public inbox for util-linux@vger.kernel.org
 help / color / mirror / Atom feed
* [PATCH] uuidd: use pkg-config to find systemd-deamon support
@ 2014-01-06 23:39 Sami Kerola
  2014-01-07  8:58 ` Karel Zak
  2014-01-07 11:20 ` Karel Zak
  0 siblings, 2 replies; 17+ messages in thread
From: Sami Kerola @ 2014-01-06 23:39 UTC (permalink / raw)
  To: util-linux; +Cc: kerolasa

Signed-off-by: Sami Kerola <kerolasa@iki.fi>
---
 configure.ac             |   3 +
 misc-utils/Makemodule.am |   3 +-
 misc-utils/sd-daemon.c   | 520 -----------------------------------------------
 misc-utils/sd-daemon.h   | 282 -------------------------
 misc-utils/uuidd.c       |   2 +-
 5 files changed, 5 insertions(+), 805 deletions(-)
 delete mode 100644 misc-utils/sd-daemon.c
 delete mode 100644 misc-utils/sd-daemon.h

diff --git a/configure.ac b/configure.ac
index 8665780..290e2c6 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1394,6 +1394,9 @@ AM_CONDITIONAL([USE_SOCKET_ACTIVATION], [test "x$enable_socket_activation" = xye
 
 AS_IF([test "x$enable_socket_activation" = xyes], [
   AC_DEFINE([USE_SOCKET_ACTIVATION], [1], [Should uuidd support socket activation?])
+  PKG_CHECK_MODULES([SYSTEMD_DAEMON], [libsystemd-daemon], [], [
+    AC_MSG_ERROR([cannot find libsystemd-daemon support])
+  ])
 ])
 
 AC_ARG_WITH([systemdsystemunitdir],
diff --git a/misc-utils/Makemodule.am b/misc-utils/Makemodule.am
index 686fb15..feaeec4 100644
--- a/misc-utils/Makemodule.am
+++ b/misc-utils/Makemodule.am
@@ -73,8 +73,7 @@ uuidd_CFLAGS = $(DAEMON_CFLAGS) $(AM_CFLAGS) -I$(ul_libuuid_incdir)
 uuidd_LDFLAGS = $(DAEMON_LDFLAGS) $(AM_LDFLAGS)
 uuidd_SOURCES = misc-utils/uuidd.c
 if USE_SOCKET_ACTIVATION
-uuidd_SOURCES += misc-utils/sd-daemon.c misc-utils/sd-daemon.h
-uuidd_LDADD += -lrt
+uuidd_LDADD += $(SYSTEMD_DAEMON_LIBS)
 endif
 if HAVE_SYSTEMD
 systemdsystemunit_DATA += \
diff --git a/misc-utils/sd-daemon.c b/misc-utils/sd-daemon.c
deleted file mode 100644
index 485b301..0000000
--- a/misc-utils/sd-daemon.c
+++ /dev/null
@@ -1,520 +0,0 @@
-/*-*- Mode: C; c-basic-offset: 8; indent-tabs-mode: nil -*-*/
-
-/***
-  Copyright 2010 Lennart Poettering
-
-  Permission is hereby granted, free of charge, to any person
-  obtaining a copy of this software and associated documentation files
-  (the "Software"), to deal in the Software without restriction,
-  including without limitation the rights to use, copy, modify, merge,
-  publish, distribute, sublicense, and/or sell copies of the Software,
-  and to permit persons to whom the Software is furnished to do so,
-  subject to the following conditions:
-
-  The above copyright notice and this permission notice shall be
-  included in all copies or substantial portions of the Software.
-
-  THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
-  EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
-  MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
-  NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
-  BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
-  ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
-  CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
-  SOFTWARE.
-***/
-
-#ifndef _GNU_SOURCE
-#  define _GNU_SOURCE
-#endif
-
-#include <sys/types.h>
-#include <sys/stat.h>
-#include <sys/socket.h>
-#include <sys/un.h>
-#include <fcntl.h>
-#include <netinet/in.h>
-#include <stdlib.h>
-#include <errno.h>
-#include <unistd.h>
-#include <string.h>
-#include <stdarg.h>
-#include <stdio.h>
-#include <stddef.h>
-#include <limits.h>
-
-#if defined(__linux__) && !defined(SD_DAEMON_DISABLE_MQ)
-#  include <mqueue.h>
-#endif
-
-#include "sd-daemon.h"
-
-#if (__GNUC__ >= 4)
-#  ifdef SD_EXPORT_SYMBOLS
-/* Export symbols */
-#    define _sd_export_ __attribute__ ((visibility("default")))
-#  else
-/* Don't export the symbols */
-#    define _sd_export_ __attribute__ ((visibility("hidden")))
-#  endif
-#else
-#  define _sd_export_
-#endif
-
-_sd_export_ int sd_listen_fds(int unset_environment) {
-
-#if defined(DISABLE_SYSTEMD) || !defined(__linux__)
-        return 0;
-#else
-        int r, fd;
-        const char *e;
-        char *p = NULL;
-        unsigned long l;
-
-        e = getenv("LISTEN_PID");
-        if (!e) {
-                r = 0;
-                goto finish;
-        }
-
-        errno = 0;
-        l = strtoul(e, &p, 10);
-
-        if (errno > 0) {
-                r = -errno;
-                goto finish;
-        }
-
-        if (!p || p == e || *p || l <= 0) {
-                r = -EINVAL;
-                goto finish;
-        }
-
-        /* Is this for us? */
-        if (getpid() != (pid_t) l) {
-                r = 0;
-                goto finish;
-        }
-
-        e = getenv("LISTEN_FDS");
-        if (!e) {
-                r = 0;
-                goto finish;
-        }
-
-        errno = 0;
-        l = strtoul(e, &p, 10);
-
-        if (errno > 0) {
-                r = -errno;
-                goto finish;
-        }
-
-        if (!p || p == e || *p) {
-                r = -EINVAL;
-                goto finish;
-        }
-
-        for (fd = SD_LISTEN_FDS_START; fd < SD_LISTEN_FDS_START + (int) l; fd ++) {
-                int flags;
-
-                flags = fcntl(fd, F_GETFD);
-                if (flags < 0) {
-                        r = -errno;
-                        goto finish;
-                }
-
-                if (flags & FD_CLOEXEC)
-                        continue;
-
-                if (fcntl(fd, F_SETFD, flags | FD_CLOEXEC) < 0) {
-                        r = -errno;
-                        goto finish;
-                }
-        }
-
-        r = (int) l;
-
-finish:
-        if (unset_environment) {
-                unsetenv("LISTEN_PID");
-                unsetenv("LISTEN_FDS");
-        }
-
-        return r;
-#endif
-}
-
-_sd_export_ int sd_is_fifo(int fd, const char *path) {
-        struct stat st_fd;
-
-        if (fd < 0)
-                return -EINVAL;
-
-        if (fstat(fd, &st_fd) < 0)
-                return -errno;
-
-        if (!S_ISFIFO(st_fd.st_mode))
-                return 0;
-
-        if (path) {
-                struct stat st_path;
-
-                if (stat(path, &st_path) < 0) {
-
-                        if (errno == ENOENT || errno == ENOTDIR)
-                                return 0;
-
-                        return -errno;
-                }
-
-                return
-                        st_path.st_dev == st_fd.st_dev &&
-                        st_path.st_ino == st_fd.st_ino;
-        }
-
-        return 1;
-}
-
-_sd_export_ int sd_is_special(int fd, const char *path) {
-        struct stat st_fd;
-
-        if (fd < 0)
-                return -EINVAL;
-
-        if (fstat(fd, &st_fd) < 0)
-                return -errno;
-
-        if (!S_ISREG(st_fd.st_mode) && !S_ISCHR(st_fd.st_mode))
-                return 0;
-
-        if (path) {
-                struct stat st_path;
-
-                if (stat(path, &st_path) < 0) {
-
-                        if (errno == ENOENT || errno == ENOTDIR)
-                                return 0;
-
-                        return -errno;
-                }
-
-                if (S_ISREG(st_fd.st_mode) && S_ISREG(st_path.st_mode))
-                        return
-                                st_path.st_dev == st_fd.st_dev &&
-                                st_path.st_ino == st_fd.st_ino;
-                else if (S_ISCHR(st_fd.st_mode) && S_ISCHR(st_path.st_mode))
-                        return st_path.st_rdev == st_fd.st_rdev;
-                else
-                        return 0;
-        }
-
-        return 1;
-}
-
-static int sd_is_socket_internal(int fd, int type, int listening) {
-        struct stat st_fd;
-
-        if (fd < 0 || type < 0)
-                return -EINVAL;
-
-        if (fstat(fd, &st_fd) < 0)
-                return -errno;
-
-        if (!S_ISSOCK(st_fd.st_mode))
-                return 0;
-
-        if (type != 0) {
-                int other_type = 0;
-                socklen_t l = sizeof(other_type);
-
-                if (getsockopt(fd, SOL_SOCKET, SO_TYPE, &other_type, &l) < 0)
-                        return -errno;
-
-                if (l != sizeof(other_type))
-                        return -EINVAL;
-
-                if (other_type != type)
-                        return 0;
-        }
-
-        if (listening >= 0) {
-                int accepting = 0;
-                socklen_t l = sizeof(accepting);
-
-                if (getsockopt(fd, SOL_SOCKET, SO_ACCEPTCONN, &accepting, &l) < 0)
-                        return -errno;
-
-                if (l != sizeof(accepting))
-                        return -EINVAL;
-
-                if (!accepting != !listening)
-                        return 0;
-        }
-
-        return 1;
-}
-
-union sockaddr_union {
-        struct sockaddr sa;
-        struct sockaddr_in in4;
-        struct sockaddr_in6 in6;
-        struct sockaddr_un un;
-        struct sockaddr_storage storage;
-};
-
-_sd_export_ int sd_is_socket(int fd, int family, int type, int listening) {
-        int r;
-
-        if (family < 0)
-                return -EINVAL;
-
-        r = sd_is_socket_internal(fd, type, listening);
-        if (r <= 0)
-                return r;
-
-        if (family > 0) {
-                union sockaddr_union sockaddr = {};
-                socklen_t l = sizeof(sockaddr);
-
-                if (getsockname(fd, &sockaddr.sa, &l) < 0)
-                        return -errno;
-
-                if (l < sizeof(sa_family_t))
-                        return -EINVAL;
-
-                return sockaddr.sa.sa_family == family;
-        }
-
-        return 1;
-}
-
-_sd_export_ int sd_is_socket_inet(int fd, int family, int type, int listening, uint16_t port) {
-        union sockaddr_union sockaddr = {};
-        socklen_t l = sizeof(sockaddr);
-        int r;
-
-        if (family != 0 && family != AF_INET && family != AF_INET6)
-                return -EINVAL;
-
-        r = sd_is_socket_internal(fd, type, listening);
-        if (r <= 0)
-                return r;
-
-        if (getsockname(fd, &sockaddr.sa, &l) < 0)
-                return -errno;
-
-        if (l < sizeof(sa_family_t))
-                return -EINVAL;
-
-        if (sockaddr.sa.sa_family != AF_INET &&
-            sockaddr.sa.sa_family != AF_INET6)
-                return 0;
-
-        if (family > 0)
-                if (sockaddr.sa.sa_family != family)
-                        return 0;
-
-        if (port > 0) {
-                if (sockaddr.sa.sa_family == AF_INET) {
-                        if (l < sizeof(struct sockaddr_in))
-                                return -EINVAL;
-
-                        return htons(port) == sockaddr.in4.sin_port;
-                } else {
-                        if (l < sizeof(struct sockaddr_in6))
-                                return -EINVAL;
-
-                        return htons(port) == sockaddr.in6.sin6_port;
-                }
-        }
-
-        return 1;
-}
-
-_sd_export_ int sd_is_socket_unix(int fd, int type, int listening, const char *path, size_t length) {
-        union sockaddr_union sockaddr = {};
-        socklen_t l = sizeof(sockaddr);
-        int r;
-
-        r = sd_is_socket_internal(fd, type, listening);
-        if (r <= 0)
-                return r;
-
-        if (getsockname(fd, &sockaddr.sa, &l) < 0)
-                return -errno;
-
-        if (l < sizeof(sa_family_t))
-                return -EINVAL;
-
-        if (sockaddr.sa.sa_family != AF_UNIX)
-                return 0;
-
-        if (path) {
-                if (length == 0)
-                        length = strlen(path);
-
-                if (length == 0)
-                        /* Unnamed socket */
-                        return l == offsetof(struct sockaddr_un, sun_path);
-
-                if (path[0])
-                        /* Normal path socket */
-                        return
-                                (l >= offsetof(struct sockaddr_un, sun_path) + length + 1) &&
-                                memcmp(path, sockaddr.un.sun_path, length+1) == 0;
-                else
-                        /* Abstract namespace socket */
-                        return
-                                (l == offsetof(struct sockaddr_un, sun_path) + length) &&
-                                memcmp(path, sockaddr.un.sun_path, length) == 0;
-        }
-
-        return 1;
-}
-
-_sd_export_ int sd_is_mq(int fd, const char *path) {
-#if !defined(__linux__) || defined(SD_DAEMON_DISABLE_MQ)
-        return 0;
-#else
-        struct mq_attr attr;
-
-        if (fd < 0)
-                return -EINVAL;
-
-        if (mq_getattr(fd, &attr) < 0)
-                return -errno;
-
-        if (path) {
-                char fpath[PATH_MAX];
-                struct stat a, b;
-
-                if (path[0] != '/')
-                        return -EINVAL;
-
-                if (fstat(fd, &a) < 0)
-                        return -errno;
-
-                strncpy(stpcpy(fpath, "/dev/mqueue"), path, sizeof(fpath) - 12);
-                fpath[sizeof(fpath)-1] = 0;
-
-                if (stat(fpath, &b) < 0)
-                        return -errno;
-
-                if (a.st_dev != b.st_dev ||
-                    a.st_ino != b.st_ino)
-                        return 0;
-        }
-
-        return 1;
-#endif
-}
-
-_sd_export_ int sd_notify(int unset_environment, const char *state) {
-#if defined(DISABLE_SYSTEMD) || !defined(__linux__) || !defined(SOCK_CLOEXEC)
-        return 0;
-#else
-        int fd = -1, r;
-        struct msghdr msghdr;
-        struct iovec iovec;
-        union sockaddr_union sockaddr;
-        const char *e;
-
-        if (!state) {
-                r = -EINVAL;
-                goto finish;
-        }
-
-        e = getenv("NOTIFY_SOCKET");
-        if (!e)
-                return 0;
-
-        /* Must be an abstract socket, or an absolute path */
-        if ((e[0] != '@' && e[0] != '/') || e[1] == 0) {
-                r = -EINVAL;
-                goto finish;
-        }
-
-        fd = socket(AF_UNIX, SOCK_DGRAM|SOCK_CLOEXEC, 0);
-        if (fd < 0) {
-                r = -errno;
-                goto finish;
-        }
-
-        memset(&sockaddr, 0, sizeof(sockaddr));
-        sockaddr.sa.sa_family = AF_UNIX;
-        strncpy(sockaddr.un.sun_path, e, sizeof(sockaddr.un.sun_path));
-
-        if (sockaddr.un.sun_path[0] == '@')
-                sockaddr.un.sun_path[0] = 0;
-
-        memset(&iovec, 0, sizeof(iovec));
-        iovec.iov_base = (char*) state;
-        iovec.iov_len = strlen(state);
-
-        memset(&msghdr, 0, sizeof(msghdr));
-        msghdr.msg_name = &sockaddr;
-        msghdr.msg_namelen = offsetof(struct sockaddr_un, sun_path) + strlen(e);
-
-        if (msghdr.msg_namelen > sizeof(struct sockaddr_un))
-                msghdr.msg_namelen = sizeof(struct sockaddr_un);
-
-        msghdr.msg_iov = &iovec;
-        msghdr.msg_iovlen = 1;
-
-        if (sendmsg(fd, &msghdr, MSG_NOSIGNAL) < 0) {
-                r = -errno;
-                goto finish;
-        }
-
-        r = 1;
-
-finish:
-        if (unset_environment)
-                unsetenv("NOTIFY_SOCKET");
-
-        if (fd >= 0)
-                close(fd);
-
-        return r;
-#endif
-}
-
-_sd_export_ int sd_notifyf(int unset_environment, const char *format, ...) {
-#if defined(DISABLE_SYSTEMD) || !defined(__linux__)
-        return 0;
-#else
-        va_list ap;
-        char *p = NULL;
-        int r;
-
-        va_start(ap, format);
-        r = vasprintf(&p, format, ap);
-        va_end(ap);
-
-        if (r < 0 || !p)
-                return -ENOMEM;
-
-        r = sd_notify(unset_environment, p);
-        free(p);
-
-        return r;
-#endif
-}
-
-_sd_export_ int sd_booted(void) {
-#if defined(DISABLE_SYSTEMD) || !defined(__linux__)
-        return 0;
-#else
-        struct stat st;
-
-        /* We test whether the runtime unit file directory has been
-         * created. This takes place in mount-setup.c, so is
-         * guaranteed to happen very early during boot. */
-
-        if (lstat("/run/systemd/system/", &st) < 0)
-                return 0;
-
-        return !!S_ISDIR(st.st_mode);
-#endif
-}
diff --git a/misc-utils/sd-daemon.h b/misc-utils/sd-daemon.h
deleted file mode 100644
index daa3f4c..0000000
--- a/misc-utils/sd-daemon.h
+++ /dev/null
@@ -1,282 +0,0 @@
-/*-*- Mode: C; c-basic-offset: 8; indent-tabs-mode: nil -*-*/
-
-#ifndef foosddaemonhfoo
-#define foosddaemonhfoo
-
-/***
-  Copyright 2010 Lennart Poettering
-
-  Permission is hereby granted, free of charge, to any person
-  obtaining a copy of this software and associated documentation files
-  (the "Software"), to deal in the Software without restriction,
-  including without limitation the rights to use, copy, modify, merge,
-  publish, distribute, sublicense, and/or sell copies of the Software,
-  and to permit persons to whom the Software is furnished to do so,
-  subject to the following conditions:
-
-  The above copyright notice and this permission notice shall be
-  included in all copies or substantial portions of the Software.
-
-  THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
-  EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
-  MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
-  NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
-  BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
-  ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
-  CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
-  SOFTWARE.
-***/
-
-#include <sys/types.h>
-#include <inttypes.h>
-
-#ifdef __cplusplus
-extern "C" {
-#endif
-
-/*
-  Reference implementation of a few systemd related interfaces for
-  writing daemons. These interfaces are trivial to implement. To
-  simplify porting we provide this reference implementation.
-  Applications are welcome to reimplement the algorithms described
-  here if they do not want to include these two source files.
-
-  The following functionality is provided:
-
-  - Support for logging with log levels on stderr
-  - File descriptor passing for socket-based activation
-  - Daemon startup and status notification
-  - Detection of systemd boots
-
-  You may compile this with -DDISABLE_SYSTEMD to disable systemd
-  support. This makes all those calls NOPs that are directly related to
-  systemd (i.e. only sd_is_xxx() will stay useful).
-
-  Since this is drop-in code we don't want any of our symbols to be
-  exported in any case. Hence we declare hidden visibility for all of
-  them.
-
-  You may find an up-to-date version of these source files online:
-
-  http://cgit.freedesktop.org/systemd/systemd/plain/src/systemd/sd-daemon.h
-  http://cgit.freedesktop.org/systemd/systemd/plain/src/libsystemd-daemon/sd-daemon.c
-
-  This should compile on non-Linux systems, too, but with the
-  exception of the sd_is_xxx() calls all functions will become NOPs.
-
-  See sd-daemon(3) for more information.
-*/
-
-#ifndef _sd_printf_attr_
-#  if __GNUC__ >= 4
-#    define _sd_printf_attr_(a,b) __attribute__ ((format (printf, a, b)))
-#  else
-#    define _sd_printf_attr_(a,b)
-#  endif
-#endif
-
-/*
-  Log levels for usage on stderr:
-
-          fprintf(stderr, SD_NOTICE "Hello World!\n");
-
-  This is similar to printk() usage in the kernel.
-*/
-#define SD_EMERG   "<0>"  /* system is unusable */
-#define SD_ALERT   "<1>"  /* action must be taken immediately */
-#define SD_CRIT    "<2>"  /* critical conditions */
-#define SD_ERR     "<3>"  /* error conditions */
-#define SD_WARNING "<4>"  /* warning conditions */
-#define SD_NOTICE  "<5>"  /* normal but significant condition */
-#define SD_INFO    "<6>"  /* informational */
-#define SD_DEBUG   "<7>"  /* debug-level messages */
-
-/* The first passed file descriptor is fd 3 */
-#define SD_LISTEN_FDS_START 3
-
-/*
-  Returns how many file descriptors have been passed, or a negative
-  errno code on failure. Optionally, removes the $LISTEN_FDS and
-  $LISTEN_PID file descriptors from the environment (recommended, but
-  problematic in threaded environments). If r is the return value of
-  this function you'll find the file descriptors passed as fds
-  SD_LISTEN_FDS_START to SD_LISTEN_FDS_START+r-1. Returns a negative
-  errno style error code on failure. This function call ensures that
-  the FD_CLOEXEC flag is set for the passed file descriptors, to make
-  sure they are not passed on to child processes. If FD_CLOEXEC shall
-  not be set, the caller needs to unset it after this call for all file
-  descriptors that are used.
-
-  See sd_listen_fds(3) for more information.
-*/
-int sd_listen_fds(int unset_environment);
-
-/*
-  Helper call for identifying a passed file descriptor. Returns 1 if
-  the file descriptor is a FIFO in the file system stored under the
-  specified path, 0 otherwise. If path is NULL a path name check will
-  not be done and the call only verifies if the file descriptor
-  refers to a FIFO. Returns a negative errno style error code on
-  failure.
-
-  See sd_is_fifo(3) for more information.
-*/
-int sd_is_fifo(int fd, const char *path);
-
-/*
-  Helper call for identifying a passed file descriptor. Returns 1 if
-  the file descriptor is a special character device on the file
-  system stored under the specified path, 0 otherwise.
-  If path is NULL a path name check will not be done and the call
-  only verifies if the file descriptor refers to a special character.
-  Returns a negative errno style error code on failure.
-
-  See sd_is_special(3) for more information.
-*/
-int sd_is_special(int fd, const char *path);
-
-/*
-  Helper call for identifying a passed file descriptor. Returns 1 if
-  the file descriptor is a socket of the specified family (AF_INET,
-  ...) and type (SOCK_DGRAM, SOCK_STREAM, ...), 0 otherwise. If
-  family is 0 a socket family check will not be done. If type is 0 a
-  socket type check will not be done and the call only verifies if
-  the file descriptor refers to a socket. If listening is > 0 it is
-  verified that the socket is in listening mode. (i.e. listen() has
-  been called) If listening is == 0 it is verified that the socket is
-  not in listening mode. If listening is < 0 no listening mode check
-  is done. Returns a negative errno style error code on failure.
-
-  See sd_is_socket(3) for more information.
-*/
-int sd_is_socket(int fd, int family, int type, int listening);
-
-/*
-  Helper call for identifying a passed file descriptor. Returns 1 if
-  the file descriptor is an Internet socket, of the specified family
-  (either AF_INET or AF_INET6) and the specified type (SOCK_DGRAM,
-  SOCK_STREAM, ...), 0 otherwise. If version is 0 a protocol version
-  check is not done. If type is 0 a socket type check will not be
-  done. If port is 0 a socket port check will not be done. The
-  listening flag is used the same way as in sd_is_socket(). Returns a
-  negative errno style error code on failure.
-
-  See sd_is_socket_inet(3) for more information.
-*/
-int sd_is_socket_inet(int fd, int family, int type, int listening, uint16_t port);
-
-/*
-  Helper call for identifying a passed file descriptor. Returns 1 if
-  the file descriptor is an AF_UNIX socket of the specified type
-  (SOCK_DGRAM, SOCK_STREAM, ...) and path, 0 otherwise. If type is 0
-  a socket type check will not be done. If path is NULL a socket path
-  check will not be done. For normal AF_UNIX sockets set length to
-  0. For abstract namespace sockets set length to the length of the
-  socket name (including the initial 0 byte), and pass the full
-  socket path in path (including the initial 0 byte). The listening
-  flag is used the same way as in sd_is_socket(). Returns a negative
-  errno style error code on failure.
-
-  See sd_is_socket_unix(3) for more information.
-*/
-int sd_is_socket_unix(int fd, int type, int listening, const char *path, size_t length);
-
-/*
-  Helper call for identifying a passed file descriptor. Returns 1 if
-  the file descriptor is a POSIX Message Queue of the specified name,
-  0 otherwise. If path is NULL a message queue name check is not
-  done. Returns a negative errno style error code on failure.
-*/
-int sd_is_mq(int fd, const char *path);
-
-/*
-  Informs systemd about changed daemon state. This takes a number of
-  newline separated environment-style variable assignments in a
-  string. The following variables are known:
-
-     READY=1      Tells systemd that daemon startup is finished (only
-                  relevant for services of Type=notify). The passed
-                  argument is a boolean "1" or "0". Since there is
-                  little value in signaling non-readiness the only
-                  value daemons should send is "READY=1".
-
-     STATUS=...   Passes a single-line status string back to systemd
-                  that describes the daemon state. This is free-from
-                  and can be used for various purposes: general state
-                  feedback, fsck-like programs could pass completion
-                  percentages and failing programs could pass a human
-                  readable error message. Example: "STATUS=Completed
-                  66% of file system check..."
-
-     ERRNO=...    If a daemon fails, the errno-style error code,
-                  formatted as string. Example: "ERRNO=2" for ENOENT.
-
-     BUSERROR=... If a daemon fails, the D-Bus error-style error
-                  code. Example: "BUSERROR=org.freedesktop.DBus.Error.TimedOut"
-
-     MAINPID=...  The main pid of a daemon, in case systemd did not
-                  fork off the process itself. Example: "MAINPID=4711"
-
-     WATCHDOG=1   Tells systemd to update the watchdog timestamp.
-                  Services using this feature should do this in
-                  regular intervals. A watchdog framework can use the
-                  timestamps to detect failed services.
-
-  Daemons can choose to send additional variables. However, it is
-  recommended to prefix variable names not listed above with X_.
-
-  Returns a negative errno-style error code on failure. Returns > 0
-  if systemd could be notified, 0 if it couldn't possibly because
-  systemd is not running.
-
-  Example: When a daemon finished starting up, it could issue this
-  call to notify systemd about it:
-
-     sd_notify(0, "READY=1");
-
-  See sd_notifyf() for more complete examples.
-
-  See sd_notify(3) for more information.
-*/
-int sd_notify(int unset_environment, const char *state);
-
-/*
-  Similar to sd_notify() but takes a format string.
-
-  Example 1: A daemon could send the following after initialization:
-
-     sd_notifyf(0, "READY=1\n"
-                   "STATUS=Processing requests...\n"
-                   "MAINPID=%lu",
-                   (unsigned long) getpid());
-
-  Example 2: A daemon could send the following shortly before
-  exiting, on failure:
-
-     sd_notifyf(0, "STATUS=Failed to start up: %s\n"
-                   "ERRNO=%i",
-                   strerror(errno),
-                   errno);
-
-  See sd_notifyf(3) for more information.
-*/
-int sd_notifyf(int unset_environment, const char *format, ...) _sd_printf_attr_(2,3);
-
-/*
-  Returns > 0 if the system was booted with systemd. Returns < 0 on
-  error. Returns 0 if the system was not booted with systemd. Note
-  that all of the functions above handle non-systemd boots just
-  fine. You should NOT protect them with a call to this function. Also
-  note that this function checks whether the system, not the user
-  session is controlled by systemd. However the functions above work
-  for both user and system services.
-
-  See sd_booted(3) for more information.
-*/
-int sd_booted(void);
-
-#ifdef __cplusplus
-}
-#endif
-
-#endif
diff --git a/misc-utils/uuidd.c b/misc-utils/uuidd.c
index d381eea..cb8bf82 100644
--- a/misc-utils/uuidd.c
+++ b/misc-utils/uuidd.c
@@ -39,7 +39,7 @@ extern int optind;
 #include "strutils.h"
 
 #ifdef USE_SOCKET_ACTIVATION
-#include "sd-daemon.h"
+#include <systemd/sd-daemon.h>
 #endif
 
 #include "nls.h"
-- 
1.8.5.2


^ permalink raw reply related	[flat|nested] 17+ messages in thread

* Re: [PATCH] uuidd: use pkg-config to find systemd-deamon support
  2014-01-06 23:39 [PATCH] uuidd: use pkg-config to find systemd-deamon support Sami Kerola
@ 2014-01-07  8:58 ` Karel Zak
  2014-01-07  9:38   ` Thomas Bächler
  2014-01-07 11:20 ` Karel Zak
  1 sibling, 1 reply; 17+ messages in thread
From: Karel Zak @ 2014-01-07  8:58 UTC (permalink / raw)
  To: Sami Kerola; +Cc: util-linux

On Mon, Jan 06, 2014 at 11:39:08PM +0000, Sami Kerola wrote:
>  configure.ac             |   3 +
>  misc-utils/Makemodule.am |   3 +-
>  misc-utils/sd-daemon.c   | 520 -----------------------------------------------
>  misc-utils/sd-daemon.h   | 282 -------------------------
>  misc-utils/uuidd.c       |   2 +-

 Seems like a good idea, unfortunately I'm not sure about it. 
 
 The systemd (udevd) depends on libblkid+libuuid, it would be 
 probably better to keep util-linux without dependencies on systemd
 libs otherwise we will introduce a cyclic dependence -- it could be
 too tricky for some build systems and users :-) I'll think about it...

    Karel

-- 
 Karel Zak  <kzak@redhat.com>
 http://karelzak.blogspot.com

^ permalink raw reply	[flat|nested] 17+ messages in thread

* Re: [PATCH] uuidd: use pkg-config to find systemd-deamon support
  2014-01-07  8:58 ` Karel Zak
@ 2014-01-07  9:38   ` Thomas Bächler
  2014-01-07  9:54     ` Karel Zak
  0 siblings, 1 reply; 17+ messages in thread
From: Thomas Bächler @ 2014-01-07  9:38 UTC (permalink / raw)
  To: Karel Zak; +Cc: util-linux, Sami Kerola

[-- Attachment #1: Type: text/plain, Size: 884 bytes --]

Am 07.01.2014 09:58, schrieb Karel Zak:
> On Mon, Jan 06, 2014 at 11:39:08PM +0000, Sami Kerola wrote:
>>  configure.ac             |   3 +
>>  misc-utils/Makemodule.am |   3 +-
>>  misc-utils/sd-daemon.c   | 520 -----------------------------------------------
>>  misc-utils/sd-daemon.h   | 282 -------------------------
>>  misc-utils/uuidd.c       |   2 +-
> 
>  Seems like a good idea, unfortunately I'm not sure about it. 
>  
>  The systemd (udevd) depends on libblkid+libuuid, it would be 
>  probably better to keep util-linux without dependencies on systemd
>  libs otherwise we will introduce a cyclic dependence -- it could be
>  too tricky for some build systems and users :-) I'll think about it...
> 
>     Karel

We already have such a cyclic dependency, since udevd depends on
libblkid+libuuid, and findmnt (from util-linux) depends on libudev.



[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 901 bytes --]

^ permalink raw reply	[flat|nested] 17+ messages in thread

* Re: [PATCH] uuidd: use pkg-config to find systemd-deamon support
  2014-01-07  9:38   ` Thomas Bächler
@ 2014-01-07  9:54     ` Karel Zak
  2014-01-07  9:55       ` Thomas Bächler
  0 siblings, 1 reply; 17+ messages in thread
From: Karel Zak @ 2014-01-07  9:54 UTC (permalink / raw)
  To: Thomas Bächler; +Cc: util-linux, Sami Kerola

On Tue, Jan 07, 2014 at 10:38:16AM +0100, Thomas Bächler wrote:
> Am 07.01.2014 09:58, schrieb Karel Zak:
> > On Mon, Jan 06, 2014 at 11:39:08PM +0000, Sami Kerola wrote:
> >>  configure.ac             |   3 +
> >>  misc-utils/Makemodule.am |   3 +-
> >>  misc-utils/sd-daemon.c   | 520 -----------------------------------------------
> >>  misc-utils/sd-daemon.h   | 282 -------------------------
> >>  misc-utils/uuidd.c       |   2 +-
> > 
> >  Seems like a good idea, unfortunately I'm not sure about it. 
> >  
> >  The systemd (udevd) depends on libblkid+libuuid, it would be 
> >  probably better to keep util-linux without dependencies on systemd
> >  libs otherwise we will introduce a cyclic dependence -- it could be
> >  too tricky for some build systems and users :-) I'll think about it...
> > 
> >     Karel
> 
> We already have such a cyclic dependency, since udevd depends on
> libblkid+libuuid, and findmnt (from util-linux) depends on libudev.

 .. and nobody complains, good point :-) Thanks!

    Karel

-- 
 Karel Zak  <kzak@redhat.com>
 http://karelzak.blogspot.com

^ permalink raw reply	[flat|nested] 17+ messages in thread

* Re: [PATCH] uuidd: use pkg-config to find systemd-deamon support
  2014-01-07  9:54     ` Karel Zak
@ 2014-01-07  9:55       ` Thomas Bächler
  2014-01-07 10:19         ` Karel Zak
  0 siblings, 1 reply; 17+ messages in thread
From: Thomas Bächler @ 2014-01-07  9:55 UTC (permalink / raw)
  To: Karel Zak; +Cc: util-linux, Sami Kerola

[-- Attachment #1: Type: text/plain, Size: 1228 bytes --]

Am 07.01.2014 10:54, schrieb Karel Zak:
> On Tue, Jan 07, 2014 at 10:38:16AM +0100, Thomas Bächler wrote:
>> Am 07.01.2014 09:58, schrieb Karel Zak:
>>> On Mon, Jan 06, 2014 at 11:39:08PM +0000, Sami Kerola wrote:
>>>>  configure.ac             |   3 +
>>>>  misc-utils/Makemodule.am |   3 +-
>>>>  misc-utils/sd-daemon.c   | 520 -----------------------------------------------
>>>>  misc-utils/sd-daemon.h   | 282 -------------------------
>>>>  misc-utils/uuidd.c       |   2 +-
>>>
>>>  Seems like a good idea, unfortunately I'm not sure about it. 
>>>  
>>>  The systemd (udevd) depends on libblkid+libuuid, it would be 
>>>  probably better to keep util-linux without dependencies on systemd
>>>  libs otherwise we will introduce a cyclic dependence -- it could be
>>>  too tricky for some build systems and users :-) I'll think about it...
>>>
>>>     Karel
>>
>> We already have such a cyclic dependency, since udevd depends on
>> libblkid+libuuid, and findmnt (from util-linux) depends on libudev.
> 
>  .. and nobody complains, good point :-) Thanks!
> 
>     Karel

Actually, now that I know you care about these things, I am complaining.
It has caused me (minor) trouble once already.



[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 901 bytes --]

^ permalink raw reply	[flat|nested] 17+ messages in thread

* Re: [PATCH] uuidd: use pkg-config to find systemd-deamon support
  2014-01-07  9:55       ` Thomas Bächler
@ 2014-01-07 10:19         ` Karel Zak
  2014-01-07 18:44           ` Mike Frysinger
  0 siblings, 1 reply; 17+ messages in thread
From: Karel Zak @ 2014-01-07 10:19 UTC (permalink / raw)
  To: Thomas Bächler; +Cc: util-linux, Sami Kerola

On Tue, Jan 07, 2014 at 10:55:35AM +0100, Thomas Bächler wrote:
> >> We already have such a cyclic dependency, since udevd depends on
> >> libblkid+libuuid, and findmnt (from util-linux) depends on libudev.
> > 
> >  .. and nobody complains, good point :-) Thanks!
> > 
> >     Karel
> 
> Actually, now that I know you care about these things, I am complaining.

 :-)))

> It has caused me (minor) trouble once already.

 It's possible (first time) to compile util-linux without libudev,
 compile systemd/udevd and then again recompile util-linux with
 libudev.

 IMHO libudev is very useful for findmnt and lsblk to get info about
 devices.

 Note that uuidd is absolutely unnecessary if you don't have huge SAP
 installation (-- it means unnecessary for 99% of the all Linux users).

    Karel

-- 
 Karel Zak  <kzak@redhat.com>
 http://karelzak.blogspot.com

^ permalink raw reply	[flat|nested] 17+ messages in thread

* Re: [PATCH] uuidd: use pkg-config to find systemd-deamon support
  2014-01-06 23:39 [PATCH] uuidd: use pkg-config to find systemd-deamon support Sami Kerola
  2014-01-07  8:58 ` Karel Zak
@ 2014-01-07 11:20 ` Karel Zak
  1 sibling, 0 replies; 17+ messages in thread
From: Karel Zak @ 2014-01-07 11:20 UTC (permalink / raw)
  To: Sami Kerola; +Cc: util-linux

On Mon, Jan 06, 2014 at 11:39:08PM +0000, Sami Kerola wrote:
>  configure.ac             |   3 +
>  misc-utils/Makemodule.am |   3 +-
>  misc-utils/sd-daemon.c   | 520 -----------------------------------------------
>  misc-utils/sd-daemon.h   | 282 -------------------------
>  misc-utils/uuidd.c       |   2 +-
>  5 files changed, 5 insertions(+), 805 deletions(-)
>  delete mode 100644 misc-utils/sd-daemon.c
>  delete mode 100644 misc-utils/sd-daemon.h

 Applied with some minor changes, thanks!

    Karel

-- 
 Karel Zak  <kzak@redhat.com>
 http://karelzak.blogspot.com

^ permalink raw reply	[flat|nested] 17+ messages in thread

* Re: [PATCH] uuidd: use pkg-config to find systemd-deamon support
  2014-01-07 10:19         ` Karel Zak
@ 2014-01-07 18:44           ` Mike Frysinger
  2014-01-07 18:51             ` Dave Reisner
  2014-01-08  9:21             ` Karel Zak
  0 siblings, 2 replies; 17+ messages in thread
From: Mike Frysinger @ 2014-01-07 18:44 UTC (permalink / raw)
  To: Karel Zak; +Cc: Thomas Bächler, util-linux, Sami Kerola

[-- Attachment #1: Type: Text/Plain, Size: 1382 bytes --]

On Tuesday 07 January 2014 05:19:35 Karel Zak wrote:
> On Tue, Jan 07, 2014 at 10:55:35AM +0100, Thomas Bächler wrote:
> > >> We already have such a cyclic dependency, since udevd depends on
> > >> libblkid+libuuid, and findmnt (from util-linux) depends on libudev.
> > >> 
> > >  .. and nobody complains, good point :-) Thanks!
> > >  
> > >     Karel
> > 
> > Actually, now that I know you care about these things, I am complaining.
> > 
>  :-)))
>  :
> > It has caused me (minor) trouble once already.
> 
>  It's possible (first time) to compile util-linux without libudev,
>  compile systemd/udevd and then again recompile util-linux with
>  libudev.
> 
>  IMHO libudev is very useful for findmnt and lsblk to get info about
>  devices.
> 
>  Note that uuidd is absolutely unnecessary if you don't have huge SAP
>  installation (-- it means unnecessary for 99% of the all Linux users).

on a semi-related matter, it'd be nice if we could easily build util-linux 
such that we only get the libraries.  the plethora of configure flags that 
control utils is a nightmare to manage, and i'm not sure there's a flag for 
every utility (i haven't checked myself, so i could easily be wrong).

if that were in play, the circular loop would be easy to address -- build 
util-linux once just for its libs, then udev/systemd, then util-linux again.
-mike

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

^ permalink raw reply	[flat|nested] 17+ messages in thread

* Re: [PATCH] uuidd: use pkg-config to find systemd-deamon support
  2014-01-07 18:44           ` Mike Frysinger
@ 2014-01-07 18:51             ` Dave Reisner
  2014-01-07 19:57               ` Mike Frysinger
  2014-01-08  9:21             ` Karel Zak
  1 sibling, 1 reply; 17+ messages in thread
From: Dave Reisner @ 2014-01-07 18:51 UTC (permalink / raw)
  To: Mike Frysinger; +Cc: Karel Zak, Thomas Bächler, util-linux, Sami Kerola

On Tue, Jan 07, 2014 at 01:44:07PM -0500, Mike Frysinger wrote:
> On Tuesday 07 January 2014 05:19:35 Karel Zak wrote:
> > On Tue, Jan 07, 2014 at 10:55:35AM +0100, Thomas Bächler wrote:
> > > >> We already have such a cyclic dependency, since udevd depends on
> > > >> libblkid+libuuid, and findmnt (from util-linux) depends on libudev.
> > > >> 
> > > >  .. and nobody complains, good point :-) Thanks!
> > > >  
> > > >     Karel
> > > 
> > > Actually, now that I know you care about these things, I am complaining.
> > > 
> >  :-)))
> >  :
> > > It has caused me (minor) trouble once already.
> > 
> >  It's possible (first time) to compile util-linux without libudev,
> >  compile systemd/udevd and then again recompile util-linux with
> >  libudev.
> > 
> >  IMHO libudev is very useful for findmnt and lsblk to get info about
> >  devices.
> > 
> >  Note that uuidd is absolutely unnecessary if you don't have huge SAP
> >  installation (-- it means unnecessary for 99% of the all Linux users).
> 
> on a semi-related matter, it'd be nice if we could easily build util-linux 
> such that we only get the libraries.  the plethora of configure flags that 
> control utils is a nightmare to manage, and i'm not sure there's a flag for 
> every utility (i haven't checked myself, so i could easily be wrong).
> 
> if that were in play, the circular loop would be easy to address -- build 
> util-linux once just for its libs, then udev/systemd, then util-linux again.
> -mike

We do this in Arch already to provide a 32bit libraries on 64bit
systems:

https://projects.archlinux.org/svntogit/community.git/tree/trunk/PKGBUILD?h=packages/lib32-util-linux

A quick grep through the git repo shows that I could get rid of the
--without-ncurses flag, too.

^ permalink raw reply	[flat|nested] 17+ messages in thread

* Re: [PATCH] uuidd: use pkg-config to find systemd-deamon support
  2014-01-07 18:51             ` Dave Reisner
@ 2014-01-07 19:57               ` Mike Frysinger
  0 siblings, 0 replies; 17+ messages in thread
From: Mike Frysinger @ 2014-01-07 19:57 UTC (permalink / raw)
  To: Dave Reisner; +Cc: Karel Zak, Thomas Bächler, util-linux, Sami Kerola

[-- Attachment #1: Type: Text/Plain, Size: 2231 bytes --]

On Tuesday 07 January 2014 13:51:54 Dave Reisner wrote:
> On Tue, Jan 07, 2014 at 01:44:07PM -0500, Mike Frysinger wrote:
> > On Tuesday 07 January 2014 05:19:35 Karel Zak wrote:
> > > On Tue, Jan 07, 2014 at 10:55:35AM +0100, Thomas Bächler wrote:
> > > > >> We already have such a cyclic dependency, since udevd depends on
> > > > >> libblkid+libuuid, and findmnt (from util-linux) depends on
> > > > >> libudev.
> > > > >> 
> > > > >  .. and nobody complains, good point :-) Thanks!
> > > > >  
> > > > >     Karel
> > > > 
> > > > Actually, now that I know you care about these things, I am
> > > > complaining.
> > > > 
> > >  :-)))
> > >  :
> > > > It has caused me (minor) trouble once already.
> > >  
> > >  It's possible (first time) to compile util-linux without libudev,
> > >  compile systemd/udevd and then again recompile util-linux with
> > >  libudev.
> > >  
> > >  IMHO libudev is very useful for findmnt and lsblk to get info about
> > >  devices.
> > >  
> > >  Note that uuidd is absolutely unnecessary if you don't have huge SAP
> > >  installation (-- it means unnecessary for 99% of the all Linux users).
> > 
> > on a semi-related matter, it'd be nice if we could easily build
> > util-linux such that we only get the libraries.  the plethora of
> > configure flags that control utils is a nightmare to manage, and i'm not
> > sure there's a flag for every utility (i haven't checked myself, so i
> > could easily be wrong).
> > 
> > if that were in play, the circular loop would be easy to address -- build
> > util-linux once just for its libs, then udev/systemd, then util-linux
> > again.
> 
> We do this in Arch already to provide a 32bit libraries on 64bit
> systems:
> 
> https://projects.archlinux.org/svntogit/community.git/tree/trunk/PKGBUILD?h
> =packages/lib32-util-linux

thanks, that illustrates my point exactly :).  those make targets are esoteric 
if not internal to automake.  relying on them externally sucks.

> A quick grep through the git repo shows that I could get rid of the
> --without-ncurses flag, too.

pretty sure there are a bunch of configure checks that apply only to utils.  
pam and ncurses both come to mind.
-mike

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

^ permalink raw reply	[flat|nested] 17+ messages in thread

* Re: [PATCH] uuidd: use pkg-config to find systemd-deamon support
  2014-01-07 18:44           ` Mike Frysinger
  2014-01-07 18:51             ` Dave Reisner
@ 2014-01-08  9:21             ` Karel Zak
  2014-01-08 20:37               ` Sami Kerola
  1 sibling, 1 reply; 17+ messages in thread
From: Karel Zak @ 2014-01-08  9:21 UTC (permalink / raw)
  To: Mike Frysinger; +Cc: Thomas Bächler, util-linux, Sami Kerola

On Tue, Jan 07, 2014 at 01:44:07PM -0500, Mike Frysinger wrote:
> On Tuesday 07 January 2014 05:19:35 Karel Zak wrote:
> > On Tue, Jan 07, 2014 at 10:55:35AM +0100, Thomas Bächler wrote:
> > > >> We already have such a cyclic dependency, since udevd depends on
> > > >> libblkid+libuuid, and findmnt (from util-linux) depends on libudev.
> > > >> 
> > > >  .. and nobody complains, good point :-) Thanks!
> > > >  
> > > >     Karel
> > > 
> > > Actually, now that I know you care about these things, I am complaining.
> > > 
> >  :-)))
> >  :
> > > It has caused me (minor) trouble once already.
> > 
> >  It's possible (first time) to compile util-linux without libudev,
> >  compile systemd/udevd and then again recompile util-linux with
> >  libudev.
> > 
> >  IMHO libudev is very useful for findmnt and lsblk to get info about
> >  devices.
> > 
> >  Note that uuidd is absolutely unnecessary if you don't have huge SAP
> >  installation (-- it means unnecessary for 99% of the all Linux users).
> 
> on a semi-related matter, it'd be nice if we could easily build util-linux 
> such that we only get the libraries.  the plethora of configure flags that 
> control utils is a nightmare to manage, and i'm not sure there's a flag for 
> every utility (i haven't checked myself, so i could easily be wrong).

not yet ;-)

> if that were in play, the circular loop would be easy to address -- build 
> util-linux once just for its libs, then udev/systemd, then util-linux again.

Yep, I know about the pain... I guess it would be enough to have
something like

 ./configure --disable-all-builds --enable-libblkid

where --enable/disable-all-builds controls the default of the
enable_<name>= variables (the current default is usually "check")
in configure.

I'll try to implement it later (I'm working on new cfdisk now) -- or
maybe some volunteer...? :-)

    Karel

-- 
 Karel Zak  <kzak@redhat.com>
 http://karelzak.blogspot.com

^ permalink raw reply	[flat|nested] 17+ messages in thread

* Re: [PATCH] uuidd: use pkg-config to find systemd-deamon support
  2014-01-08  9:21             ` Karel Zak
@ 2014-01-08 20:37               ` Sami Kerola
  2014-01-08 20:44                 ` Mike Frysinger
  0 siblings, 1 reply; 17+ messages in thread
From: Sami Kerola @ 2014-01-08 20:37 UTC (permalink / raw)
  To: Karel Zak; +Cc: Mike Frysinger, Thomas Bächler, util-linux

On 8 January 2014 09:21, Karel Zak <kzak@redhat.com> wrote:
> On Tue, Jan 07, 2014 at 01:44:07PM -0500, Mike Frysinger wrote:
>> On Tuesday 07 January 2014 05:19:35 Karel Zak wrote:
>> > On Tue, Jan 07, 2014 at 10:55:35AM +0100, Thomas Bächler wrote:
>> > > >> We already have such a cyclic dependency, since udevd depends on
>> > > >> libblkid+libuuid, and findmnt (from util-linux) depends on libudev.
>> > > >>
>> > > >  .. and nobody complains, good point :-) Thanks!
>> > > >
>> > > >     Karel
>> > >
>> > > Actually, now that I know you care about these things, I am complaining.
>> > >
>> >  :-)))
>> >  :
>> > > It has caused me (minor) trouble once already.
>> >
>> >  It's possible (first time) to compile util-linux without libudev,
>> >  compile systemd/udevd and then again recompile util-linux with
>> >  libudev.
>> >
>> >  IMHO libudev is very useful for findmnt and lsblk to get info about
>> >  devices.
>> >
>> >  Note that uuidd is absolutely unnecessary if you don't have huge SAP
>> >  installation (-- it means unnecessary for 99% of the all Linux users).
>>
>> on a semi-related matter, it'd be nice if we could easily build util-linux
>> such that we only get the libraries.  the plethora of configure flags that
>> control utils is a nightmare to manage, and i'm not sure there's a flag for
>> every utility (i haven't checked myself, so i could easily be wrong).
>
> not yet ;-)
>
>> if that were in play, the circular loop would be easy to address -- build
>> util-linux once just for its libs, then udev/systemd, then util-linux again.
>
> Yep, I know about the pain... I guess it would be enough to have
> something like
>
>  ./configure --disable-all-builds --enable-libblkid
>
> where --enable/disable-all-builds controls the default of the
> enable_<name>= variables (the current default is usually "check")
> in configure.
>
> I'll try to implement it later (I'm working on new cfdisk now) -- or
> maybe some volunteer...? :-)

I could try to fiddle around with autotools in next days/weeks.

The one thing I am thinking to change immediately is
's/--disable-all-builds/--enable-minimal/'.  Unfortunately, due I'm just
a random contributor without distribution packaging experience at all I
will need few hints from real pros.  Most important question being what
the minimal should include?  I did think that option could be
--enable-libaries-only, but is that too minimal to be practical?

p.s.  Karel, the consider the cal(1) enhancements being temporarily
postponed.  Well, they weren't import anyway, so IMHO not a big loss.

-- 
Sami Kerola
http://www.iki.fi/kerolasa/

^ permalink raw reply	[flat|nested] 17+ messages in thread

* Re: [PATCH] uuidd: use pkg-config to find systemd-deamon support
  2014-01-08 20:37               ` Sami Kerola
@ 2014-01-08 20:44                 ` Mike Frysinger
  2014-01-08 22:28                   ` Sami Kerola
  0 siblings, 1 reply; 17+ messages in thread
From: Mike Frysinger @ 2014-01-08 20:44 UTC (permalink / raw)
  To: kerolasa; +Cc: Karel Zak, Thomas Bächler, util-linux

[-- Attachment #1: Type: Text/Plain, Size: 2933 bytes --]

On Wednesday 08 January 2014 15:37:30 Sami Kerola wrote:
> On 8 January 2014 09:21, Karel Zak <kzak@redhat.com> wrote:
> > On Tue, Jan 07, 2014 at 01:44:07PM -0500, Mike Frysinger wrote:
> >> On Tuesday 07 January 2014 05:19:35 Karel Zak wrote:
> >> > On Tue, Jan 07, 2014 at 10:55:35AM +0100, Thomas Bächler wrote:
> >> > > >> We already have such a cyclic dependency, since udevd depends on
> >> > > >> libblkid+libuuid, and findmnt (from util-linux) depends on
> >> > > >> libudev.
> >> > > >> 
> >> > > >  .. and nobody complains, good point :-) Thanks!
> >> > > >  
> >> > > >     Karel
> >> > > 
> >> > > Actually, now that I know you care about these things, I am
> >> > > complaining.
> >> > > 
> >> >  :-)))
> >> >  :
> >> > > It has caused me (minor) trouble once already.
> >> >  
> >> >  It's possible (first time) to compile util-linux without libudev,
> >> >  compile systemd/udevd and then again recompile util-linux with
> >> >  libudev.
> >> >  
> >> >  IMHO libudev is very useful for findmnt and lsblk to get info about
> >> >  devices.
> >> >  
> >> >  Note that uuidd is absolutely unnecessary if you don't have huge SAP
> >> >  installation (-- it means unnecessary for 99% of the all Linux
> >> >  users).
> >> 
> >> on a semi-related matter, it'd be nice if we could easily build
> >> util-linux such that we only get the libraries.  the plethora of
> >> configure flags that control utils is a nightmare to manage, and i'm
> >> not sure there's a flag for every utility (i haven't checked myself, so
> >> i could easily be wrong).
> > 
> > not yet ;-)
> > 
> >> if that were in play, the circular loop would be easy to address --
> >> build util-linux once just for its libs, then udev/systemd, then
> >> util-linux again.
> > 
> > Yep, I know about the pain... I guess it would be enough to have
> > something like
> > 
> >  ./configure --disable-all-builds --enable-libblkid
> > 
> > where --enable/disable-all-builds controls the default of the
> > enable_<name>= variables (the current default is usually "check")
> > in configure.
> > 
> > I'll try to implement it later (I'm working on new cfdisk now) -- or
> > maybe some volunteer...? :-)
> 
> I could try to fiddle around with autotools in next days/weeks.
> 
> The one thing I am thinking to change immediately is
> 's/--disable-all-builds/--enable-minimal/'.  Unfortunately, due I'm just
> a random contributor without distribution packaging experience at all I
> will need few hints from real pros.  Most important question being what
> the minimal should include?  I did think that option could be
> --enable-libaries-only, but is that too minimal to be practical?

i don't think general flags like "minimal" are useful.  a knob that'd disable 
all binaries would though as then we could use the existing library knobs to 
select the exact ones we want to build/install.
-mike

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

^ permalink raw reply	[flat|nested] 17+ messages in thread

* Re: [PATCH] uuidd: use pkg-config to find systemd-deamon support
  2014-01-08 20:44                 ` Mike Frysinger
@ 2014-01-08 22:28                   ` Sami Kerola
  2014-01-09  8:54                     ` Karel Zak
  2014-01-09 16:56                     ` Mike Frysinger
  0 siblings, 2 replies; 17+ messages in thread
From: Sami Kerola @ 2014-01-08 22:28 UTC (permalink / raw)
  To: Mike Frysinger; +Cc: util-linux, Sami Kerola, Karel Zak, Thomas Bächler

Hi Mike, and others,

So the change that's needed could be something like

--->8----
From: Sami Kerola <kerolasa@iki.fi>
Date: Wed, 8 Jan 2014 22:16:05 +0000
Subject: [PATCH] build-sys: add ./configure --enable-libraries-only option

The option is meant to help distributors to bootstrap systems out of
scratch.

Reference: http://www.spinics.net/lists/util-linux-ng/index.html#08695
Signed-off-by: Sami Kerola <kerolasa@iki.fi>
---
 Makefile.am               | 12 ++++++++++++
 configure.ac              | 10 +++++++++-
 login-utils/Makemodule.am | 11 -----------
 3 files changed, 21 insertions(+), 12 deletions(-)

diff --git a/Makefile.am b/Makefile.am
index 17f4c33..24a7212 100644
--- a/Makefile.am
+++ b/Makefile.am
@@ -79,6 +79,7 @@ include libblkid/Makemodule.am
 include libmount/Makemodule.am
 include libfdisk/Makemodule.am
 
+if HAVE_COMMANDS
 include schedutils/Makemodule.am
 include text-utils/Makemodule.am
 include term-utils/Makemodule.am
@@ -92,6 +93,7 @@ include fdisks/Makemodule.am
 include bash-completion/Makemodule.am
 
 include tests/Makemodule.am
+endif	# HAVE_COMMANDS
 
 #
 # Don't rely on configure.ac AC_CONFIG_FILES for install paths.
@@ -191,6 +193,16 @@ checksmatch:
 changelog: $(CHANGELOG_FILE)
 
 install-exec-hook: $(INSTALL_EXEC_HOOKS)
+if HAVE_COMMANDS
+if BUILD_SU
+if MAKEINSTALL_DO_SETUID
+	chmod 4755 $(DESTDIR)$(bindir)/su
+endif
+endif	# BUILD_SU
+if BUILD_VIPW
+	cd $(DESTDIR)$(usrsbin_execdir) && ln -sf vipw vigr
+endif
+endif	# HAVE_COMMANDS
 
 uninstall-hook: $(UNINSTALL_HOOKS)
 
diff --git a/configure.ac b/configure.ac
index 6bf5048..d08589e 100644
--- a/configure.ac
+++ b/configure.ac
@@ -704,6 +704,11 @@ AS_IF([test "x$with_utempter" = xyes], [
   AM_CONDITIONAL([HAVE_UTEMPTER], [false])
 ])
 
+AC_ARG_ENABLE([binaries],
+  AS_HELP_STRING([--disable-binaries], [build only libraries]),
+  [], [enable_libraries_only=yes]
+)
+AM_CONDITIONAL(HAVE_COMMANDS, test "x$enable_libraries_only" = xyes)
 
 AC_ARG_ENABLE([most-builds],
   AS_HELP_STRING([--enable-most-builds], [build everything other than experimental code]),
@@ -1570,7 +1575,10 @@ libblkid/src/blkid.h
 libmount/docs/Makefile
 libmount/docs/version.xml
 libmount/src/libmount.h
-po/Makefile.in
+])
+
+AS_IF([test "x$enable_libraries_only" = xyes], [
+  AC_CONFIG_FILES([po/Makefile.in])
 ])
 
 AC_OUTPUT
diff --git a/login-utils/Makemodule.am b/login-utils/Makemodule.am
index aca028a..c57aca7 100644
--- a/login-utils/Makemodule.am
+++ b/login-utils/Makemodule.am
@@ -196,14 +196,3 @@ test_logindefs_SOURCES = \
 	login-utils/logindefs.c \
 	login-utils/logindefs.h
 test_logindefs_CPPFLAGS = -DTEST_PROGRAM $(AM_CPPFLAGS)
-
-
-install-exec-hook:
-if BUILD_SU
-if MAKEINSTALL_DO_SETUID
-	chmod 4755 $(DESTDIR)$(bindir)/su
-endif
-endif
-if BUILD_VIPW
-	cd $(DESTDIR)$(usrsbin_execdir) && ln -sf vipw vigr
-endif
-- 
1.8.5.2

^ permalink raw reply related	[flat|nested] 17+ messages in thread

* Re: [PATCH] uuidd: use pkg-config to find systemd-deamon support
  2014-01-08 22:28                   ` Sami Kerola
@ 2014-01-09  8:54                     ` Karel Zak
  2014-01-09 16:56                     ` Mike Frysinger
  1 sibling, 0 replies; 17+ messages in thread
From: Karel Zak @ 2014-01-09  8:54 UTC (permalink / raw)
  To: Sami Kerola; +Cc: Mike Frysinger, util-linux, Thomas Bächler

On Wed, Jan 08, 2014 at 10:28:06PM +0000, Sami Kerola wrote:
> So the change that's needed could be something like

It would be really better to have --disable-all-builds, then we can
support things like:

    --disable-all-builds --enable-unshare

how many utils are not covered by any enable_<name>= ? I think all we
need is to just add AC_ARG_ENABLE() to all utils. 

(Well, for stuff like schedutils we have AC_ARG_ENABLE([schedutils])
 and it's enough, it's unnecessary to have AC_ARG_ENABLE for taskset,
 ionice, ..) 

    Karel

-- 
 Karel Zak  <kzak@redhat.com>
 http://karelzak.blogspot.com

^ permalink raw reply	[flat|nested] 17+ messages in thread

* Re: [PATCH] uuidd: use pkg-config to find systemd-deamon support
  2014-01-08 22:28                   ` Sami Kerola
  2014-01-09  8:54                     ` Karel Zak
@ 2014-01-09 16:56                     ` Mike Frysinger
  2014-01-10 14:29                       ` Sami Kerola
  1 sibling, 1 reply; 17+ messages in thread
From: Mike Frysinger @ 2014-01-09 16:56 UTC (permalink / raw)
  To: Sami Kerola; +Cc: util-linux, Karel Zak, Thomas Bächler

[-- Attachment #1: Type: Text/Plain, Size: 1605 bytes --]

On Wednesday 08 January 2014 17:28:06 Sami Kerola wrote:
> --- a/Makefile.am
> +++ b/Makefile.am
> @@ -79,6 +79,7 @@ include libblkid/Makemodule.am
>  include libmount/Makemodule.am
>  include libfdisk/Makemodule.am
> 
> +if HAVE_COMMANDS
>  include schedutils/Makemodule.am
>  include text-utils/Makemodule.am
>  include term-utils/Makemodule.am
> @@ -92,6 +93,7 @@ include fdisks/Makemodule.am
>  include bash-completion/Makemodule.am
> 
>  include tests/Makemodule.am
> +endif	# HAVE_COMMANDS

aren't there some tests for the libs ?

> --- a/configure.ac
> +++ b/configure.ac
> @@ -704,6 +704,11 @@ AS_IF([test "x$with_utempter" = xyes], [
>    AM_CONDITIONAL([HAVE_UTEMPTER], [false])
>  ])
> 
> +AC_ARG_ENABLE([binaries],
> +  AS_HELP_STRING([--disable-binaries], [build only libraries]),
> +  [], [enable_libraries_only=yes]
> +)
> +AM_CONDITIONAL(HAVE_COMMANDS, test "x$enable_libraries_only" = xyes)

i'm not sure about this inversion of naming.  why can't it just use 
enable_binaries everywhere ?

> --- a/login-utils/Makemodule.am
> +++ b/login-utils/Makemodule.am
> @@ -196,14 +196,3 @@ test_logindefs_SOURCES = \
>  	login-utils/logindefs.c \
>  	login-utils/logindefs.h
>  test_logindefs_CPPFLAGS = -DTEST_PROGRAM $(AM_CPPFLAGS)
> -
> -
> -install-exec-hook:
> -if BUILD_SU
> -if MAKEINSTALL_DO_SETUID
> -	chmod 4755 $(DESTDIR)$(bindir)/su
> -endif
> -endif
> -if BUILD_VIPW
> -	cd $(DESTDIR)$(usrsbin_execdir) && ln -sf vipw vigr
> -endif

why does this need to move ?  this module isn't included when commands are 
disabled ...
-mike

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

^ permalink raw reply	[flat|nested] 17+ messages in thread

* Re: [PATCH] uuidd: use pkg-config to find systemd-deamon support
  2014-01-09 16:56                     ` Mike Frysinger
@ 2014-01-10 14:29                       ` Sami Kerola
  0 siblings, 0 replies; 17+ messages in thread
From: Sami Kerola @ 2014-01-10 14:29 UTC (permalink / raw)
  To: Mike Frysinger; +Cc: util-linux, Karel Zak, Thomas Bächler

On 9 January 2014 16:56, Mike Frysinger <vapier@gentoo.org> wrote:
> On Wednesday 08 January 2014 17:28:06 Sami Kerola wrote:
>>  include tests/Makemodule.am
>> +endif        # HAVE_COMMANDS
>
> aren't there some tests for the libs ?

There are some. Taking in consideration Karel's proposal to have
--disable-all-builds that could be combined with
--enable-<per-subdirectory> the tests cannot be skipped in build root
Makefile.am level.

>> +AM_CONDITIONAL(HAVE_COMMANDS, test "x$enable_libraries_only" = xyes)
>
> i'm not sure about this inversion of naming.  why can't it just use
> enable_binaries everywhere ?

Yes, I agree the names could have been better but I did not bother
making the very first hastily wrote patch 100% correct, which is good
since --enable-libraries-only or --disable-commands does not seem to
be what is needed.

>> --- a/login-utils/Makemodule.am
>> +++ b/login-utils/Makemodule.am
>> -install-exec-hook:
>> -if BUILD_SU
>> -if MAKEINSTALL_DO_SETUID
>> -     chmod 4755 $(DESTDIR)$(bindir)/su
>> -endif
>> -endif
>> -if BUILD_VIPW
>> -     cd $(DESTDIR)$(usrsbin_execdir) && ln -sf vipw vigr
>> -endif
>
> why does this need to move ?  this module isn't included when commands are
> disabled ...

automake gave warning the install-exec-hook was if'd out. I think the
sudo suid & vipw will need to use INSTALL_EXEC_HOOKS, and this is
patch has to precede --disable-all-builds change.

-- 
Sami Kerola
http://www.iki.fi/kerolasa/

^ permalink raw reply	[flat|nested] 17+ messages in thread

end of thread, other threads:[~2014-01-10 14:29 UTC | newest]

Thread overview: 17+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2014-01-06 23:39 [PATCH] uuidd: use pkg-config to find systemd-deamon support Sami Kerola
2014-01-07  8:58 ` Karel Zak
2014-01-07  9:38   ` Thomas Bächler
2014-01-07  9:54     ` Karel Zak
2014-01-07  9:55       ` Thomas Bächler
2014-01-07 10:19         ` Karel Zak
2014-01-07 18:44           ` Mike Frysinger
2014-01-07 18:51             ` Dave Reisner
2014-01-07 19:57               ` Mike Frysinger
2014-01-08  9:21             ` Karel Zak
2014-01-08 20:37               ` Sami Kerola
2014-01-08 20:44                 ` Mike Frysinger
2014-01-08 22:28                   ` Sami Kerola
2014-01-09  8:54                     ` Karel Zak
2014-01-09 16:56                     ` Mike Frysinger
2014-01-10 14:29                       ` Sami Kerola
2014-01-07 11:20 ` Karel Zak

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox