Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Messages - mhorst

Pages: 1 2 [3] 4 5
Users / Re: Diskless MD, Mythtv frontend fails to start.
« on: December 19, 2010, 12:43:57 pm »
I've had an issue like this. It was caused by my packages not being up to date, see
Especially note the part about changing the 600 to 500 if necessary.

Alternatively check how the mythfrontend is started when you use the orbiter (use ps axf | grep myth), and then start it yourself using the same command. That way you get the output from mythfrontend to your terminal and you might be able to determine what is going on.
(You might have to add the -display <display> option to get the frontend to open on the right display though.)

In my case the command used to start mythfrontend included "-w mythfrontend", but the -w option doesn't take any arguments so the program interpreted the mythfrontend part after the -w as a plugin that had to be loaded. This plugin does not exist, hence the weird behavior.
After that I looked into the trac tickets, found that this issue had been solved a long time ago, and that lead me to the conclusion that updates weren't installed properly.

Users / Re: soundsystem with ethernet
« on: August 18, 2010, 09:58:10 am »
I think might be of help.
Look at the Proliphix NT series thermostat for an example. The thermostat uses the HTTP protocol, but by selecting a different port and leaving out the HTTP headers you should be able to communicate over telnet.

Users / Re: DLNA experiences?
« on: August 17, 2010, 09:17:13 pm »
Interesting. I guess I'll just have to find a TV with good DLNA support and then see how far I get. (Until I have it working I can at least keep the current MD).

I found a suggestion somewhere that instead of watching live TV, it might be possible to start a recording of the channel I want to watch and then watch the recording (in progress) via DLNA. But I haven't seen anybody reporting a success with this method yet.

Anyway, I know some python, but I really should finish what I start ( before diving into another project :)

Users / Re: DLNA experiences?
« on: August 16, 2010, 10:58:44 am »
Thanks for the info. I'll check to see what the different DLNA enabled TVs support in terms of playback and pausing.
For the lost functionality I was indeed thinking of a wireless orbiter or something.

I have two more questions:
1) Does the MythTV PnP server also enable you to watch live TV? In my current setup I haven't bothered to run a coax cable to my TV, I always watch TV via the MD. Would that still be possible, perhaps by selecting a channel on a wireless orbiter?
2) What about subtitles for movies on my harddisk? Is this also TV dependent, or can MediaTomb somehow send them along with the media stream?


Users / DLNA experiences?
« on: August 14, 2010, 01:43:49 pm »

Its time for me to buy a new TV set and I noticed some of the new TVs support DLNA. As I understand it such TVs can be connected to the network and should be able to play media directly from the MediaTomb server that is integrated with LinuxMCE.
So I was wondering if anybody has experience with such a setup?

Particularly, assuming I'd use a separate device as orbiter, would it still be necessary to have an MD connected to the TV?

Ok, I'll make a trac ticket as soon as I've got a patch on a clean checkout.

Question though: should that checkout be on the trunk or on the LinuxMCE-810 branch?

Also, I always like to check whether my code at least builds before sending it out. So, is there a way to build with my patch, without the builder replacing the #DEFINE version ?
I quickly tried doing a make on a clean checkout (without my patch), but I get a message that ../include/version.h does not exist. (Which is correct, there is a version.h.prep file though...)

Ok I got a bit futher. I modified PlutoDHCP. It now automatically includes the domain name from the web-admin's network settings pageinto the dhcpd.conf file. I attached the svn diff with my changes to this post, for anybody who's interested.

I found some issues when I tried following the instructions on The avenard repro key was not installed by one of the scripts, resulting in an error, and the build did not complete succesfully.
I was able to work around that because the PlutoDHCP component which I needed to change did build.

I also have some questions:
1) I made my changes on the "LinuxMCE-810" branch, because that is the branch that I got after following the instructions on the wiki page. Is this correct? I actually expected to be working on the "trunk" branch instead.
2) What's with the following lines in the diff?
Code: [Select]
-#define  VERSION "<=version=>"
+#define  VERSION ""
I never changed that line, and I guess the version is supposed to be filled in automatically. Should I manually remove them before a commit or is this done automatically?
3) How can I ensure that new files will be included in an existing package? For the bind config files I will create a new template file, and that will have to be included in the pluto-boot-scripts package once its finished. There's probably a config file somewhere, but I haven't found it yet.
4) Is there a way to build an individual package? For the PlutoDHCP I just manually installed the binary, and since it is already part of a package I am quite sure that it will be included and installed as normal. I am not so sure about a new file that is added to a package however, so I'd like to test no. 3), but I would like to do that by building the individual package.

No, I did NOT change the primary key. Upon re-reading my previous post, however, I can see how you can interpret it that way :)

I changed the foreign key in the Device table of the pluto_main database. My device with PK_Device=78 now has FK_DeviceTemplate=2127 (instead of FK_DeviceTemplate=2190). That's the only thing I changed, and I kinda considered that the obvious thing to do (I'll try express myself better next time :))

I was just surprised that the foreign keys in other tables do not change automatically when the primary key is changed.

I updated my system yesterday and it stopped working.

Took me a while to track it down; the id of the device template for my Proliphix driver had changed, during the update, but the template id referenced by already installed devices was not updated. Is this supposed to happen?

I had to do a manual mysql query to fix it, because according to the web-admin there was no such device (despite the fact that my thermostat was there in the list of devices :) ).


I checked what happens if I set 
Code: [Select]
option domain-name "linuxmce.lan"; in the /etc/dhcp3/dhcpd.conf file.
After that I can actually resolve http://dcerouter on my laptop, so that looks good.
I now know what changes should make to the config files, now it is a matter of getting the LinuxMCE code to do that.

I saw that the domain name is also configurable in the web-admin, so I figured it would be a good first step to add the domain name to the DHCP template file, and after that create a template for the bind files.

The dhcp template contains several keys like %CORE_INTERNAL_ADDRESS%, so I wanted to check where this template is filled in and see if there is already a key for the domain name. If that would be the case the DHCP changes would be a simple matter of updating the template.
However, it seems that the template is not used; my search lead me to PlutoDHCP.cpp file in the source code, and that source code contains a function that produces the text for the dhcpd.conf file from scratch. The output of this function seems to be the same as the template (except for the comments on the first 3 lines), but it does not actually use the template at all.

I guess that means I should not alter the template but just the PlutoDHCP.cpp file, or am I missing something?

Anyway, is there an example of how the template files should be used? I saw that the ddclient.conf.template is filled in the script. I can create a similar approach for the bind9 config files (shell scripts that use the ReplaceVars function), if that is the proper way to go.

I'll take a look, but the amount time I have is erratic, so if somebody else is interested too; go right ahead (you can probably use what I have so far as a starting point).

By the way, if possible I would like to expand this feature so that we can associate names with MAC addresses or something, such that the names of devices can also be used to resolve them.

Anyhow, I saw that LinuxMCE already includes a DNS server (bind9), so I read some HOWTOs and played with the config files a bit.
I added two zones to /etc/bind/named.conf.local:
Code: [Select]
// Do any local configuration here

// Consider adding the 1918 zones here, if they are not used in your
// organization
//include "/etc/bind/zones.rfc1918";

zone "linuxmce.lan" IN {
type master;
file "zone.linuxmce.lan";

zone "" {
type master;
file "revp.linuxmce.lan";

I also added the files /var/cache/bind/zone.linuxmce.lan:
Code: [Select]
; dns zone for linuxmce. lan
$ORIGIN linuxmce.lan.
@ IN SOA dcerouter hostmaster (
2009051401      ; serial (YYYMMDD of last change, followed by a sequence number)
8H ; refresh
4H ; retry
4W ; expire
1D ) ; minimum
NS dcerouter
; DCE router address
dcerouter A
; Addresses of other devices
moon54 A

and I added the file /var/cache/bind/revp.linuxmce.lan:
Code: [Select]
; dns zone for linuxmce. lan
$ORIGIN linuxmce.lan.
@ IN SOA dcerouter hostmaster (
2009051401 ; serial (YYYMMDD of last change, followed by a sequence number)
8H ; refresh
4H ; retry
4W ; expire
1D ) ; minimum
NS dcerouter
; DCE router address
1 PTR dcerouter
; Addresses of other devices
2 PTR moon54

I can now resolve dcerouter.linuxmce.lan and moon54.linuxmce.lan from my laptop and it will direct me to the correct IP address.
To be able to remove the ".linuxmce.lan", I think I have to reconfigure the DHCP server on the core to provide the "linuxmce.lan" domain name to its clients. I haven't gotten around to trying that yet.
And, of course, if this all works the new files I created in /var/cache/bind will have to be updated automatically rather than manually.

Also, it seems that the core itself considers itself to be in the domain provided by my ISP, and uses the DNS server of my ISP directly. So currently the only place in the network where I cannot resolve the new names is from the core itself. (Of course the core has a hosts file, so that doesn't really matter, but I would like to have the name resolution work the same on every system in the network.)

So that's my progress so far.

Users / Re: Why many CATx cables?
« on: April 25, 2010, 10:14:44 pm »
If you have to option, I'd run multiple cables while you're at it.

I didn't have the option in my house; I had to run the cables through existing conduits. I now have a small local router in my living room and one in my upstairs study. They use power over ethernet (802.3af) to get their power so I didn't have to put them near any power outlet and just tucked them away in a corner.
Works like a charm, and so far I haven't run out of connections.

I've already updated the page, but the page I reference there can still use some work.
I'll see what I can do.

BTW: I've only made an ethernet based driver, and have only added ethernet based examples to the wiki page. If somebody has experience with writing drivers for the other communication methods (RS232, USB, etc.), feel free to add the drivers you wrote to the examples section of the page, and note any differences in creation. (For RS232 drivers the device data probably has to contain the COM port to be used by the core, instead of the TCP port, but I have no idea which id that field has.)

You're welcome :)

It was actually quite fun, I'm planning on contributing more in the future.

Yep. Did something go wrong?

Pages: 1 2 [3] 4 5