Release version 0.0.1.7 not working

[mightyoakbob]

Hi Pete,
Just tried the release version 1.7 and the player does not show up in Kazoo and there is no webserver.
Gone back to 1.6 beta 1 which works fine.

Thanks.

Bob.

[PeteManchester]

Hi Bob, Do you have any additional info from when you test v1.7, such as log files etc? Thanks, Pete.

…

On Fri, 12 Apr 2024 at 12:30, mightyoakbob ***@***.***> wrote: Hi Pete, Just tried the release version 1.7 and the player does not show up in Kazoo and there is no webserver. Gone back to 1.6 beta 1 which works fine. Thanks. Bob. — Reply to this email directly, view it on GitHub <#107>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AA5RVJYHI42B6NUNIV3A7FDY47AT5AVCNFSM6AAAAABGD7B5OGVHI2DSMVQWIX3LMV43ASLTON2WKOZSGIZTSOJSGIZDANY> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[mightyoakbob]

Hi Pete,

I should have thought you would need the log file, sorry about that.

Had to build the whole player again today but at least it proves the fault is solid.

Here it is.
MediaPlayerLog.zip

Thanks.

Bob.

[mightyoakbob]

Hi Pete,

I don't know if this helps but I've been looking at the new code in v1.7 and comparing with what we had before. I've found something that to me looks wrong but I would expect that's probably my lack of understanding.

In the previous working version we had the line:
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$SCRIPTPATH/mediaplayer_lib/ohNet/linux/armv6hf

That path points to a directory "armv6hf" that exists.

In the 1.7 version we have the line:

export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$SCRIPTPATH/mediaplayer_lib/ohNet/linux/$arch

So I'm guessing that in MY CASE the variable arch should contain the value "armv6hf" but it doesn't.

I've worked out (I think) that the instruction
arch=uname -m should set the value but if I do this..

arch=`uname -m`
echo "$arch"

I get:

armv7l
The character after the 7 is 0x6C lower case letter L.
There is an armv7 folder but not one that ends in lower case L.

But unfortunately things are more complex. Although the above results from my playing look reasonable to me I tried re-booting the pi logging in and just reading the varaible.

echo "$arch"

It's empty!! Nothing in it. I rebooted and tried again, still empty but I can set it as described. I'm confused again and misunderstanding aren't I.

Incidentally, I've been computing and even writing stuff in assembler for 40 years and that's the first time I've ever encountered the character ` (0x60).

Cheers,

Bob.

[mightyoakbob]

More...
In the files run_systemd.sh and run.sh I have commented out the line:

arch=uname -m``

and replaced it with:

arch="armv7"

That gets this version to run both the webserver and the player.

So to me there are two problems.

1. When run as part of a start up script arch=uname -m`` doesn't seem to do anything.
2. If the command is used at the cli then it sets it to armv7l which has no directory of the same name.

For some reason the "code" mark up isn't working properly on this post.

Bob.

[PeteManchester]

Thanks Bob, Well done.. Looks like your OS is a different version (armv71) that I don't have the library files for. There are a number of ways to fix it, probably the easiest and best is the one you have done. Other ways could be: 1. Create a new directory called armv71 and copy the contents of armv7 into that. 2. Try and download the correct files for armv71 from the Linn repository, but I don't see any for armv71. 3. Modify the scripts to set the arch variable to be armv7 if the result of uname -m is 'armv71' For example: #andrum993 Generate lib path dynamically, since Java can't do this itself now :( arch=`uname -m` if [ $arch = 'armv7l' ]; then echo "OS is 'armv7l' changing to 'armv7'" arch='armv7' echo $arch fi echo $arch From your previous email, when you tried this: "But unfortunately things are more complex. Although the above results from my playing look reasonable to me I tried re-booting the pi logging in and just reading the varaible. echo "$arch"" It didn't work, because the variable 'arch' only exists in the context of the script, it is not a system variable. Thanks again, Pete.

…

On Sat, 20 Apr 2024 at 13:33, mightyoakbob ***@***.***> wrote: More... In the files run_systemd.sh and run.sh I have commented out the line: arch=uname -m`` and replaced it with: arch="armv7" That gets this version to run both the webserver and the player. So to me there are two problems. 1. When run as part of a start up script arch=uname -m`` doesn't seem to do anything. 2. If the command is used at the cli then it sets it to armv7l which has no directory of the same name. For some reason the "code" mark up isn't working properly on thei post. Bob. — Reply to this email directly, view it on GitHub <#107 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AA5RVJ47VNESCSNLKRJNYHLY6JOCJAVCNFSM6AAAAABGD7B5OGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDANRXGY2TSOBXGQ> . You are receiving this because you commented.Message ID: ***@***.***>

[mightyoakbob]

Hi Pete,

Chuffed if I've manged to help us understand what was going wrong.

I've been doing a fair bit of googling hoping to understand the scripting a bit better now that I think I understand the issue.

Thanks for explaining why the arch vaiable was empty.

Where would I find the Linn repositories?

I use pi3 devices because they are quite cheap secondhand and have wi-fi. They are arm7 so any ideas what can we do about arm8 and later devices?

So in an attempt to both get the best selection for arm Raspberry pi use and at the same time allow the code to still work on other platforms I had a go and came up with this code:

#andrum993 Generate lib path dynamically, since Java can't do this itself now :(
arch=`uname -m`
case $arch in
    armv5*)
    arch="armv5sf"
    ;;
    armv6hf*)
    arch="armv6hf"
    ;;
    armv6*)
    arch="armv6sf"
    ;;
    armv7*)
    arch="armv7"
    ;;
esac
echo "$arch"
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$SCRIPTPATH/mediaplayer_lib/ohNet/linux/$arch

All "cases" have a wildcard end, seeing as odd things get added apparently.
Specifically armv6hf is dealt with allowing all other armv6 types to be grouped into armv5sf.
I'm sure you get the idea.

I've just tried it and FANFARE it works!
What do you think? Over the top?

This is what you get when people start to google too much. :-)

Cheers,

Bob.

[PeteManchester]

That's great, I've added your code to the repository. Thanks for your help. Pete.

…

On Mon, 22 Apr 2024 at 14:22, mightyoakbob ***@***.***> wrote: Hi Pete, Chuffed if I've manged to help us understand what was going wrong. I've been doing a fair bit of googling hoping to understand the scripting a bit better now that I think I understand the issue. Thanks for explaining why the arch vaiable was empty. Where would I find the Linn repositories? I use pi3 devices because they are quite cheap secondhand and have wi-fi. They are arm7 so any ideas what can we do about arm8 and later devices? So in an attempt to both get the best selection for arm Raspberry pi use and at the same time allow the code to still work on other platforms I had a go and came up with this code: #andrum993 Generate lib path dynamically, since Java can't do this itself now :( arch=`uname -m` case $arch in armv5*) arch="armv5sf" ;; armv6hf*) arch="armv6hf" ;; armv6*) arch="armv6sf" ;; armv7*) arch="armv7" ;; esac echo "$arch" export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$SCRIPTPATH/mediaplayer_lib/ohNet/linux/$arch All "cases" have a wildcard end, seeing as odd things get added apparently. Specifically armv6hf is dealt with allowing all other armv6 types to be grouped into armv5sf. I'm sure you get the idea. I've just tried it and FANFARE it works! What do you think? Over the top? This is what you get when people start to google too much. :-) Cheers, Bob. — Reply to this email directly, view it on GitHub <#107 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AA5RVJY7SAYWVOIW2CINDRLY6UFIJAVCNFSM6AAAAABGD7B5OGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDANRZGQYDSMZWGE> . You are receiving this because you commented.Message ID: ***@***.***>