Errors in both /etc/init.d/LCDd script and /etc/LCDd.conf that break lcdproc

[ironstorm]

Version: 0.5.0-1

The driver path in LCDd.conf does not correspond to the driver locations, causing lcdproc not to start, here's the patch:
--- /etc/LCDd.conf.orig 2006-09-30 23:54:12.000000000 -0400
+++ /etc/LCDd.conf 2006-09-30 23:54:31.000000000 -0400
@@ -76,7 +76,7 @@
 # the driver modules and will thus not be able to
 # function properly.
 # NOTE: Always place a slash as last character !
-DriverPath=server/drivers/
+DriverPath=/usr/lib/lcdproc/

-s and -f require a boolean value in /etc/init.d/LCDd otherwise LCDd refuses to start, here's the patch:
--- /etc/init.d/LCDd.orig 2006-10-01 01:20:56.000000000 -0400
+++ /etc/init.d/LCDd 2006-10-01 01:19:55.000000000 -0400
@@ -14,7 +14,7 @@
 NAME=LCDd
 DESC=LCDd
 DESC="LCDd"
-DAEMON_OPTS="-s -f -c /etc/LCDd.conf"
+DAEMON_OPTS="-s true -f true -c /etc/LCDd.conf"

 test -x $DAEMON || exit 0

BTW, thanks for including imon support (even though the pkg desc doesn't mention it) in the build options... :-)

-G

[Gnu_Raiz (gnu-raiz)]

I am surprised that no one has confirmed this bug. To me this is a critical bug for those who use the Lcdproc package.

I just wanted to mention that his patch to LCDd works, I was pulling my hair out trying to find a work around. For me i would start LCDd with the -d option using my lcd driver. Then I would try to attach lcdproc to it, but it would fail, at least I know that it is a problem with lcdproc.

In fact I will be filing another bug report to lcdproc as it still fails, I am using AMD64. I have to restart LCDd a few times before lcdproc shows up, and doesn't fail.

[TEN (launchpad-20-ten)]

lcdproc-0.5.1-3-feisty in Ubuntu 7.04 "Feisty Fawn" still has this same problem (unsurprisingly, for its make process as described in the INSTALL file and for Ubuntu e.g. by http://vdr-portal.de/board/thread.php?threadid=52517&page=3#post614045 actually builds it this way - and moreover it should be replaced by lcdproc-0.5.2 for the reasons stated there and at http://www.lcdproc.org/download.php3, i.e. notably malfunctions of serialVFD).

[Mario Limonciello (superm1)]

I'm attaching a debdiff resolving this issue and subscribing u-u-s to upload it.

[Joel Ebel (jbebel)]

Perhaps in 0.5.0 it was, but in 0.5.1, the -f option isn't boolean. Having true after it results in it not starting. In fact, the existing options of -s 1 -f -c /etc/LCDd.conf worked fine as of 0.5.1. The driver path does still need to be changed though.

[Mario Limonciello (superm1)]

lcdproc (0.5.1-3ubuntu1) gutsy; urgency=low

  * Add 01_fix_driverpath.dpatch to fix default DriverPath. (LP: #63488)
  * debian/rules:
    - Add dpatch include.
  * debian/lcdproc.LCDd.init:
    - Make sure that DAEMON_OPTS has boolean values.
  * debian/control:
    - Update maintainer field to <email address hidden>
    - Add dpatch to build deps

 -- Mario Limonciello <email address hidden> Sun, 29 Jul 2007 18:14:41 -0500

[Barry deFreese (bddebian)]

Uploaded. Thank you for the patch!

[Joel Ebel (jbebel)]

Wait! Did you read my last post? I think this patch fixes one problem and makes another.

[Barry deFreese (bddebian)]

No, I did not understand that post.

[ironstorm (ironstorm-gmail)]

What Joel was referring to is that according to the LCDproc users guide, -f does not take a boolean parameter while -s does...

So according to http://lcdproc.sourceforge.net/docs/stable-0-5-x-user.html#lcdd-commandline-options

it should be
+DAEMON_OPTS="-s true -f -c /etc/LCDd.conf"

[Joel Ebel (jbebel)]

Right. Or not edit that line at all. It worked before. Only the driverpath needed to be fixed.

[TEN (launchpad-20-ten)]

For the Gutsy Gibbon climbing closer, a move to the new stable LCDproc v0.5.2 is recommended, which has fixes to several important drivers (as discussed above) and has been available from http://www.lcdproc.org/download.php3 since April 2007.

[Barry deFreese (bddebian)]

I have packaged up 0.5.2 and posted it here. http://www.bddebian.com/packages/ubuntu/lcdproc/ Is there any chance one of you could test this out before Thursday's UVF and I can upload if it works. I kept the driver path fix since it still seemed incorrect. Thanks!

[Joel Ebel (jbebel)]

Barry,

It looks like the patch isn't applying to the new LCDd.conf. The package still has the old, non-working DriverPath set. Other than that, this package appears to work fine. The startup scrips has working options.

Joel

[Barry deFreese (bddebian)]

That was my fault it was a packaging error. I have uploaded the new version to Gutsy now. Thanks for the feedback!

[Daniel Holbach (dholbach)]

Marking as 'fix released' then.