Re: [RFC PATCH 1/5] maintenance: package systemd units

From: Max Gautier <mg@max.gautier.name>
Date: Mar 21, 2024, 13:38 UTC
Message-ID: <Zfw4XNJdZqgZhvOv@framework>
In-Reply-To: <ZfwqCv889UdI0mU6@tanuki>

On Thu, Mar 21, 2024 at 01:37:30PM +0100, Patrick Steinhardt wrote:

Show 49 quoted lines

> On Mon, Mar 18, 2024 at 04:31:15PM +0100, Max Gautier wrote:
> 
> It would be great to document _why_ we want to package the systemd units
> alongside with Git.
> 
> > Signed-off-by: Max Gautier <mg@max.gautier.name>
> > ---
> >  Makefile                              |  4 ++++
> >  systemd/user/git-maintenance@.service | 16 ++++++++++++++++
> >  systemd/user/git-maintenance@.timer   |  9 +++++++++
> >  3 files changed, 29 insertions(+)
> >  create mode 100644 systemd/user/git-maintenance@.service
> >  create mode 100644 systemd/user/git-maintenance@.timer
> > 
> > diff --git a/Makefile b/Makefile
> > index 4e255c81f2..276b4373c6 100644
> > --- a/Makefile
> > +++ b/Makefile
> > @@ -619,6 +619,7 @@ htmldir = $(prefix)/share/doc/git-doc
> >  ETC_GITCONFIG = $(sysconfdir)/gitconfig
> >  ETC_GITATTRIBUTES = $(sysconfdir)/gitattributes
> >  lib = lib
> > +libdir = $(prefix)/lib
> >  # DESTDIR =
> >  pathsep = :
> >  
> > @@ -1328,6 +1329,8 @@ BUILTIN_OBJS += builtin/verify-tag.o
> >  BUILTIN_OBJS += builtin/worktree.o
> >  BUILTIN_OBJS += builtin/write-tree.o
> >  
> > +SYSTEMD_USER_UNITS := $(wildcard systemd/user/*)
> > +
> >  # THIRD_PARTY_SOURCES is a list of patterns compatible with the
> >  # $(filter) and $(filter-out) family of functions. They specify source
> >  # files which are taken from some third-party source where we want to be
> > @@ -3469,6 +3472,7 @@ install: all
> >  	$(INSTALL) -m 644 $(SCRIPT_LIB) '$(DESTDIR_SQ)$(gitexec_instdir_SQ)'
> >  	$(INSTALL) $(INSTALL_STRIP) $(install_bindir_xprograms) '$(DESTDIR_SQ)$(bindir_SQ)'
> >  	$(INSTALL) $(BINDIR_PROGRAMS_NO_X) '$(DESTDIR_SQ)$(bindir_SQ)'
> > +	$(INSTALL) -Dm 644 -t '$(DESTDIR_SQ)$(libdir)/systemd/user' $(SYSTEMD_USER_UNITS)
> 
> I wonder whether we want to unconditionally install those units. Many of
> the platforms that we support don't even have systemd available, so
> certainly it wouldn't make any sense to install it on those platforms.
> 
> Assuming that this is something we want in the first place I thus think
> that we should at least make this conditional and add some platform
> specific quirk to "config.mak.uname".
>

We probably want that (conditional install) but I'm not sure where that should go ; I'm not super familiar with autoconf.

I just noticed than man 7 daemon (shipped by systemd) propose the following snippet for installing systemd system services (should be easy enough to adapt for user ervices, I think):

PKG_PROG_PKG_CONFIG()
AC_ARG_WITH([systemdsystemunitdir],
    [AS_HELP_STRING([--with-systemdsystemunitdir=DIR], [Directory for systemd service files])],,
    [with_systemdsystemunitdir=auto])
AS_IF([test "x$with_systemdsystemunitdir" = "xyes" -o "x$with_systemdsystemunitdir" = "xauto"], [
    def_systemdsystemunitdir=$($PKG_CONFIG --variable=systemdsystemunitdir systemd)

    AS_IF([test "x$def_systemdsystemunitdir" = "x"],
  [AS_IF([test "x$with_systemdsystemunitdir" = "xyes"],
   [AC_MSG_ERROR([systemd support requested but pkg-config unable to query systemd package])])
   with_systemdsystemunitdir=no],
  [with_systemdsystemunitdir="$def_systemdsystemunitdir"])])
AS_IF([test "x$with_systemdsystemunitdir" != "xno"],
     [AC_SUBST([systemdsystemunitdir], [$with_systemdsystemunitdir])])
AM_CONDITIONAL([HAVE_SYSTEMD], [test "x$with_systemdsystemunitdir" != "xno"])

Would something like that work ?

-- 
Max Gautier