDOM) != 0 ? FILE_FLAG_RANDOM_ACCESS : 0),
+                NULL);
+  if (handle == INVALID_HANDLE_VALUE)
+    switch (GetLastError ())
+      {
+      /* Some of these errors probably cannot happen with the specific flags
+         that we pass to CreateFile.  But who knows...  */
+      case ERROR_FILE_NOT_FOUND: /* The last component of rname does not exist.  */
+      case ERROR_PATH_NOT_FOUND: /* Some directory component in rname does not exist.  */
+      case ERROR_BAD_PATHNAME:   /* rname is such as '\\server'.  */
+      case ERROR_BAD_NETPATH:    /* rname is such as '\\nonexistentserver\share'.  */
+      case ERROR_BAD_NET_NAME:   /* rname is such as '\\server\nonexistentshare'.  */
+      case ERROR_INVALID_NAME:   /* rname contains wildcards, misplaced colon, etc.  */
+      case ERROR_DIRECTORY:
+        errno = ENOENT;
+        break;
+
+      case ERROR_ACCESS_DENIED:  /* rname is such as 'C:\System Volume Information\foo'.  */
+      case ERROR_SHARING_VIOLATION: /* rname is such as 'C:\pagefile.sys'.  */
+                                    /* XXX map to EACCES or EPERM? */
+        errno = EACCES;
+        break;
+
+      case ERROR_OUTOFMEMORY:
+        errno = ENOMEM;
+        break;
+
+      case ERROR_WRITE_PROTECT:
+        errno = EROFS;
+        break;
+
+      case ERROR_WRITE_FAULT:
+      case ERROR_READ_FAULT:
+      case ERROR_GEN_FAILURE:
+        errno = EIO;
+        break;
+
+      case ERROR_BUFFER_OVERFLOW:
+      case ERROR_FILENAME_EXCED_RANGE:
+        errno = ENAMETOOLONG;
+        break;
+
+      case ERROR_DELETE_PENDING: /* XXX map to EACCES or EPERM? */
+        errno = EPERM;
+        break;
+
+      default:
+        errno = EINVAL;
+        break;
+      }
+
+  if (malloca_rname != NULL)
+    {
+      int saved_errno = errno;
+      freea (malloca_rname);
+      errno = saved_errno;
+    }
+  return handle;
+}
+
+/* Executes an 'open' action.
+   Returns 0 upon success.  In case of failure, -1 is returned, with errno set.
+ */
+static int
+do_open (struct inheritable_handles *inh_handles, int newfd,
+         const char *filename, const char *directory,
+         int flags, mode_t mode)
+{
+  if (!(newfd >= 0 && newfd < _getmaxstdio ()))
+    {
+      errno = EBADF;
+      return -1;
+    }
+  if (grow_inheritable_handles (inh_handles, newfd) < 0)
+    return -1;
+  if (do_close (inh_handles, newfd, true) < 0)
+    return -1;
+  if (filename == NULL)
+    {
+      errno = EINVAL;
+      return -1;
+    }
+  char *filename_to_free = NULL;
+  if (directory != NULL && IS_RELATIVE_FILE_NAME (filename))
+    {
+      char *real_filename = concatenated_filename (directory, filename, NULL);
+      if (real_filename == NULL)
+        {
+          errno = ENOMEM;
+          return -1;
+        }
+      filename = real_filename;
+      filename_to_free = real_filename;
+    }
+  HANDLE handle = open_handle (filename, flags, mode);
+  if (handle == INVALID_HANDLE_VALUE)
+    {
+      free (filename_to_free);
+      return -1;
+    }
+  free (filename_to_free);
+  inh_handles->ih[newfd].handle = handle;
+  inh_handles->ih[newfd].flags =
+    ((flags & O_APPEND) != 0 ? 32 : 0) | KEEP_OPEN_IN_CHILD;
+  return 0;
+}
+
+/* Executes a 'dup2' action.
+   Returns 0 upon success.  In case of failure, -1 is returned, with errno set.
+ */
+static int
+do_dup2 (struct inheritable_handles *inh_handles, int oldfd, int newfd,
+         HANDLE curr_process)
+{
+  if (!(oldfd >= 0 && oldfd < inh_handles->count
+        && inh_handles->ih[oldfd].handle != INVALID_HANDLE_VALUE))
+    {
+      errno = EBADF;
+      return -1;
+    }
+  if (!(newfd >= 0 && newfd < _getmaxstdio ()))
+    {
+      errno = EBADF;
+      return -1;
+    }
+  if (newfd != oldfd)
+    {
+      if (grow_inheritable_handles (inh_handles, newfd) < 0)
+        return -1;
+      if (do_close (inh_handles, newfd, true) < 0)
+        return -1;
+      /* We may need to duplicate the handle, so that a forthcoming do_close
+         action on oldfd has no effect on newfd.  */
+# if SPAWN_INTERNAL_OPTIMIZE_DUPLICATEHANDLE
+      /* But try to not do it now; delay it if possible.  In many cases, the
+         DuplicateHandle call can be optimized away.  */
+      if ((inh_handles->ih[oldfd].flags & DELAYED_DUP2_NEWFD) != 0)
+        if (do_delayed_dup2 (oldfd, inh_handles, curr_process) < 0)
+          return -1;
+      if ((inh_handles->ih[oldfd].flags & DELAYED_DUP2_OLDFD) != 0)
+        {
+          /* Check invariants.  */
+          int dup2_newfd = inh_handles->ih[oldfd].linked_fd;
+          if (!((inh_handles->ih[dup2_newfd].flags & DELAYED_DUP2_NEWFD) != 0
+                && oldfd == inh_handles->ih[dup2_newfd].linked_fd
+                && inh_handles->ih[oldfd].handle == inh_handles->ih[dup2_newfd].handle))
+            abort ();
+          /* We can't delay two or more dup2 calls from the same oldfd.  */
+          if (!DuplicateHandle (curr_process, inh_handles->ih[oldfd].handle,
+                                curr_process, &inh_handles->ih[newfd].handle,
+                                0, TRUE, DUPLICATE_SAME_ACCESS))
+            {
+              errno = EBADF; /* arbitrary */
+              return -1;
+            }
+          inh_handles->ih[newfd].flags =
+            (unsigned char) inh_handles->ih[oldfd].flags | KEEP_OPEN_IN_CHILD;
+        }
+      else
+        {
+          /* Delay the dup2 (oldfd, newfd) action.  */
+          inh_handles->ih[oldfd].flags |= DELAYED_DUP2_OLDFD;
+          inh_handles->ih[oldfd].linked_fd = newfd;
+          inh_handles->ih[newfd].handle = inh_handles->ih[oldfd].handle;
+          inh_handles->ih[newfd].flags = DELAYED_DUP2_NEWFD;
+          inh_handles->ih[newfd].linked_fd = oldfd;
+        }
+# else
+      if (!DuplicateHandle (curr_process, inh_handles->ih[oldfd].handle,
+                            curr_process, &inh_handles->ih[newfd].handle,
+                            0, TRUE, DUPLICATE_SAME_ACCESS))
+        {
+          errno = EBADF; /* arbitrary */
+          return -1;
+        }
+      inh_handles->ih[newfd].flags =
+        (unsigned char) inh_handles->ih[oldfd].flags | KEEP_OPEN_IN_CHILD;
+# endif
+    }
+  return 0;
+}
+
+int
+__spawni (pid_t *pid, const char *prog_filename,
+          const posix_spawn_file_actions_t *file_actions,
+          const posix_spawnattr_t *attrp, const char *const prog_argv[],
+          const char *const envp[], int use_path)
+{
+  /* Validate the arguments.  */
+  if (prog_filename == NULL
+      || (attrp != NULL
+          && ((attrp->_flags & ~POSIX_SPAWN_SETPGROUP) != 0
+              || attrp->_pgrp != 0
+              || ! sigisempty (&attrp->_sd)
+              || ! sigisempty (&attrp->_ss)
+              || attrp->_sp.sched_priority != 0
+              || attrp->_policy != 0)))
+    return EINVAL;
+
+  /* Process group handling:
+     Native Windows does not have the concept of process group, but it has the
+     concept of a console attached to a process.
+     So, we interpret the three cases as follows:
+       - Flag POSIX_SPAWN_SETPGROUP not set: Means, the child process is in the
+         same process group as the parent process.  We interpret this as a
+         request to reuse the same console.
+       - Flag POSIX_SPAWN_SETPGROUP set with attrp->_pgrp == 0: Means the child
+         process starts a process group of its own.  See
+         <https://pubs.opengroup.org/onlinepubs/9699919799/functions/posix_spawnattr_getpgroup.html>
+         <https://pubs.opengroup.org/onlinepubs/9699919799/functions/setpgrp.html>
+         We interpret this as a request to detach from the current console.
+       - Flag POSIX_SPAWN_SETPGROUP set with attrp->_pgrp != 0: Means the child
+         process joins another, existing process group.  See
+         <https://pubs.opengroup.org/onlinepubs/9699919799/functions/posix_spawnattr_getpgroup.html>
+         <https://pubs.opengroup.org/onlinepubs/9699919799/functions/setpgid.html>
+         We don't support this case; it produces error EINVAL above.  */
+  /* <https://docs.microsoft.com/en-us/windows/win32/procthread/process-creation-flags>  */
+  DWORD process_creation_flags =
+    (attrp != NULL && (attrp->_flags & POSIX_SPAWN_SETPGROUP) != 0 ? DETACHED_PROCESS : 0);
+
+  char *argv_mem_to_free;
+  const char **argv = prepare_spawn (prog_argv, &argv_mem_to_free);
+  if (argv == NULL)
+    return errno; /* errno is set here */
+  argv++;
+
+  /* Compose the command.  */
+  char *command = compose_command (argv);
+  if (command == NULL)
+    {
+      free (argv_mem_to_free);
+      return ENOMEM;
+    }
+
+  /* Copy *ENVP into a contiguous block of memory.  */
+  char *envblock;
+  if (envp == NULL)
+    envblock = NULL;
+  else
+    {
+      envblock = compose_envblock (envp);
+      if (envblock == NULL)
+        {
+          free (command);
+          free (argv_mem_to_free);
+          return ENOMEM;
+        }
+    }
+
+  /* Set up the array of handles to inherit.
+     Duplicate each handle, so that a spawn_do_close action (below) has no
+     effect on the file descriptors of the current process.  Alternatively,
+     we could store, for each handle, a bit that tells whether it is shared
+     with the current process.  But this is simpler.  */
+  struct inheritable_handles inh_handles;
+  if (init_inheritable_handles (&inh_handles, true) < 0)
+    goto failed_1;
+
+  /* Directory in which to execute the new process.  */
+  const char *directory = NULL;
+
+  /* Execute the file_actions, modifying the inh_handles instead of the
+     file descriptors of the current process.  */
+  if (file_actions != NULL)
+    {
+      HANDLE curr_process = GetCurrentProcess ();
+      int cnt;
+
+      for (cnt = 0; cnt < file_actions->_used; ++cnt)
+        {
+          struct __spawn_action *action = &file_actions->_actions[cnt];
+
+          switch (action->tag)
+            {
+            case spawn_do_close:
+              {
+                int fd = action->action.close_action.fd;
+                if (do_close (&inh_handles, fd, false) < 0)
+                  goto failed_2;
+              }
+              break;
+
+            case spawn_do_open:
+              {
+                int newfd = action->action.open_action.fd;
+                const char *filename = action->action.open_action.path;
+                int flags = action->action.open_action.oflag;
+                mode_t mode = action->action.open_action.mode;
+                if (do_open (&inh_handles, newfd, filename, directory,
+                             flags, mode)
+                    < 0)
+                  goto failed_2;
+              }
+              break;
+
+            case spawn_do_dup2:
+              {
+                int oldfd = action->action.dup2_action.fd;
+                int newfd = action->action.dup2_action.newfd;
+                if (do_dup2 (&inh_handles, oldfd, newfd, curr_process) < 0)
+                  goto failed_2;
+              }
+              break;
+
+            case spawn_do_chdir:
+              {
+                char *newdir = action->action.chdir_action.path;
+                if (directory != NULL && IS_RELATIVE_FILE_NAME (newdir))
+                  {
+                    newdir = concatenated_filename (directory, newdir, NULL);
+                    if (newdir == NULL)
+                      {
+                        errno = ENOMEM;
+                        goto failed_2;
+                      }
+                  }
+                directory = newdir;
+              }
+              break;
+
+            case spawn_do_fchdir:
+              /* Not supported in this implementation.  */
+              errno = EINVAL;
+              goto failed_2;
+            }
+        }
+
+# if SPAWN_INTERNAL_OPTIMIZE_DUPLICATEHANDLE
+      /* Do the remaining delayed dup2 invocations.  */
+      if (do_remaining_delayed_dup2 (&inh_handles, curr_process) < 0)
+        goto failed_2;
+# endif
+    }
+
+  /* Close the handles in inh_handles that are not meant to be preserved in the
+     child process, and reduce inh_handles.count to the minimum needed.  */
+  shrink_inheritable_handles (&inh_handles);
+
+  /* CreateProcess
+     <https://docs.microsoft.com/en-us/windows/win32/api/processthreadsapi/nf-processthreadsapi-createprocessa>  */
+  /* STARTUPINFO
+     <https://docs.microsoft.com/en-us/windows/win32/api/processthreadsapi/ns-processthreadsapi-startupinfoa>  */
+  STARTUPINFO sinfo;
+  sinfo.cb = sizeof (STARTUPINFO);
+  sinfo.lpReserved = NULL;
+  sinfo.lpDesktop = NULL;
+  sinfo.lpTitle = NULL;
+  if (compose_handles_block (&inh_handles, &sinfo) < 0)
+    goto failed_2;
+
+  /* Perform the PATH search now, considering the final DIRECTORY.  */
+  char *resolved_prog_filename_to_free = NULL;
+  {
+    const char *resolved_prog_filename =
+      find_in_given_path (prog_filename, use_path ? getenv ("PATH") : "",
+                          directory, false);
+    if (resolved_prog_filename == NULL)
+      goto failed_3;
+    if (resolved_prog_filename != prog_filename)
+      resolved_prog_filename_to_free = (char *) resolved_prog_filename;
+    prog_filename = resolved_prog_filename;
+  }
+
+  PROCESS_INFORMATION pinfo;
+  if (!CreateProcess (prog_filename, command, NULL, NULL, TRUE,
+                      process_creation_flags, envblock, directory, &sinfo,
+                      &pinfo))
+    {
+      DWORD error = GetLastError ();
+
+      free (resolved_prog_filename_to_free);
+      free (sinfo.lpReserved2);
+      close_inheritable_handles (&inh_handles);
+      free_inheritable_handles (&inh_handles);
+      free (envblock);
+      free (command);
+      free (argv_mem_to_free);
+
+      return convert_CreateProcess_error (error);
+    }
+
+  if (pinfo.hThread)
+    CloseHandle (pinfo.hThread);
+
+  free (resolved_prog_filename_to_free);
+  free (sinfo.lpReserved2);
+  close_inheritable_handles (&inh_handles);
+  free_inheritable_handles (&inh_handles);
+  free (envblock);
+  free (command);
+  free (argv_mem_to_free);
+
+  if (pid != NULL)
+    *pid = (intptr_t) pinfo.hProcess;
+  return 0;
+
+ failed_3:
+  {
+    int saved_errno = errno;
+    free (sinfo.lpReserved2);
+    close_inheritable_handles (&inh_handles);
+    free_inheritable_handles (&inh_handles);
+    free (envblock);
+    free (command);
+    free (argv_mem_to_free);
+    return saved_errno;
+  }
+
+ failed_2:
+  {
+    int saved_errno = errno;
+    close_inheritable_handles (&inh_handles);
+    free_inheritable_handles (&inh_handles);
+    free (envblock);
+    free (command);
+    free (argv_mem_to_free);
+    return saved_errno;
+  }
+
+ failed_1:
+  free (envblock);
+  free (command);
+  free (argv_mem_to_free);
+  return errno;
+}
+
+#else
+
+
+/* The warning "warning: 'vfork' is deprecated: Use posix_spawn or fork" seen
+   on macOS 12 is pointless, as we use vfork only when it is safe or when the
+   user has explicitly requested it.  Silence this warning.  */
+#if __GNUC__ >= 3
+# pragma GCC diagnostic ignored "-Wdeprecated-declarations"
+#endif
+
+/* Spawn a new process executing PATH with the attributes describes in *ATTRP.
+   Before running the process perform the actions described in FILE-ACTIONS. */
+int
+__spawni (pid_t *pid, const char *file,
+          const posix_spawn_file_actions_t *file_actions,
+          const posix_spawnattr_t *attrp, const char *const argv[],
+          const char *const envp[], int use_path)
+{
+  pid_t new_pid;
+  char *path, *p, *name;
+  size_t len;
+  size_t pathlen;
+
+  /* Do this once.  */
+  short int flags = attrp == NULL ? 0 : attrp->_flags;
+
+  /* Avoid gcc warning
+       "variable 'flags' might be clobbered by 'longjmp' or 'vfork'"  */
+  (void) &flags;
+
+  /* Generate the new process.  */
+#if HAVE_VFORK
+  if ((flags & POSIX_SPAWN_USEVFORK) != 0
+      /* If no major work is done, allow using vfork.  Note that we
+         might perform the path searching.  But this would be done by
+         a call to execvp(), too, and such a call must be OK according
+         to POSIX.  */
+      || ((flags & (POSIX_SPAWN_SETSIGMASK | POSIX_SPAWN_SETSIGDEF
+                    | POSIX_SPAWN_SETSCHEDPARAM | POSIX_SPAWN_SETSCHEDULER
+                    | POSIX_SPAWN_SETPGROUP | POSIX_SPAWN_RESETIDS)) == 0
+          && file_actions == NULL))
+    new_pid = vfork ();
+  else
+#endif
+    new_pid = fork ();
+
+  if (new_pid != 0)
+    {
+      if (new_pid < 0)
+        return errno;
+
+      /* The call was successful.  Store the PID if necessary.  */
+      if (pid != NULL)
+        *pid = new_pid;
+
+      return 0;
+    }
+
+  /* Set signal mask.  */
+  if ((flags & POSIX_SPAWN_SETSIGMASK) != 0
+      && sigprocmask (SIG_SETMASK, &attrp->_ss, NULL) != 0)
+    _exit (SPAWN_ERROR);
+
+  /* Set signal default action.  */
+  if ((flags & POSIX_SPAWN_SETSIGDEF) != 0)
+    {
+      /* We have to iterate over all signals.  This could possibly be
+         done better but it requires system specific solutions since
+         the sigset_t data type can be very different on different
+         architectures.  */
+      int sig;
+      struct sigaction sa;
+
+      memset (&sa, '\0', sizeof (sa));
+      sa.sa_handler = SIG_DFL;
+
+      for (sig = 1; sig <= NSIG; ++sig)
+        if (sigismember (&attrp->_sd, sig) != 0
+            && sigaction (sig, &sa, NULL) != 0)
+          _exit (SPAWN_ERROR);
+
+    }
+
+#if (_LIBC ? defined _POSIX_PRIORITY_SCHEDULING : HAVE_SCHED_SETPARAM && HAVE_SCHED_SETSCHEDULER)
+  /* Set the scheduling algorithm and parameters.  */
+  if ((flags & (POSIX_SPAWN_SETSCHEDPARAM | POSIX_SPAWN_SETSCHEDULER))
+      == POSIX_SPAWN_SETSCHEDPARAM)
+    {
+      if (sched_setparam (0, &attrp->_sp) == -1)
+        _exit (SPAWN_ERROR);
+    }
+  else if ((flags & POSIX_SPAWN_SETSCHEDULER) != 0)
+    {
+      if (sched_setscheduler (0, attrp->_policy,
+                              (flags & POSIX_SPAWN_SETSCHEDPARAM) != 0
+                              ? &attrp->_sp : NULL) == -1)
+        _exit (SPAWN_ERROR);
+    }
+#endif
+
+  /* Set the process group ID.  */
+  if ((flags & POSIX_SPAWN_SETPGROUP) != 0
+      && setpgid (0, attrp->_pgrp) != 0)
+    _exit (SPAWN_ERROR);
+
+  /* Set the effective user and group IDs.  */
+  if ((flags & POSIX_SPAWN_RESETIDS) != 0
+      && (local_seteuid (getuid ()) != 0
+          || local_setegid (getgid ()) != 0))
+    _exit (SPAWN_ERROR);
+
+  /* Execute the file actions.  */
+  if (file_actions != NULL)
+    {
+      int cnt;
+
+      for (cnt = 0; cnt < file_actions->_used; ++cnt)
+        {
+          struct __spawn_action *action = &file_actions->_actions[cnt];
+
+          switch (action->tag)
+            {
+            case spawn_do_close:
+              if (close_not_cancel (action->action.close_action.fd) != 0)
+                /* Signal the error.  */
+                _exit (SPAWN_ERROR);
+              break;
+
+            case spawn_do_open:
+              {
+                int new_fd = open_not_cancel (action->action.open_action.path,
+                                              action->action.open_action.oflag
+                                              | O_LARGEFILE,
+                                              action->action.open_action.mode);
+
+                if (new_fd == -1)
+                  /* The 'open' call failed.  */
+                  _exit (SPAWN_ERROR);
+
+                /* Make sure the desired file descriptor is used.  */
+                if (new_fd != action->action.open_action.fd)
+                  {
+                    if (dup2 (new_fd, action->action.open_action.fd)
+                        != action->action.open_action.fd)
+                      /* The 'dup2' call failed.  */
+                      _exit (SPAWN_ERROR);
+
+                    if (close_not_cancel (new_fd) != 0)
+                      /* The 'close' call failed.  */
+                      _exit (SPAWN_ERROR);
+                  }
+              }
+              break;
+
+            case spawn_do_dup2:
+              if (dup2 (action->action.dup2_action.fd,
+                        action->action.dup2_action.newfd)
+                  != action->action.dup2_action.newfd)
+                /* The 'dup2' call failed.  */
+                _exit (SPAWN_ERROR);
+              break;
+
+            case spawn_do_chdir:
+              if (chdir (action->action.chdir_action.path) < 0)
+                /* The 'chdir' call failed.  */
+                _exit (SPAWN_ERROR);
+              break;
+
+            case spawn_do_fchdir:
+              if (fchdir (action->action.fchdir_action.fd) < 0)
+                /* The 'fchdir' call failed.  */
+                _exit (SPAWN_ERROR);
+              break;
+            }
+        }
+    }
+
+  if (! use_path || strchr (file, '/') != NULL)
+    {
+      /* The FILE parameter is actually a path.  */
+      execve (file, (char * const *) argv, (char * const *) envp);
+
+      /* Oh, oh.  'execve' returns.  This is bad.  */
+      _exit (SPAWN_ERROR);
+    }
+
+  /* We have to search for FILE on the path.  */
+  path = getenv ("PATH");
+  if (path == NULL)
+    {
+#if HAVE_CONFSTR
+      /* There is no 'PATH' in the environment.
+         The default search path is the current directory
+         followed by the path 'confstr' returns for '_CS_PATH'.  */
+      len = confstr (_CS_PATH, (char *) NULL, 0);
+      path = (char *) alloca (1 + len);
+      path[0] = ':';
+      (void) confstr (_CS_PATH, path + 1, len);
+#else
+      /* Pretend that the PATH contains only the current directory.  */
+      path = "";
+#endif
+    }
+
+  len = strlen (file) + 1;
+  pathlen = strlen (path);
+  name = alloca (pathlen + len + 1);
+  /* Copy the file name at the top.  */
+  name = (char *) memcpy (name + pathlen + 1, file, len);
+  /* And add the slash.  */
+  *--name = '/';
+
+  p = path;
+  do
+    {
+      char *startp;
+
+      path = p;
+      p = strchrnul (path, ':');
+
+      if (p == path)
+        /* Two adjacent colons, or a colon at the beginning or the end
+           of 'PATH' means to search the current directory.  */
+        startp = name + 1;
+      else
+        startp = (char *) memcpy (name - (p - path), path, p - path);
+
+      /* Try to execute this name.  If it works, execv will not return.  */
+      execve (startp, (char * const *) argv, (char * const *) envp);
+
+      switch (errno)
+        {
+        case EACCES:
+        case ENOENT:
+        case ESTALE:
+        case ENOTDIR:
+          /* Those errors indicate the file is missing or not executable
+             by us, in which case we want to just try the next path
+             directory.  */
+          break;
+
+        default:
+          /* Some other error means we found an executable file, but
+             something went wrong executing it; return the error to our
+             caller.  */
+          _exit (SPAWN_ERROR);
+        }
+    }
+  while (*p++ != '\0');
+
+  /* Return with an error.  */
+  _exit (SPAWN_ERROR);
+}
+
+#endif