libboost_regex 1.49 on UDOObuntu/Debian Wheezy

Discussion in 'General Programming Discussion' started by Hans55, Jan 8, 2015.

  1. Hans55

    Hans55 New Member

    Joined:
    Jan 8, 2015
    Messages:
    3
    Likes Received:
    0
    Hi

    I'm new here in the forum and also on Udoo/Linux. So first, hello everybody :) I'm an electrical engineering student from Switzerland.

    I have a problem using the PiAUISuite voice control from Steve Hickson (https://github.com/StevenHickson/PiAUISuite)
    First, I tried with UDOObuntu but when I want to run the voicecommand program, it gives me this error:
    "voice command: error while loading shared libraries: libboost_regex.so.1.49.0: cannot open shared object file: No such file or directory"
    I found out that this version of boost isnt available on Lubuntu 12.04.

    So I tried the same with Debian Wheezy, Now I can download the 1.49 version of boost, but the error is still there.

    What can I do to solve this problem? Is there a file where I can change the boost version? Because on UDOObuntu I have installed 1.46.1. Or is the version given by the developer of this program?

    I also can't figure out how to make rc.local run at boot on Debian Wheezy. On UDOObuntu it works perfectly so I would prefer using UDOObuntu if I can manage the boost problem.

    I hope somebody can help me. And as mentioned before, unfortunately I'm not a Linux Pro :roll:

    Thanks a lot and best regards :)
     
  2. Hans55

    Hans55 New Member

    Joined:
    Jan 8, 2015
    Messages:
    3
    Likes Received:
    0
    Nobody can help me? :shock:
     
  3. plugwash

    plugwash New Member

    Joined:
    Jan 20, 2015
    Messages:
    2
    Likes Received:
    0
    Hi

    Can you tell me what exactly you did? did you try and compile voicecommand from source or did you use a binary from somewhere?

    If a library is not being loaded for no obvious reason despite apparently existing it may be that you are trying to run a hardfloat binary on a softfloat system or vice-versa. Libraries using the wrong float ABI will be ignored.

    It's technically possible to have both hardfloat and softfloat libraries on the same system using multiarch but I don't know how well that combination works in pratice.
     
  4. Hans55

    Hans55 New Member

    Joined:
    Jan 8, 2015
    Messages:
    3
    Likes Received:
    0
    Hi

    Thanks for your reply. You are the moderator from the Pi forum right? :)

    I downloaded the PiAUSuite from here:
    https://github.com/StevenHickson/PiAUIS ... iceCommand

    And I installed directly the binaries from this download. So when I compile the source code by myself, it should work?
    As mentioned, I'm new on Linux, so can you tell me how to do it?

    And how can find out if I have a hard- or softfloat system? Tha makefile says, that it is for hardfloat...

    Thanks and best regards
    Hans
     
  5. plugwash

    plugwash New Member

    Joined:
    Jan 20, 2015
    Messages:
    2
    Likes Received:
    0
    Yes i'm the same plugwash who is among other things a mod on the raspberry pi forum.

    The command "dpkg --print-architecture" can be used to check if your system is using the softfloat or hardfloat ABIs. If it says "armel" then your system is using the soft float port as it's primary architecture if it says "armhf" then your system is using the hard float ABI as it's primary architecture.

    I haven't been following IMX6 graphical stuff recently but my understanding is that freescale still haven't made their graphics libraries available for hardfloat systems.

    In principle it's possible to mix armel stuff and armhf stuff on the same system using multiarch ( https://wiki.debian.org/Multiarch/ ) but as I said before I have no idea how well this works in practice.

    Normally to compile software you use the command "make" though it looks like in this case you will also need to edit the makefile as the author has hard-coded the ABI choice.

    Note that there are three settings for -mfloat-abi

    "soft": this uses software floating point and the soft float abi
    "softfp": this uses hardware floating point but still uses the soft float abi. This is probablly what you want to build the code for your system.
    "hard": this uses hardware floating point and the hard float ABI.

    You might also want to change the -mcpu setting to something more appropriate to your system (I think "cortex-a9" is a valid value and probablly what you want).

    P.S. to answer a question you asked elsewhere details of the debian-arm mailing list can be found at https://lists.debian.org/debian-arm/ . Theres also an irc channel #debian-arm on the oftc irc network.
     

Share This Page