[PATCH BlueZ 0/3] build: Fix distcheck while installing org.bluez.obex.service with --disable-systemd

[PATCH BlueZ 0/3] build: Fix distcheck while installing org.bluez.obex.service with --disable-systemd

[Alfred Wingate]

Pkg-config variables give absolute paths with system prefixes, causing problems
for distcheck which wants to install everything in its own prefix. To cope with
this lets give distcheck its own configure arguments with its own prefix variable
already included.

https://www.gnu.org/software/automake/manual/html_node/DISTCHECK_005fCONFIGURE_005fFLAGS.html

This allows reverting the previous workaround that stopped installing the obex
dbus service on non systemd targets.

https://github.com/bluez/bluez/issues/806

Alfred Wingate (3):
  build: Support setting the cups_serverbin directory
  build: Fix distcheck by hardcoding non absolute paths
  build: obexd: Revert "Fix make distcheck"

 Makefile.am    | 8 ++++++++
 Makefile.obexd | 7 ++++---
 configure.ac   | 4 +++-
 3 files changed, 15 insertions(+), 4 deletions(-)


base-commit: 9faf1244a93217775e31ed5ef3b656d4430b63e1
-- 
2.51.1

[PATCH BlueZ 1/3] build: Support setting the cups_serverbin directory

[Alfred Wingate]

This will be needed in the follow up change that tries to fix distcheck
properly.

Signed-off-by: Alfred Wingate <parona@protonmail.com>
---
 configure.ac | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/configure.ac b/configure.ac
index ebb01c4ae..adb089e91 100644
--- a/configure.ac
+++ b/configure.ac
@@ -264,7 +264,9 @@ AC_SUBST(UDEV_DIR, [${path_udevdir}])
 AC_ARG_ENABLE(cups, AS_HELP_STRING([--disable-cups],
                 [disable CUPS printer support]), [enable_cups=${enableval}])
 AM_CONDITIONAL(CUPS, test "${enable_cups}" != "no")
-if (test "${enable_cups}" != "no"); then
+AC_ARG_WITH([cups_serverbin], AS_HELP_STRING([--with-cups_serverbin=DIR],
+			[path to cups server binary directory]), [cups_serverbin=${withval}])
+if (test "${enable_cups}" != "no" && test -z "${cups_serverbin}"); then
 	AC_MSG_CHECKING([cups directory])
 	cups_serverbin=`$PKG_CONFIG cups --variable=cups_serverbin`
 	AC_MSG_RESULT([${cups_serverbin}])
-- 
2.51.1

[PATCH BlueZ 2/3] build: Fix distcheck by hardcoding non absolute paths

[Alfred Wingate]

To make sure it doesn't fail lets hardcode every path that could come
from pkg-config.
(I managed to hit issues from dbus **and** cups as I have them installed)

You can learn more about this type of issue here.
https://bugzilla.redhat.com/show_bug.cgi?id=1123963

See-Also: 9cc8b97b8 ("build: obexd: Fix make distcheck")
Signed-off-by: Alfred Wingate <parona@protonmail.com>
---
 Makefile.am | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/Makefile.am b/Makefile.am
index 94f625db4..3e25c01dc 100644
--- a/Makefile.am
+++ b/Makefile.am
@@ -2,6 +2,14 @@
 AM_MAKEFLAGS = --no-print-directory
 AM_CPPFLAGS =
 
+# pkg-config gives absolute paths that do not work with distcheck that wants to use its own $(prefix)
+AM_DISTCHECK_CONFIGURE_FLAGS = --with-dbusconfdir=$${dc_install_base}/usr/share \
+			       --with-dbussystembusdir=$${dc_install_base}/usr/share/dbus-1/system-services \
+			       --with-dbussessionbusdir=$${dc_install_base}/usr/share/dbus-1/services \
+			       --with-systemdsystemunitdir=$${dc_install_base}/usr/lib/systemd/system \
+			       --with-systemduserunitdir=$${dc_install_base}/usr/lib/systemd/user \
+			       --with-cups_serverbin=$${dc_install_base}/usr/libexec/cups
+
 lib_LTLIBRARIES =
 
 noinst_LIBRARIES =
-- 
2.51.1

[PATCH BlueZ 3/3] build: obexd: Revert "Fix make distcheck"

[Alfred Wingate]

Bug: https://github.com/bluez/bluez/issues/806
Reverts: 9cc8b97b8 ("build: obexd: Fix make distcheck")
Signed-off-by: Alfred Wingate <parona@protonmail.com>
---
 Makefile.obexd | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/Makefile.obexd b/Makefile.obexd
index b59cfaf8f..7ad74e128 100644
--- a/Makefile.obexd
+++ b/Makefile.obexd
@@ -4,9 +4,6 @@ if OBEX
 if SYSTEMD
 systemduserunit_DATA += obexd/src/obex.service
 
-dbussessionbusdir = $(DBUS_SESSIONBUSDIR)
-dbussessionbus_DATA = obexd/src/org.bluez.obex.service
-
 obexd-add-service-symlink:
 	$(LN_S) -f obex.service $(DESTDIR)$(SYSTEMD_USERUNITDIR)/dbus-org.bluez.obex.service
 
@@ -17,6 +14,10 @@ obexd-add-service-symlink:
 obexd-remove-service-symlink:
 endif
 
+dbussessionbusdir = $(DBUS_SESSIONBUSDIR)
+dbussessionbus_DATA = obexd/src/org.bluez.obex.service
+
+
 obex_plugindir = $(libdir)/obex/plugins
 
 obexd_builtin_modules =
-- 
2.51.1

RE: build: Fix distcheck while installing org.bluez.obex.service with --disable-systemd

[bluez.test.bot]

[-- Attachment #1: Type: text/plain, Size: 1262 bytes --]

This is automated email and please do not reply to this email!

Dear submitter,

Thank you for submitting the patches to the linux bluetooth mailing list.
This is a CI test results with your patch series:
PW Link:https://patchwork.kernel.org/project/bluetooth/list/?series=1014817

---Test result---

Test Summary:
CheckPatch                    PENDING   0.24 seconds
GitLint                       PENDING   0.25 seconds
BuildEll                      PASS      20.07 seconds
BluezMake                     PASS      2657.25 seconds
MakeCheck                     PASS      19.99 seconds
MakeDistcheck                 PASS      185.42 seconds
CheckValgrind                 PASS      237.78 seconds
CheckSmatch                   PASS      311.08 seconds
bluezmakeextell               PASS      129.46 seconds
IncrementalBuild              PENDING   0.33 seconds
ScanBuild                     PASS      921.52 seconds

Details
##############################
Test: CheckPatch - PENDING
Desc: Run checkpatch.pl script
Output:

##############################
Test: GitLint - PENDING
Desc: Run gitlint
Output:

##############################
Test: IncrementalBuild - PENDING
Desc: Incremental build with the patches in the series
Output:



---
Regards,
Linux Bluetooth

Re: [PATCH BlueZ 2/3] build: Fix distcheck by hardcoding non absolute paths

[Luiz Augusto von Dentz]

Hi Alfred,

On Thu, Oct 23, 2025 at 1:06 AM Alfred Wingate <parona@protonmail.com> wrote:
>
> To make sure it doesn't fail lets hardcode every path that could come
> from pkg-config.
> (I managed to hit issues from dbus **and** cups as I have them installed)

Well if we hardcode the paths then what is the point of checking them
to begin with? I assume we are retrieving them from pkg-config because
they are configurable, and if they are configurable then we cannot
hardcode them. Maybe the point is to make these paths available even
when systemd is not? In that case AM_DISTCHECK_CONFIGURE_FLAGS should
be conditional to not having systemd, or perhaps it just the prefix
being used, in that case can you just do
$${dc_install_base}/$(<original_path>)?


> You can learn more about this type of issue here.
> https://bugzilla.redhat.com/show_bug.cgi?id=1123963
>
> See-Also: 9cc8b97b8 ("build: obexd: Fix make distcheck")
> Signed-off-by: Alfred Wingate <parona@protonmail.com>
> ---
>  Makefile.am | 8 ++++++++
>  1 file changed, 8 insertions(+)
>
> diff --git a/Makefile.am b/Makefile.am
> index 94f625db4..3e25c01dc 100644
> --- a/Makefile.am
> +++ b/Makefile.am
> @@ -2,6 +2,14 @@
>  AM_MAKEFLAGS = --no-print-directory
>  AM_CPPFLAGS =
>
> +# pkg-config gives absolute paths that do not work with distcheck that wants to use its own $(prefix)
> +AM_DISTCHECK_CONFIGURE_FLAGS = --with-dbusconfdir=$${dc_install_base}/usr/share \
> +                              --with-dbussystembusdir=$${dc_install_base}/usr/share/dbus-1/system-services \
> +                              --with-dbussessionbusdir=$${dc_install_base}/usr/share/dbus-1/services \
> +                              --with-systemdsystemunitdir=$${dc_install_base}/usr/lib/systemd/system \
> +                              --with-systemduserunitdir=$${dc_install_base}/usr/lib/systemd/user \
> +                              --with-cups_serverbin=$${dc_install_base}/usr/libexec/cups
> +
>  lib_LTLIBRARIES =
>
>  noinst_LIBRARIES =
> --
> 2.51.1
>
>


--
Luiz Augusto von Dentz

Re: [PATCH BlueZ 2/3] build: Fix distcheck by hardcoding non absolute paths

[Alfred Wingate]

Hello Luiz,

On Friday, October 31st, 2025 at 6:08 PM, Luiz Augusto von Dentz <luiz.dentz@gmail.com> wrote:

> Hi Alfred,
>
> On Thu, Oct 23, 2025 at 1:06 AM Alfred Wingate parona@protonmail.com wrote:
>
> > To make sure it doesn't fail lets hardcode every path that could come
> > from pkg-config.
> > (I managed to hit issues from dbus and cups as I have them installed)
>
>
> Well if we hardcode the paths then what is the point of checking them
> to begin with? I assume we are retrieving them from pkg-config because
> they are configurable, and if they are configurable then we cannot
> hardcode them. Maybe the point is to make these paths available even
> when systemd is not? In that case AM_DISTCHECK_CONFIGURE_FLAGS should
> be conditional to not having systemd, or perhaps it just the prefix
> being used, in that case can you just do
> $${dc_install_base}/$(<original_path>)?

Firstly system paths from pkg-config are fundamentally incompatible with how distcheck works.
distcheck installs them into its own prefix (not DESTDIR!). So when pkg-config gives it a system path
it will simply not work with that assumption. This is why you can configure arguments for distcheck itself
and systemd's own manual page also mentions this requirement[1].

> Additionally, to ensure that make distcheck continues to work, it is recommended to add the following to
> the top-level Makefile.am file in automake(1)-based projects:
> AM_DISTCHECK_CONFIGURE_FLAGS = \
>   --with-systemdsystemunitdir=$$dc_install_base/$(systemdsystemunitdir)

Secondly this affects --enable-systemd as well, its just not apparent as its explicitly disabled for distcheck[2].

Thirdly yes it would be nicer to use $${dc_install_base}/$(<original_path>). I didn't think it was possible so I didn't try,
but looking at the systemd manual it appears to be doable.

https://www.gnu.org/software//automake/manual/html_node/DISTCHECK_005fCONFIGURE_005fFLAGS.html
> The user can still extend or override the flags provided there by defining the DISTCHECK_CONFIGURE_FLAGS
> variable, on the command line when invoking make.

Also I just noticed this, but the DISTCHECK_CONFIGURE_FLAGS already in bluez should be rolled into AM_DISTCHECK_CONFIGURE_FLAGS as well.
Will have to address it.

[1] https://www.freedesktop.org/software/systemd/man/latest/daemon.html#Installing%20systemd%20Service%20Files
[2] https://github.com/bluez/bluez/blob/32d740d87970fa476644aeea267663ea3c74644a/Makefile.am#L753

--
Alfred Wingate

Re: [PATCH BlueZ 0/3] build: Fix distcheck while installing org.bluez.obex.service with --disable-systemd

[patchwork-bot+bluetooth]

Hello:

This series was applied to bluetooth/bluez.git (master)
by Luiz Augusto von Dentz <luiz.von.dentz@intel.com>:

On Thu, 23 Oct 2025 05:05:09 +0000 you wrote:
> Pkg-config variables give absolute paths with system prefixes, causing problems
> for distcheck which wants to install everything in its own prefix. To cope with
> this lets give distcheck its own configure arguments with its own prefix variable
> already included.
> 
> https://www.gnu.org/software/automake/manual/html_node/DISTCHECK_005fCONFIGURE_005fFLAGS.html
> 
> [...]

Here is the summary with links:
  - [BlueZ,1/3] build: Support setting the cups_serverbin directory
    (no matching commit)
  - [BlueZ,2/3] build: Fix distcheck by hardcoding non absolute paths
    (no matching commit)
  - [BlueZ,3/3] build: obexd: Revert "Fix make distcheck"
    https://git.kernel.org/pub/scm/bluetooth/bluez.git/?id=ebd6f9fefa26

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html