gmane.comp.freedesktop.hal

[PATCH 1/2] Don't close file descriptor twice

Thomas Jarosch — 2011-10-05T20:14:46

Detected by "cppcheck"

Please CC: comments.

Signed-off-by: Thomas Jarosch <thomas.jarosch< at >intra2net.com>
---
 hald/linux/probing/probe-video4linux.c |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/hald/linux/probing/probe-video4linux.c b/hald/linux/probing/probe-video4linux.c
index b055720..9faf009 100644
--- a/hald/linux/probing/probe-video4linux.c
+++ b/hald/linux/probing/probe-video4linux.c
< at >< at > -147,6 +147,7 < at >< at > main (int argc, char *argv[])
 libhal_device_free_changeset (cset);
 
 close (fd);
+fd = 0;
 
 ret = 0;

[PATCH 2/2] Fix mismatching allocation/deallocation

Thomas Jarosch — 2011-10-05T20:15:17

Detected by "cppcheck"

Please CC: comments.

Signed-off-by: Thomas Jarosch <thomas.jarosch< at >intra2net.com>
---
 hald/freebsd/probing/freebsd_dvd_rw_utils.c |   14 +++++++-------
 1 files changed, 7 insertions(+), 7 deletions(-)

diff --git a/hald/freebsd/probing/freebsd_dvd_rw_utils.c b/hald/freebsd/probing/freebsd_dvd_rw_utils.c
index c004916..130d8a3 100644
--- a/hald/freebsd/probing/freebsd_dvd_rw_utils.c
+++ b/hald/freebsd/probing/freebsd_dvd_rw_utils.c
< at >< at > -45,7 +45,7 < at >< at > scsi_command_new_from_cdrom (HFPCDROM *cdrom)
 static void
 scsi_command_free (ScsiCommand * cmd)
 {
-free (cmd);
+g_free (cmd);
 }
 
 static void
< at >< at > -107,7 +107,7 < at >< at > get_dvd_r_rw_profile (HFPCDROM *cdrom)
 if (scsi_command_transport (cmd, READ, list, len)) {
 /* GET CONFIGURATION failed */
 scsi_command_free (cmd);
-free (list);
+g_free (list);
 return -1;
 }
 
< at >< at > -170,7 +170,7 < at >< at > get_dvd_r_rw_profile (HFPCDROM *cdrom)
 }
 
 scsi_command_free (cmd);
-free (list);
+g_free (list);
 
 return retval;
 
< at >< at > -230,7 +230,7 < at >< at > pull_page2a_from_cdrom (HFPCDROM *cdrom)
 if (scsi_command_transport (cmd, READ, page2A, len)) {
 /* MODE SENSE failed */
 scsi_command_free (cmd);
-free (page2A);
+g_free (page2A);
 return NULL;
 }
 
< at >< at > -330,7 +330,7 < at >< at > get_write_speeds (const unsigned char *p, int length, int max_speed)
 }
 
 free_tmpspeeds:
-free (tmpspeeds);
+g_free (tmpspeeds);
 
 return result;
 }
< at >< at > -462,7 +462,7 < at >< at > get_disc_capacity_cdr (HFPCDROM *cdrom, guint64 *size)
 
 if (scsi_command_transport (cmd, READ, atip, len)) {
 /* READ TOC failed */
-free (atip);
+g_free (atip);
 goto done;
 }
 
< at >< at > -473,7 +473,7 < at >< at > get_disc_capacity_cdr (HFPCDROM *cdrom, guint64 *size)
 }
 retval = 0;
 
-free (atip);
+g_free (atip);
  done:
 scsi_command_free (cmd);

Re: hal & UPS with serial port

Richard Hughes — 2011-09-15T16:11:12

2011/9/14 Grzesiek Sójka <pld< at >sojka.co>:

Just uninstall hal. Hal is dead upstream and no longer maintained.

Richard.
_______________________________________________
hal mailing list
hal< at >lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/hal

hal & UPS with serial port

Grzesiek Sójka — 2011-09-14T18:25:42

Hi there.

Recently I updated to hal-0.5.14 (I do not remember the previous 
version). The problem is that after starting hald my UPS (connected via 
serial port) immediately switches off.

Please help. I already lost some data.

Regards

Greg

Debian

John Larysz — 2011-07-25T20:01:14

I am trying to get hal to configure on an  armel architecture machine with
build 2.6.29-00236-g4f8dbbb-dirty. It's not going well.

 

How can I tell if hal is even running? I can't see anything in ps -ef and I
can't find any error messages.

 

 

     John

HAL mount result differs if invoked by HAL

潘图 — 2011-06-26T14:18:12

Hi,

in order to have my msdos device being able to use correct charset i am
using a fuse module. Here is what I do:

root> mount_msdosfs -L zh_CN.GBK /dev/msdosfs/device /mnt/tmp
root> fusexmp_fh
-oallow_other,modules=iconv:subdir,from_code=GBK,subdir=/mnt/tmp
/media/CW-P7EOS-X

This allows me to have chinese/japanese Names in the Folder/File Name
even if created on windows.

If I do not use the fusexmp_fh, but only the mount_msdosfs, all
filenames with chin/jap names in it will be displayed as (invalid encoding).

No I have created a wrapper with the name mount_msdosfs and move the
actual binary to another place. This wrapper does the 2 commands above
automatically.

So when I am under my gnome environment and plug in my USB Device with
msdos filesystem, HAL calls the mount_msdosfs (which is now the wrapper)
and this script does the two commands above.



The only problem is, that the result differ, if invoked by HAL or if
invoked manually.

If I call the wrapper script, everything works well, if HAL does the
thing (with the exact same paramter), both locations are mounted as
expected, but in the /media/CW-P7EOS-X mount all Files/Folders with
chin/jap. Name in it do not appear.

I have absolutely no idea why it is like this. I can do it by hand with
the exact same command and parameter and everything goes well, but HAL
give strange results.

So I am looking for a place, where I can see any HAL errors.

Here are my System Details:

FreeBSD host.domain 9.0-CURRENT FreeBSD 9.0-CURRENT #3: Mon May 23
07:20:47 CST 2011
root< at >trinity.the-wire.de:/usr/obj/usr/src/sys/TRINKERNEL  i386

HAL package version: 0.5.14

Any help on this is appreciated, I am trying to write a fully functional
wrapper, that remove the charset Problem of msdos devices between
windows (utf-16 i think) and *bsd systems, so that we can use chin/jap
names in folders and files even if been created under windows.

Best wishes
Patu

Hey

2011-05-31T22:21:53

Hello  I just wanted to share this opportunity with you, I've been making 200-300 dollars a day and I started only a week ago. Check out this news article and it will show you how to get started, it's definitely easy enough for you :)! http://news7cnbc.com/money

need help with hald on MacOS

Cuga — 2011-05-16T00:43:25

I need help debugging a problem with hald!

I'm building an X11 window manager under XQuartz/MacOS 10.6.7. I'm not using Fink or Macports. I'm installing everything in POSIX-compliant locations (/usr/local).

Dbus-1.5.0 and hal-0.5.11 both compile and install fine, but I can't get hald to run. It throws messages and quits.

The X11 wm I'm using requires dbus-session, and optionally uses dbus-system and hal for hardware-specific modules. Dbus is running properly. The X11 wm is successfully connecting to dbus-session which is being launched by launchd. 

I can launch dbus-daemon --system, either by hand at the terminal, or with launchd. I get the same results either way. The pid and socket files are created as expected and the daemon runs in the background without error. However hald does not start up properly after this. The hald man page suggests starting hald by hand in verbose mode, which I have tried, with results shown below. hald quits and gives me some messages I don't understand. I've tried re-compiling both dbus and hald with different config options, but I get the same results, so I'm stumped.

I noticed the directory /usr/local/etc/hal/ contains the expected directories, but they are all empty. The build directory had many fdi files that didn't seem to install. Not sure if that's normal or a source of the problem.

Here are is some system information:

$ ps aux | grep dbus
david   143   0.0  0.0  2447012    760   ??  S     6:56PM   0:00.00 /usr/local/bin/dbus-daemon --nofork --session
messagebus    48   0.0  0.0  2435548    728   ??  Ss    6:55PM   0:00.01 /usr/local/bin/dbus-daemon --nofork --system

$ env | grep DBUS
DBUS_LAUNCHD_SESSION_BUS_SOCKET=/tmp/launch-9irO8p/unix_domain_listener

$ ls -al /tmp/launch-9irO8p/unix_domain_listener
srwx------  1 david  wheel  0 May 13 18:56 /tmp/launch-9irO8p/unix_domain_listener

$ ls -al /var/run/dbus*
-rw-r--r--  1 root  daemon  3 May 13 18:56 /var/run/dbus_pid
srwxrwxrwx  1 root  daemon  0 May 13 18:56 /var/run/dbus_system_bus_socket

$ ls -al /var/cache | grep hald
drwxr-xr-x   2 haldaemon  haldaemon   68 May 14 11:05 hald

$ /usr/local/sbin/hald --daemon=no --verbose=yes
16:57:42.030 [I] hald.c:669: hal 0.5.11
16:57:42.031 [I] hald.c:734: Will not daemonize
16:57:42.031 [I] hald_dbus.c:5381: local server is listening at unix:path=/tmp/dbus-QcRvOECZN8,guid=f638aeb56c645e37d6dbc6164dd06876
16:57:42.074 [I] hald_runner.c:301: Runner has pid 404
Runner started - allowed paths are '/usr/local/libexec:/usr/local/lib/hal/scripts:/usr/local/bin'
16:57:42.087 [I] hald_runner.c:182: runner connection is 0x101000670
16:57:42.089 [D] util_helper.c:124: drop_privileges: could not set group id
[ note: then it quits and gives me the shell prompt ]

$ sudo /usr/local/sbin/hald --daemon=no --verbose=yes
19:08:08.339 [I] hald.c:669: hal 0.5.11
19:08:08.339 [I] hald.c:734: Will not daemonize
19:08:08.339 [I] hald_dbus.c:5381: local server is listening at unix:path=/tmp/dbus-h4OzVqCnRk,guid=a2f53e5960b12f13b8a3625b4dcde408
Dynamic session lookup supported but failed: launchd did not provide a socket path, verify that org.freedesktop.dbus-session.plist is loaded!
19:08:08.343 [E] hald_dbus.c:5747: dbus_bus_get(): Not enough memory
[ note: then it quits and gives me the shell prompt ]
[ note: org.freedesktop.dbus-session.plist is indeed loaded and DBUS_LAUNCHD_SESSION_BUS_SOCKET is set as shown above! ]

dbus configured with:
./configure --disable-static --with-dbus-user=messagebus --enable-console-owner-file --disable-asserts --disable-tests --enable-checks --enable-kqueue --enable-launchd --disable-userdb-cache --disable-ansi --with-xml=expat --disable-xml-docs --disable-doxygen-docs --disable-dependency-tracking --without-x ac_cv_func_poll=no --with-system-pid-file=/var/run/dbus_pid --with-system-socket=/var/run/dbus_system_bus_socket --with-session-socket-dir=/tmp --enable-verbose-mode --localstatedir=/var

hal configured with:
./configure --with-hal-user=haldaemon --with-hal-group=haldaemon --disable-pci-ids --disable-usb-ids --disable-pnp-ids --disable-pmu --disable-policy-kit --disable-console-kit --disable-dependency-tracking --disable-static --with-socket-dir=/tmp --with-localstatedir=/var --with-macbookpro --with-cpufreq

Let me know if any other system info would be useful.

Thanks in advance,
Dave

Re: [PATCH] ConsoleKit.conf: fix explicit missing destination

Lennart Poettering — 2011-05-10T17:30:41


Doesn't apply on current git. Can you rebase please?



Lennart

Keymapping

David Woodfall — 2011-04-24T02:07:33

Hi, I have a usr9600 phone which acts as an extra keyboard and I'm
trying to map keys 1-0 to keypad 1-0:

         <append key="input.keymap.data" type="strlist">0x02:KP1</append>
         <append key="input.keymap.data" type="strlist">0x03:KP2</append>
         <append key="input.keymap.data" type="strlist">0x04:KP3</append>
         <append key="input.keymap.data" type="strlist">0x05:KP4</append>
         <append key="input.keymap.data" type="strlist">0x06:KP5</append>
         <append key="input.keymap.data" type="strlist">0x07:KP6</append>
         <append key="input.keymap.data" type="strlist">0x08:KP7</append>
         <append key="input.keymap.data" type="strlist">0x09:KP8</append>
         <append key="input.keymap.data" type="strlist">0x0a:KP9</append>
         <append key="input.keymap.data" type="strlist">0x0b:KP0</append>

lshal shows the mappings but they aren't working. I got the codes with 
scankey -s and I've also tried using xev which show different keycodes.

I've also tried <merge> instead of <append>

Any ideas why this isn't working?

Cheers

Dave

[PATCH] ConsoleKit.conf: fix explicit missing destination

Auke Kok — 2011-04-11T21:04:22

The last rule in the dbus system.d conf does not explicitly list the 
destination, which would give problems if a potential other target has 
the same path. Caught by an automated MeeGo build check.

Signed-off-by: Auke Kok <auke-jan.h.kok< at >intel.com>

--- a/data/ConsoleKit.conf2011-04-11 14:01:34.813582580 -0700
+++ b/data/ConsoleKit.conf2011-04-11 14:01:53.255582574 -0700
< at >< at > -140,7 +140,8 < at >< at >
      <allow send_destination="org.freedesktop.ConsoleKit"
             send_interface="org.freedesktop.ConsoleKit.Session"
             send_member="GetIdleSinceHint"/>
-    <allow send_interface="org.freedesktop.ConsoleKit.Session"
+    <allow send_destination="org.freedesktop.ConsoleKit"
+           send_interface="org.freedesktop.ConsoleKit.Session"
             send_member="SetIdleHint"/>
    </policy>

Re: keymapping

Robby Workman — 2011-04-24T04:59:05

On Sun, 24 Apr 2011 05:42:59 +0100
David Woodfall <dave< at >dawoodfall.net> wrote:



Oh, yeah, still used hal in 13.1.  Not sure what's going on
there, and to be frank (and likely disappointing), I really
don't have much inclination to work on that since we're on
the verge of a new release (and hal is mostly dead anyway,
especially wrt xorg)...

-RW

Re: Re: keymapping

David Woodfall — 2011-04-24T04:42:59

On (23:19 23/04/11), Robby Workman <rw< at >rlworkman.net> put forth the proposition:

Even the xorg in slack 13.1? (By the way xorg/hal does see the
keyboard and set layout etc. just not the maps)

Re: keymapping

Robby Workman — 2011-04-24T04:19:46

On Sun, 24 Apr 2011 03:27:46 +0100
David Woodfall <dave< at >dawoodfall.net> wrote:



It's because recent Xorg uses udev instead of hal for this.

-RW

keymapping

David Woodfall — 2011-04-24T02:27:46

Hi, I have a usr9600 phone which acts as an extra keyboard and I'm
trying to map keys 1-0 to keypad 1-0:

         <append key="input.keymap.data" type="strlist">0x02:KP1</append>
         <append key="input.keymap.data" type="strlist">0x03:KP2</append>
         <append key="input.keymap.data" type="strlist">0x04:KP3</append>
         <append key="input.keymap.data" type="strlist">0x05:KP4</append>
         <append key="input.keymap.data" type="strlist">0x06:KP5</append>
         <append key="input.keymap.data" type="strlist">0x07:KP6</append>
         <append key="input.keymap.data" type="strlist">0x08:KP7</append>
         <append key="input.keymap.data" type="strlist">0x09:KP8</append>
         <append key="input.keymap.data" type="strlist">0x0a:KP9</append>
         <append key="input.keymap.data" type="strlist">0x0b:KP0</append>

lshal shows the mappings but they aren't working. I got the codes with
scankey -s and I've also tried using xev which show different
keycodes.

I've also tried <merge> instead of <append>

Any ideas why this isn't working?

Re: What replaces this part of HAL?

Martin Pitt — 2011-02-24T14:51:49

Hello Tom,

Tom Horsley [2011-02-24  8:43 -0500]:

You should instead listen to the uevent which adds (or changes) the
/dev/input/eventXX device node. That's what hal catches as well, and
it just translates that into a D-BUS signal.


Again, these will create a new USB device (/dev/bus/usb/..) with an
ID_GPHOTO2 property.


It's all udev with uevents (which is also HAL's main input source,
apart from some CD-ROM polling which went into udisks).

Also, "is going away" is a bit of an understatement, it's really quite
dead. GNOME switched over a year ago, KDE and XFCE finished the
transition a couple of months back (see [1]). The major distros don't
even install it by default any more, or at least switched hal to D-BUS
activation so that it doesn't run all the time and cost extra boot
time/CPU/battery power, and just gets started for the odd old program.

Martin

[1] https://wiki.ubuntu.com/Halsectomy

What replaces this part of HAL?

Tom Horsley — 2011-02-24T13:43:37

I keep reading that HAL is going away, yet I have programs
that use bits and pieces of HAL, and I can't for the life
of me deduce what the "official" replacement for this
functionality is. For instance:

When my USB mouse is plugged/unplugged, I get DBUS messages
which allow me to re-run my xinput settings for the mouse
to properly configure the buttons and acceleration.

When my PTP cameras are plugged/unplugged, I also get DBUS
messages (with the HAL udi string), so I can do hal queries
to dig up properties like the USB bus and device numbers
so I can open the camera with libgphoto2.

So that's my question: If HAL really is going away, what
is supposed to replace this kind of notification mechanism
for newly plugged devices?

Re: hald-addon-cpufreq broken on newer kernels

Ozan Çağlayan — 2011-02-17T20:54:53

Ignore hald-addon-cpufreq, that was a misnomer. The problem is in the 
ACPI parsing stuff.

hald-addon-cpufreq broken on newer kernels

Ozan Çağlayan — 2011-02-17T20:52:31

Hi,

I don't know if this deserves a fix but some removals in /proc on newer 
kernels causes hald-addon-cpufreq to not correctly identify processors.

In:
http://cgit.freedesktop.org/hal/tree/hald/linux/acpi.c

< at > line 559.

The linux.acpi_path does no longer provide the "info" file.

Regards,

Re: hal Digest, Vol 63, Issue 3

tom z — 2010-12-31T01:23:58

thanks a lot

On Fri, Dec 31, 2010 at 4:00 AM, <hal-request< at >lists.freedesktop.org> wrote:

Re: help hal VS udisks

Maciej Grela — 2010-12-30T16:15:31

2010/12/30 tom z <maillist.tom< at >gmail.com>:

A lot of HALs functionality (detecting & assigning properties to
devices) was moved to udev. Some procedures that could be called for
devices ("callouts" iirc) are now available via dbus from udisks and
upower.

Br,
Maciej Grela