Loading ...
Sorry, an error occurred while loading the content.

Re: [linuxham] Update + trying to get fldigi to work

Expand Messages
  • Rick Kunath
    Hello Rick, Haven t heard from you in a while, and glad you are making some headway. More comments inline... ... I am curious whether you got accelerated video
    Message 1 of 16 , Jun 3, 2007
    • 0 Attachment
      Hello Rick,

      Haven't heard from you in a while, and glad you are making some headway.

      More comments inline...

      On Monday 04 June 2007 1:22:44 pm Rick wrote:
      > Although I was never able to solve the problem of Linux not being able
      > to run on the open source Intel video on my Emachines computer, I have
      > recently purchased a fairly high end HP computer with MS Windows Vista.

      I am curious whether you got accelerated video going on the new HP machine?
      The latest Mandriva 2007.1 Spring release does handle the accelerated video
      on the Intel chipset, so that still might be an option for the Emachine.
      Though you have to read the errata on the Intel video for the release and
      turn it on. Intel HD audio also works on the 2007.1 Spring release.

      > Very nice interface appearance, although I have no idea what all the
      > excitement is about with Aero. By the same token, I have no idea why all
      > the excitement about Beryl either.

      I have to agree there. I don't like either myself.

      > I kept trying many different Linux distributions on the new HP and
      > finally did find that the newest Mepis 6.5 version can load the new
      > Nvidia drivers as a Live disk. While it can not equal the font rendering
      > and general sophistication of the MS product, it works far better than
      > the earlier problems I had with my widescreen LCD display. It now can
      > sort of work in native mode, but not quite. By that I mean that it will
      > display the common 1680x1050 rez that is used on most all 22" widescreen
      > monitors and although Mepis says it is sending it at 50 Hz, the monitor
      > claims it is at 60 Hz. (There is no 60 Hz option, nor is the monitor
      > listed (yet) in the drop down menu for Samsung Monitors).

      DVI or VGA to the monitor? And, the display of a real install (not a live-CD
      looks really nice on an Nvidia card. Is Mepis using the non-free libfreetype
      package?

      > When I use the
      > monitor with XP, the monitor menu display also shows NP and with Mepis
      > it is PP. I have not been able to figure out what those terms mean, but
      > it may explain why the screen rendering is not the best.

      What are the font anti-aliasing settings? This makes a lot of difference in
      font quality, as does the non-free libfreetype package. I don't know if Mepis
      supports either of these. Maybe someone else will know.

      > For now, I
      > removed Windows Vista and installed Mepis on the hard drive and continue
      > to have the two computers side by side with a KVM switch.

      Is font quality still an issue?

      > In the past I tried setting up fldigi when I had a Linux installation
      > (Ubuntu perhaps?) on the Emachnines and I was able to at least boot up
      > the program and copy signals with it. On the HP I had no luck with
      > getting the binary to work so decided to bite the bullet and find out
      > about all this source compiling and installation.
      >
      > After figuring out permission isses, I was able to download and
      > seemingly compile the flltk and hamlib sources and hopefully installed
      > them correctly.

      What *exactly* did you do here? also, how are the fftw (2.x and 3.x) libs
      compiled, and with what configure switches?

      > Then I tried to install Fldigi but after I am done
      > nothing happens when I click on the executable. If I do a make clean, I
      > notice that the executable is removed from my folder.

      Yes, that is correct for a make clean. This cleans out all of the temporary
      installation files, removes the executable, and makes ready for a new
      compilation.

      Have you opened a console in the fldigi directory and tried executing
      with ./fldigi from the command line? Some distros require this (Mandriva
      doesn't.)

      > Since nothing is crashing or showing any errors, I am not sure what to
      > do next to troubleshoot.

      Try the above and let us know what happens. If you still are stuck, try
      grabbing the hkj version of gMFSK and building that with the
      usual ./configure, make, su to root, make install routine. The configure
      output may clue you in to some missing packages, missing devels for packages,
      or other issues.

      > Is it correct that when you work with the
      > fldigi source code, after you untar the source, you do not do a
      > ./configure? You just do a make and not even a make install?

      Yes. That is correct for FLdigi.

      Rick Kunath, k9ao
    • w1hkj
      ... Yes. I would recommend the following. First, the relationship between COM ports and ttyS ports (case is IMPORTANT) COM1 == /dev/ttyS0 COM2 == /dev/ttyS1
      Message 2 of 16 , Jun 3, 2007
      • 0 Attachment
        Rick wrote:

        Everything was done according to your suggestions through the
        installation of both fltk and hamlib. While there is no sudo from what I
        can tell on Mepis (it is now Ubuntu based), I just went from the
        terminal to su and did the same thing. The /etc/ld.so.conf file was in
        place with the correct pointer to /usr/local/lib.

        I did the ldconfig command from root and exited root and ran the
        ./fldigi from the terminal in side the directory and it brought up the
        program.

        Then tried entering the parameters for the configuration but can not get
        anything to function as yet. Rig is selected as ICOM 756 Pro II, I have
        no idea about how to fill in the lines for /dev/ttySx and /dev/tty/USBx
        but put in the numbers in the help document screen shot and tried
        changing them but no luck. I have the +V set on DTR to power my homebrew
        2 transistor interface. I am using the old Radio Shack USB to COM
        adapter and I believe it is supported in the Linux kernel. Is there
        something else I have to do with that?

        Other suggestions?

        Yes.  I would recommend the following.

        First, the relationship between COM ports and ttyS ports (case is IMPORTANT)

        COM1 ==> /dev/ttyS0
        COM2 ==> /dev/ttyS1

        the USB-serial dongle is probably
        /dev/ttyUSB0

        Are you using the USB-serial dongle for the CAT interface to the rig?  It sounds like you have a CI-V DTR cheater interface.  I have the same thing here for the Icom 746 PRO.

        I would recommend that you begin with a rig.xml file that will probably work with your 756 PRO and forget about hamlib.  I will send you the rig.xml file as a direct attachment to your mrfarm email address.  It will be configured for the 756 PRO and /dev/ttyUSB0.  I will give you additional instructions in that email.

        73, Dave, W1HKJ
      • w1hkj
        Rick, Both of the precompiled fldigi binaries have been tested on Mepis and have been found to work FB. You probably could have avoided the build blues, but
        Message 3 of 16 , Jun 3, 2007
        • 0 Attachment
          Rick,

          Both of the precompiled fldigi binaries have been tested on Mepis and
          have been found to work FB. You probably could have avoided the build
          blues, but think of all the education you would have missed ;>)

          The recommendation to "make clean" is entirely dependent on the source
          code tree and the make file. It is always wise to preview the makefile
          to see what is happening. You would have seen that the "clean" deletes
          all of the object files and the executable. An install option could
          have been added to the makefile. I chose not to since many Linux hams
          already have a directory with binary files for amateur radio use. Most
          of the Linux distros will recognize $HOME/bin as a directory in which
          "local" binaries are stored.

          If you have a $HOME/bin but the fldigi executable there and your shell
          will always find it. You can also create a desktop icon that points to
          that executable and then change the icon attributes to use the
          "fldigi.png" file that is included in the source tree (../Install
          directory) and also can be found in the binary packages.

          73, Dave, W1HKJ
        • Rick Kunath
          Comments inline... ... This is with the Nvidia card, I assume? ... Certainly odd. I can t say as I have ever seen that one before. Does the blurring appear in
          Message 4 of 16 , Jun 3, 2007
          • 0 Attachment
            Comments inline...

            On Monday 04 June 2007 6:41:53 pm Rick wrote:
            > 1. I am not familiar with the accelerated video for Intel and spent a
            > lot of time on their site. One of the boot choices with Mepis was to use
            > the latest Nvidia driver and that brought up the higher quality video.

            This is with the Nvidia card, I assume?

            > I
            > was pretty elated and thought I was finally going to have acceptable
            > fonts. After installing Mepis, I downloaded the MS fonts to improve font
            > quality, but I get kind of a streaking effect every few inches on the
            > screen where it is fairly good quality and then an inch of lower quality
            > fonts that appear to be a kind of blurring artifact. Large images and
            > pictures, including very high resolution ones seem fine.

            Certainly odd. I can't say as I have ever seen that one before. Does the
            blurring appear in a saved full-screen screenshot, or only during live screen
            useage? It wouldn't be video tearing would it?

            > 2. The monitor is standard VGA since the computer doesn't have DVI
            > output and even if it does, I am using a KVM switch which only accepts VGA.
            >
            > I checked and have the libfreetype6 package installed

            Is the package built using autohinting (free), or the better (and patented)
            bytecode interpreter?

            What is the actual hardware DPI resolution of your LCD display? 96 DPI or
            maybe 100 DPI? What is the default setting for your Xfontserver set to, i.e.
            are the 75 DPI fonts set to load first or are the 100 DPI fonts higher in the
            load list in /etc/X11/fs/config? A good installer should have figured out
            the correct settings and applied then during install, but if this is wrong,
            you'll have sub-par fonts. This wouldn't cause the blurring though.

            > 3. I tried changing the font anti-aliasing and have it turned off with
            > smaller sizes since it did not help. Font quality is the main thing that
            > makes Linux unusable for practical work.

            At least in your distro. I haven't had font quality issues with Linux for many
            years. My fonts here have looked better than MS for a while now.

            > I have other things that I need
            > to work on such as getting it to print to my networked computer.

            Samba should handle that for you. Here I moved away from MS/Linux
            machine-based printer sharing and installed a separate small print server
            that operates as an LPD network print server. It smoothly integrates with my
            MS and Linux boxes, and is not tied to any one machine.

            > 4. The compiling of fltk and hamlib appears to have gone OK now that I
            > can at least boot up the fldigi program.
            >
            > 5. I am confused why some program installations recommend doing a make
            > clean after installation of the files if it deletes the executable.
            > (Linux Questions recommendations for installing from source code).

            It doesn't delete the executable installed by the make install command
            elsewhere in the file system (/usr, /usr/local, etc.) It will clean out the
            local build directory tree. I never issue a make clean after a build. I have
            plenty of hard drive space. A make clean can be useful to flush included
            tarball files not built from scratch on your particular box as a part of the
            initial build, i.e., ./configure (--add-whatever-options-you-need), make
            clean, make, su to root, make install, doing this sequence if there are build
            errors without the make clean. You'd use this instead of the
            usual ./configure, make, su to root, make install sequence.

            Often commands are needed to be passed to ./configure and doing
            a ./configure --help will get you a list of anything special and the standard
            switches. You might need to set some of these. You would have used some of
            these on your FLTK install if the binary on your distro was not built with
            everything you needed.

            > This whole source code thing is of course a double edged sword.
            > Tremendously powerful and really quite necessary if you want the latest
            > software updates since there is not one version of Linux and you need to
            > tailor make it for your distribution. But it is also something that
            > almost no computer users would ever be willing to do except for the real
            > diehard types like those of us on this group. Maybe we could get to the
            > point of having ham radio meetings to explain this stuff to the average
            > ham, but that is probably not realistic

            Sounds like a good idea. Generally, once an app stabilizes, gets feature
            complete, and needs less testing and cross distro testing, binary packages
            for the major distros can be offered that install easily.

            > 6. During the installation of fltk I got a error that it could not find
            > the right files and did a search on the internet for fltk + the error
            > message and found out that there were several possible dependency issues
            > and was able to compile once I installed libx11-dev, although one source
            > indicated it needed to be libx11-devel. These are the kinds of things
            > that make it nearly impossible for any serious adoption of Linux at this
            > level. It has to be more like installing MS software.

            It is with a properly functioning package installer. The exact package will be
            distro specific, so your package installer should be used to query the
            package list for the appropriate package. There are both command line as well
            as full-featured GUI based managers on some distros. In Mandriva 2007.1's
            case, had you asked the package installer for package names containing a
            portion of the above missing package, you'd have been presented with a list
            of installed packages, as well as not installed packages. This would have
            included obvious references to missing devel packages as well. You'd check
            off a few boxes, click OK, and be done with it.

            In the case of source compiled apps, you would have to identify the package
            you needed yourself. But this isn't too hard with practice. Had someone
            supplied you with an rpm for fldigi, the rpm itself would have instructed the
            package installer to automatically (requiring you to pre-approve the
            additions list) install any additional dependent packages that it needed to
            run. You'd have been presented with a list of packages that needed to be
            installed along with your rpm, been asked to OK the additional packages, and
            the lot would have been installed along with nice icons and a menu entry.
            Eventually, if the universal binary won't work on a specific distro, rpm
            packages can be created.

            You did great on this so far. And it will get lots easier with time. You did
            just right on hunting up the missing package depends. Hopefully, with a few
            more tweaks in place, you will start to get to the point where you feel your
            install is really shaping up.

            Rick Kunath, k9ao
          • Ken N9VV
            Hi I have a dual-core D940 1GB (recent vintage Dell PC) that is triple booted into XP + SimplyMEPIS (for Beryl support) and now DigiPup. I had to struggle with
            Message 5 of 16 , Jun 3, 2007
            • 0 Attachment
              Hi I have a dual-core D940 1GB (recent vintage Dell PC) that is
              triple booted into XP + SimplyMEPIS (for Beryl support) and now DigiPup.

              I had to struggle with it's messed up grub installer and then
              discovered that it will recognize the partition and try to boot from
              it but just dies a quiet death after a few spins of the HD.

              The DigiPup LiveCD came up like gangbusters but I was unable to get
              a single sound out of my mobo sound system (that works perfectly
              under SimplyMEPIS, Windows, and Ubuntu 7.04).

              Fldigi looks like a lot of fun if I can ever get over the ALSA-OSS
              hump. Several guys have suggested I buy a cheap sound card and use
              it instead of the mobo SigmaTel audio. I think there might be Linux
              support for the Delta-44(?). I will give it another go tomorrow when
              I am awake.

              This certainly must be why so few Hams are migrating to Linux
              instead of Vista ;-( It is a big hassle - it is like trying to
              build equipment instead of buying a commercial unit and just getting
              on the air! Perhaps the Linux guys spend a lot more time on their
              PC's than on the air in a real QSO(?).

              de Ken N9VV
            • Rick Kunath
              ... How are you triple-booting? Messing about with lilo or Grub, or using a dedicated bootloader and chaining (this is what I do)? ... The SigmaTel Intel sound
              Message 6 of 16 , Jun 3, 2007
              • 0 Attachment
                Ken N9VV wrote:
                > Hi I have a dual-core D940 1GB (recent vintage Dell PC) that is
                > triple booted into XP + SimplyMEPIS (for Beryl support) and now DigiPup.

                How are you triple-booting? Messing about with lilo or Grub, or using a
                dedicated bootloader and chaining (this is what I do)?

                > Fldigi looks like a lot of fun if I can ever get over the ALSA-OSS
                > hump. Several guys have suggested I buy a cheap sound card and use
                > it instead of the mobo SigmaTel audio. I think there might be Linux
                > support for the Delta-44(?). I will give it another go tomorrow when
                > I am awake.

                The SigmaTel Intel sound on Intel motherboards I have used works out of
                the box on Mandriva 2007.1, but I don't know about your Dell
                motherboard. Dell has been known to use a similarly named product but
                have different actual hardware. Add the alsa-oss emulation package and
                you are all set. Support for that card is recent though. As to card
                support it depends on the ALSA version and included kernel modules built
                into the distro. A well supported sound card is a good investment though.

                See what the card identifier is for your card and see if it matches up
                with the supported SigmaTel cards on the ALSA site. The site will tell
                you what ALSA version began supporting the card, so you can see what a
                particular distro includes, ALSA release version wise.

                > This certainly must be why so few Hams are migrating to Linux
                > instead of Vista ;-( It is a big hassle - it is like trying to
                > build equipment instead of buying a commercial unit and just getting
                > on the air!

                Linux is still about supported hardware. If you don't have it, you are
                in for a harder time. It's the same for Vista.

                > Perhaps the Linux guys spend a lot more time on their
                > PC's than on the air in a real QSO(?).

                Depends on the distro. Pop in the Mandriva powerpack DVD, answer a few
                questions, in about a half-hour you will have working 3d acceleration,
                multimedia, web browsing and email, lots more. Grab the FLdigi binary,
                unpack it, click the executable, and you're off.

                I had it going in about that on a test Ubuntu installation too. Though I
                didn't have the working 3d acceleration and multimedia with all the
                proprietary codecs working then.

                It all depends on your hardware and the chosen distro, and what is
                included, driver wise.

                Rick Kunath, k9ao
              • w1hkj
                1. I found that I had entered /dev/tty/USB0 instead of /dev/ttyUSB0. Ans. Linux treats every serail port in a uniform way. All device drivers are located in
                Message 7 of 16 , Jun 4, 2007
                • 0 Attachment
                  1. I found that I had entered /dev/tty/USB0 instead of /dev/ttyUSB0.

                  Ans. Linux treats every serail port in a uniform way.  All device drivers are located in the /dev directory.  The tty part of the name indicates a terminal device that obeys certain character oriented rules for the data stream.  The Sx (S0, S1 etc) indicate that the hardware is a serial port  ergo ttyS0, ttyS1 etc.  If the device name is appended with the USBx it indicates the the hardware is a USB type which give the names ttyUSB0, ttyUSB1 etc.  The determination of which physical device is designated by 0, 1, 2 etc is made during boot when the OS discovers what hardware is on the machine.  If you replace a USB device the OS can find most on a plug and play basis.  If you added a COM type of serial device that discovery must be made at boot time.

                  2. Is this /dev/ttyUSBx the USB side of the port only or is this associated with the COM port?

                  Ans.  You can learn more about which physical device is associated with which logical device  on an Ubuntu based system using the Device Manager gui application.  My mother board sound system, an AC'97 audio controller from Intel, shows up as /dev/dsp1 and /dev/mixer1 for the sound and mixer.  The SignaLink USB interface is a USB Audio interface, Texas Insturments Japan, and is logical device(s) /dev/dsp and /dev/mixer

                  3. If I am using Hamlib, do I need to enter a /dev/tty/Sx? Or just check on the highlight box for Hamlib on the left side?

                  Ans.   Remember that the device name is /dev/ttySx.  If you are using hamlib and not the rig.xml (rigcat) access method to gain access to the rig then you will need to specify the port associated with the serial device that hamlib will find the rig.  That is entered on the right hand side of the RigCtl panel.  A caveat ... the hamlib interface code for the Icom rigs was written with the expectation that you would have a real CI-V interface and not one which steals power from the DTR or RTS pin.  rigcat i/o recognizes the need to retain DTR at +12V when the rig.xml file so specifies.  rigcat can therefore be used with your homebrew interface and not hamlib.

                  4. What is normal response time for the updating of the rig display frequency? It often takes a very long time, >5 to 10 seconds to update
                  but sometimes very quickly.
                  5. Is this possibly due to the fact that I am not providing the right power to the transistor circuitry because I don't have the /dev/tty/Sx
                  on the right setting?

                  Ans.  The rig is queried every 50 msec for new data.  I have tested this on an IC-756PRO-II and have seen no problems associated with latency.  I am controlling the rig via /dev/ttyS0 which is physically connected to a homebrew CI-V interface which steals power from DTR.  If you are using the rig.xml file which I provided on a private email I suggest leaving the left hand side of the RigCtl panel blank with the exception of the selection of RigCAT for push to talk.

                  6. I still can not get the audio to connect up to the program. I am pretty sure I am inputting it OK on the Line in and Line out. I have the OSS simulation installed for my ALSA audio.

                  Ans.  Try the audio editor Audacity ( http://audacity.sourceforge.net ) to see if you're sound system is working OK with the OS.

                  Dave, W1HKJ

                • Rick
                  Everything was done according to your suggestions through the installation of both fltk and hamlib. While there is no sudo from what I can tell on Mepis (it is
                  Message 8 of 16 , Jun 4, 2007
                  • 0 Attachment
                    Everything was done according to your suggestions through the
                    installation of both fltk and hamlib. While there is no sudo from what I
                    can tell on Mepis (it is now Ubuntu based), I just went from the
                    terminal to su and did the same thing. The /etc/ld.so.conf file was in
                    place with the correct pointer to /usr/local/lib.

                    I did the ldconfig command from root and exited root and ran the
                    ./fldigi from the terminal in side the directory and it brought up the
                    program.

                    Then tried entering the parameters for the configuration but can not get
                    anything to function as yet. Rig is selected as ICOM 756 Pro II, I have
                    no idea about how to fill in the lines for /dev/ttySx and /dev/tty/USBx
                    but put in the numbers in the help document screen shot and tried
                    changing them but no luck. I have the +V set on DTR to power my homebrew
                    2 transistor interface. I am using the old Radio Shack USB to COM
                    adapter and I believe it is supported in the Linux kernel. Is there
                    something else I have to do with that?

                    Other suggestions?

                    73,

                    Rick, KV9U

                    w1hkj wrote:
                    > Congratulations on moving closer to having Linux and fldigi running on
                    > your wide screen display Rick.
                    >
                    > I have to assume that you have fltk built and installed with the
                    > required switches as per the on-line help for fldigi.
                    >
                    > I also am going to assume that you have hamlib built and installed.
                    > You know that the final stage in building and installing both fltk and
                    > hamlib is to perform the following (from the top source code directory
                    > in the corresponding directory).
                    >
                    > sudo make install
                    >
                    > then make sure that the file *//etc/ld.so.conf/* has the line
                    >
                    > /usr/local/lib
                    >
                    > contained therein.
                    >
                    > then execute
                    >
                    > sudo ldconfig
                    >
                    > That tells the operating system where to find the newly installed
                    > dynamic libraries.
                    >
                    > Now. When you make fldigi you apparantly got an executable or would
                    > have received error messages that complained of missing files etc.
                    >
                    > Don't execute the program by clicking on the file from the gui.
                    > Instead execute from within the directory in which the executable
                    > resides. Do this from a terminal window:
                    >
                    > ./fldigi
                    >
                    > The ./ tells the shell to look for the program in the current directory.
                    >
                    > Tell me what messages are displayed after trying the execution.
                    >
                    >
                    >
                    >
                  • Rick
                    1. I am not familiar with the accelerated video for Intel and spent a lot of time on their site. One of the boot choices with Mepis was to use the latest
                    Message 9 of 16 , Jun 4, 2007
                    • 0 Attachment
                      1. I am not familiar with the accelerated video for Intel and spent a
                      lot of time on their site. One of the boot choices with Mepis was to use
                      the latest Nvidia driver and that brought up the higher quality video. I
                      was pretty elated and thought I was finally going to have acceptable
                      fonts. After installing Mepis, I downloaded the MS fonts to improve font
                      quality, but I get kind of a streaking effect every few inches on the
                      screen where it is fairly good quality and then an inch of lower quality
                      fonts that appear to be a kind of blurring artifact. Large images and
                      pictures, including very high resolution ones seem fine.

                      2. The monitor is standard VGA since the computer doesn't have DVI
                      output and even if it does, I am using a KVM switch which only accepts VGA.

                      I checked and have the libfreetype6 package installed

                      3. I tried changing the font anti-aliasing and have it turned off with
                      smaller sizes since it did not help. Font quality is the main thing that
                      makes Linux unusable for practical work. I have other things that I need
                      to work on such as getting it to print to my networked computer.

                      4. The compiling of fltk and hamlib appears to have gone OK now that I
                      can at least boot up the fldigi program.

                      5. I am confused why some program installations recommend doing a make
                      clean after installation of the files if it deletes the executable.
                      (Linux Questions recommendations for installing from source code).

                      This whole source code thing is of course a double edged sword.
                      Tremendously powerful and really quite necessary if you want the latest
                      software updates since there is not one version of Linux and you need to
                      tailor make it for your distribution. But it is also something that
                      almost no computer users would ever be willing to do except for the real
                      diehard types like those of us on this group. Maybe we could get to the
                      point of having ham radio meetings to explain this stuff to the average
                      ham, but that is probably not realistic.

                      6. During the installation of fltk I got a error that it could not find
                      the right files and did a search on the internet for fltk + the error
                      message and found out that there were several possible dependency issues
                      and was able to compile once I installed libx11-dev, although one source
                      indicated it needed to be libx11-devel. These are the kinds of things
                      that make it nearly impossible for any serious adoption of Linux at this
                      level. It has to be more like installing MS software. A lot of this is
                      relative. I remember when I first got into computing with a C-64 and
                      shortly after an Apple //e. The learning curve was pretty large and
                      would not appeal to most people. I ran an audiovisual repair facility
                      for many years and eventually repaired some of the computers and also
                      worked extensively with Corvus networking systems, so if you know about
                      such things, it was about as user unfriendly as you can get. Just
                      loading software unto the server was a really big deal.

                      73,

                      Rick, KV9U




                      Rick Kunath wrote:
                      > Hello Rick,
                      >
                      > Haven't heard from you in a while, and glad you are making some headway.
                      >
                      > More comments inline...
                      >
                      > On Monday 04 June 2007 1:22:44 pm Rick wrote:
                      >
                      >> Although I was never able to solve the problem of Linux not being able
                      >> to run on the open source Intel video on my Emachines computer, I have
                      >> recently purchased a fairly high end HP computer with MS Windows Vista.
                      >>
                      >
                      > I am curious whether you got accelerated video going on the new HP machine?
                      > The latest Mandriva 2007.1 Spring release does handle the accelerated video
                      > on the Intel chipset, so that still might be an option for the Emachine.
                      > Though you have to read the errata on the Intel video for the release and
                      > turn it on. Intel HD audio also works on the 2007.1 Spring release.
                      >
                      >
                      >> Very nice interface appearance, although I have no idea what all the
                      >> excitement is about with Aero. By the same token, I have no idea why all
                      >> the excitement about Beryl either.
                      >>
                      >
                      > I have to agree there. I don't like either myself.
                      >
                      >
                      >> I kept trying many different Linux distributions on the new HP and
                      >> finally did find that the newest Mepis 6.5 version can load the new
                      >> Nvidia drivers as a Live disk. While it can not equal the font rendering
                      >> and general sophistication of the MS product, it works far better than
                      >> the earlier problems I had with my widescreen LCD display. It now can
                      >> sort of work in native mode, but not quite. By that I mean that it will
                      >> display the common 1680x1050 rez that is used on most all 22" widescreen
                      >> monitors and although Mepis says it is sending it at 50 Hz, the monitor
                      >> claims it is at 60 Hz. (There is no 60 Hz option, nor is the monitor
                      >> listed (yet) in the drop down menu for Samsung Monitors).
                      >>
                      >
                      > DVI or VGA to the monitor? And, the display of a real install (not a live-CD
                      > looks really nice on an Nvidia card. Is Mepis using the non-free libfreetype
                      > package?
                      >
                      >
                      >> When I use the
                      >> monitor with XP, the monitor menu display also shows NP and with Mepis
                      >> it is PP. I have not been able to figure out what those terms mean, but
                      >> it may explain why the screen rendering is not the best.
                      >>
                      >
                      > What are the font anti-aliasing settings? This makes a lot of difference in
                      > font quality, as does the non-free libfreetype package. I don't know if Mepis
                      > supports either of these. Maybe someone else will know.
                      >
                      >
                      >> For now, I
                      >> removed Windows Vista and installed Mepis on the hard drive and continue
                      >> to have the two computers side by side with a KVM switch.
                      >>
                      >
                      > Is font quality still an issue?
                      >
                      >
                      >> In the past I tried setting up fldigi when I had a Linux installation
                      >> (Ubuntu perhaps?) on the Emachnines and I was able to at least boot up
                      >> the program and copy signals with it. On the HP I had no luck with
                      >> getting the binary to work so decided to bite the bullet and find out
                      >> about all this source compiling and installation.
                      >>
                      >> After figuring out permission isses, I was able to download and
                      >> seemingly compile the flltk and hamlib sources and hopefully installed
                      >> them correctly.
                      >>
                      >
                      > What *exactly* did you do here? also, how are the fftw (2.x and 3.x) libs
                      > compiled, and with what configure switches?
                      >
                      >
                      >> Then I tried to install Fldigi but after I am done
                      >> nothing happens when I click on the executable. If I do a make clean, I
                      >> notice that the executable is removed from my folder.
                      >>
                      >
                      > Yes, that is correct for a make clean. This cleans out all of the temporary
                      > installation files, removes the executable, and makes ready for a new
                      > compilation.
                      >
                      > Have you opened a console in the fldigi directory and tried executing
                      > with ./fldigi from the command line? Some distros require this (Mandriva
                      > doesn't.)
                      >
                      >
                      >> Since nothing is crashing or showing any errors, I am not sure what to
                      >> do next to troubleshoot.
                      >>
                      >
                      > Try the above and let us know what happens. If you still are stuck, try
                      > grabbing the hkj version of gMFSK and building that with the
                      > usual ./configure, make, su to root, make install routine. The configure
                      > output may clue you in to some missing packages, missing devels for packages,
                      > or other issues.
                      >
                      >
                      >> Is it correct that when you work with the
                      >> fldigi source code, after you untar the source, you do not do a
                      >> ./configure? You just do a make and not even a make install?
                      >>
                      >
                      > Yes. That is correct for FLdigi.
                      >
                      > Rick Kunath, k9ao
                      >
                      >
                      >
                      >
                      >
                      >
                      >
                      >
                      >
                      >
                    • Rick
                      Hi Again, 1. The only computer I have been working with the past few weeks has been the new HP since I discovered that Mepis 6.5 (but not the earlier 6.0)
                      Message 10 of 16 , Jun 4, 2007
                      • 0 Attachment
                        Hi Again,

                        1. The only computer I have been working with the past few weeks has
                        been the new HP since I discovered that Mepis 6.5 (but not the earlier
                        6.0) would drive the Nvidia. The computer is the Pavilion a1730N. I got
                        it because I think it has one of the best price points at this time with
                        an AMD Athlon 4600+ 64 X2 dual core, 2 gig memory, 320 gig HD, and the
                        Nvidia GeForce 6150LE graphics. I had more positive comments about
                        Nvidia video for Linux than most other products.

                        2. The blurring of the font is fairly subtle, but quite distressing when
                        using to do things such as view a web page or a document that has a lot
                        of writing, especially with small font sizes. It sort of looks like a
                        smudging effect and drastically reduces readability. I would like to
                        know what the NP and PP refer to in the video menu Information as that
                        seems to be the only difference between XP and Mepis driving the
                        identical screen.

                        I could not find the hardware font size. I did not find any /etc/X11/fs
                        file so possibly different from other versions of Linux? I have the X11
                        setup to display 96 dpi. The other choices are 75 and 120 and they don't
                        really make much difference for web pages if you accept the web page font.

                        3. If your fonts look better than MS OS's that seems contrary to what
                        even the Linux devotees admit about the fonts. As a Mandriva user, this
                        may be a better version of Linux that attends to that detail? Many Linux
                        references are honest and point out the problems with the bad fonts and
                        show how to bring in the Windows fonts to get higher quality and so I
                        followed their recommendations and it did improve things over the
                        initial fonts. These are some of the things that are really impacting
                        Linux the most. There are certain things that have to work out of the
                        box and it just does not seem to be there at this time. That doesn't
                        mean that they won't get there eventually. Vista has better font
                        rendering than XP due to some ultra tweaked fonts. I bet they spent a
                        few million with that alone. I wish the energies that are spent on Beryl
                        and other eye candy, would focus on basic enhancements though but I
                        suppose it is exciting to work on the new cool things.

                        6. I don't know what to believe about rpm's after the recent blow up by
                        ESR over Red Hat and his conversion to Ubuntu due to dependency hell. At
                        this point this is still a toy for me or I would have let it go a long
                        time ago and gotten some actual work done.:) I am still intrigued by the
                        one program that is only available on Linux (PSKmail) but it sounds like
                        something else is in the works for a cross platform type of program.

                        73,

                        Rick, KV9U



                        Rick Kunath wrote:
                        > Comments inline...
                        >
                        > On Monday 04 June 2007 6:41:53 pm Rick wrote:
                        >
                        >> 1. I am not familiar with the accelerated video for Intel and spent a
                        >> lot of time on their site. One of the boot choices with Mepis was to use
                        >> the latest Nvidia driver and that brought up the higher quality video.
                        >>
                        >
                        > This is with the Nvidia card, I assume?
                        >
                        >
                        >> I
                        >> was pretty elated and thought I was finally going to have acceptable
                        >> fonts. After installing Mepis, I downloaded the MS fonts to improve font
                        >> quality, but I get kind of a streaking effect every few inches on the
                        >> screen where it is fairly good quality and then an inch of lower quality
                        >> fonts that appear to be a kind of blurring artifact. Large images and
                        >> pictures, including very high resolution ones seem fine.
                        >>
                        >
                        > Certainly odd. I can't say as I have ever seen that one before. Does the
                        > blurring appear in a saved full-screen screenshot, or only during live screen
                        > useage? It wouldn't be video tearing would it?
                        >
                        >
                        >> 2. The monitor is standard VGA since the computer doesn't have DVI
                        >> output and even if it does, I am using a KVM switch which only accepts VGA.
                        >>
                        >> I checked and have the libfreetype6 package installed
                        >>
                        >
                        > Is the package built using autohinting (free), or the better (and patented)
                        > bytecode interpreter?
                        >
                        > What is the actual hardware DPI resolution of your LCD display? 96 DPI or
                        > maybe 100 DPI? What is the default setting for your Xfontserver set to, i.e.
                        > are the 75 DPI fonts set to load first or are the 100 DPI fonts higher in the
                        > load list in /etc/X11/fs/config? A good installer should have figured out
                        > the correct settings and applied then during install, but if this is wrong,
                        > you'll have sub-par fonts. This wouldn't cause the blurring though.
                        >
                        >
                        >> 3. I tried changing the font anti-aliasing and have it turned off with
                        >> smaller sizes since it did not help. Font quality is the main thing that
                        >> makes Linux unusable for practical work.
                        >>
                        >
                        > At least in your distro. I haven't had font quality issues with Linux for many
                        > years. My fonts here have looked better than MS for a while now.
                        >
                        >
                        >> I have other things that I need
                        >> to work on such as getting it to print to my networked computer.
                        >>
                        >
                        > Samba should handle that for you. Here I moved away from MS/Linux
                        > machine-based printer sharing and installed a separate small print server
                        > that operates as an LPD network print server. It smoothly integrates with my
                        > MS and Linux boxes, and is not tied to any one machine.
                        >
                        >
                        >> 4. The compiling of fltk and hamlib appears to have gone OK now that I
                        >> can at least boot up the fldigi program.
                        >>
                        >> 5. I am confused why some program installations recommend doing a make
                        >> clean after installation of the files if it deletes the executable.
                        >> (Linux Questions recommendations for installing from source code).
                        >>
                        >
                        > It doesn't delete the executable installed by the make install command
                        > elsewhere in the file system (/usr, /usr/local, etc.) It will clean out the
                        > local build directory tree. I never issue a make clean after a build. I have
                        > plenty of hard drive space. A make clean can be useful to flush included
                        > tarball files not built from scratch on your particular box as a part of the
                        > initial build, i.e., ./configure (--add-whatever-options-you-need), make
                        > clean, make, su to root, make install, doing this sequence if there are build
                        > errors without the make clean. You'd use this instead of the
                        > usual ./configure, make, su to root, make install sequence.
                        >
                        > Often commands are needed to be passed to ./configure and doing
                        > a ./configure --help will get you a list of anything special and the standard
                        > switches. You might need to set some of these. You would have used some of
                        > these on your FLTK install if the binary on your distro was not built with
                        > everything you needed.
                        >
                        >
                        >> This whole source code thing is of course a double edged sword.
                        >> Tremendously powerful and really quite necessary if you want the latest
                        >> software updates since there is not one version of Linux and you need to
                        >> tailor make it for your distribution. But it is also something that
                        >> almost no computer users would ever be willing to do except for the real
                        >> diehard types like those of us on this group. Maybe we could get to the
                        >> point of having ham radio meetings to explain this stuff to the average
                        >> ham, but that is probably not realistic
                        >>
                        >
                        > Sounds like a good idea. Generally, once an app stabilizes, gets feature
                        > complete, and needs less testing and cross distro testing, binary packages
                        > for the major distros can be offered that install easily.
                        >
                        >
                        >> 6. During the installation of fltk I got a error that it could not find
                        >> the right files and did a search on the internet for fltk + the error
                        >> message and found out that there were several possible dependency issues
                        >> and was able to compile once I installed libx11-dev, although one source
                        >> indicated it needed to be libx11-devel. These are the kinds of things
                        >> that make it nearly impossible for any serious adoption of Linux at this
                        >> level. It has to be more like installing MS software.
                        >>
                        >
                        > It is with a properly functioning package installer. The exact package will be
                        > distro specific, so your package installer should be used to query the
                        > package list for the appropriate package. There are both command line as well
                        > as full-featured GUI based managers on some distros. In Mandriva 2007.1's
                        > case, had you asked the package installer for package names containing a
                        > portion of the above missing package, you'd have been presented with a list
                        > of installed packages, as well as not installed packages. This would have
                        > included obvious references to missing devel packages as well. You'd check
                        > off a few boxes, click OK, and be done with it.
                        >
                        > In the case of source compiled apps, you would have to identify the package
                        > you needed yourself. But this isn't too hard with practice. Had someone
                        > supplied you with an rpm for fldigi, the rpm itself would have instructed the
                        > package installer to automatically (requiring you to pre-approve the
                        > additions list) install any additional dependent packages that it needed to
                        > run. You'd have been presented with a list of packages that needed to be
                        > installed along with your rpm, been asked to OK the additional packages, and
                        > the lot would have been installed along with nice icons and a menu entry.
                        > Eventually, if the universal binary won't work on a specific distro, rpm
                        > packages can be created.
                        >
                        > You did great on this so far. And it will get lots easier with time. You did
                        > just right on hunting up the missing package depends. Hopefully, with a few
                        > more tweaks in place, you will start to get to the point where you feel your
                        > install is really shaping up.
                        >
                        > Rick Kunath, k9ao
                        >
                        >
                      • Rick
                        I knew about the case sensitive issue with the tty ports from a previous post. I seem to recall that the COM port is 3 or 5 with that interface when I used it
                        Message 11 of 16 , Jun 4, 2007
                        • 0 Attachment
                          I knew about the case sensitive issue with the tty ports from a previous
                          post. I seem to recall that the COM port is 3 or 5 with that interface
                          when I used it with XP so I wonder if it would be hard wired that way?

                          I tried using S0 through S9 I think it was but only changed them and
                          clicked on the initialize button. Maybe I have to do more?

                          The USB adapter then plugs into my HB 2 transistor interface. This seems
                          to be getting a bit complicated. I would prefer trying to get this to
                          work in what I would consider a normal manner like I would with anything
                          I have done over the years with windows/C-64/Apple //e.

                          Is there a way to tell what COM port is being used? and USB port? Or
                          whatever the device associated with /dev/ttyUSB0?

                          I do not have my Linux machine connected for e-mail but may try and
                          sneaker net this file from this XP machine to the Linux machine later on.

                          Thanks for the help.

                          73,

                          Rick, KV9U


                          w1hkj wrote:
                          > Yes. I would recommend the following.
                          >
                          > First, the relationship between COM ports and ttyS ports (case is
                          > IMPORTANT)
                          >
                          > COM1 ==> /dev/ttyS0
                          > COM2 ==> /dev/ttyS1
                          >
                          > the USB-serial dongle is probably
                          > /dev/ttyUSB0
                          >
                          > Are you using the USB-serial dongle for the CAT interface to the rig?
                          > It sounds like you have a CI-V DTR cheater interface. I have the same
                          > thing here for the Icom 746 PRO.
                          >
                          > I would recommend that you begin with a rig.xml file that will
                          > probably work with your 756 PRO and forget about hamlib. I will send
                          > you the rig.xml file as a direct attachment to your mrfarm email
                          > address. It will be configured for the 756 PRO and /dev/ttyUSB0. I
                          > will give you additional instructions in that email.
                          >
                          > 73, Dave, W1HKJ
                          > ------------------------------------------------------------------------
                          >
                          > No virus found in this incoming message.
                          > Checked by AVG Free Edition.
                          > Version: 7.5.472 / Virus Database: 269.8.7/830 - Release Date: 6/3/2007 12:47 PM
                          >
                        • Rick
                          1. I found that I had entered /dev/tty/USB0 instead of /dev/ttyUSB0. It turns out that you can quickly tell which command to use since the rig on the task bar
                          Message 12 of 16 , Jun 4, 2007
                          • 0 Attachment
                            1. I found that I had entered /dev/tty/USB0 instead of /dev/ttyUSB0. It
                            turns out that you can quickly tell which command to use since the rig
                            on the task bar will become active as soon as you enter the right
                            combination and initialize. Maybe this is in the docs and I missed it? I
                            discovered that the old USB to COM adapter from Radio Shack is operating
                            at /dev/ttyUSB0 and my newer no name ultra low cost unit (sold
                            everywhere:) requires /dev/ttyUSB1.

                            2. Is this /dev/ttyUSBx the USB side of the port only or is this
                            associated with the COM port?

                            3. If I am using Hamlib, do I need to enter a /dev/tty/Sx? Or just check
                            on the highlight box for Hamlib on the left side?

                            4. What is normal response time for the updating of the rig display
                            frequency? It often takes a very long time, >5 to 10 seconds to update
                            but sometimes very quickly. The transmit takes about 5 seconds to key up
                            after clicking on either of TX buttons and it does not seem to be able
                            to return to RX. Curiously, when I click on the list of stored
                            frequencies, it always switches instantly to the correct freq. Same for
                            mode changes.

                            5. Is this possibly due to the fact that I am not providing the right
                            power to the transistor circuitry because I don't have the /dev/tty/Sx
                            on the right setting?

                            6. I still can not get the audio to connect up to the program. I am
                            pretty sure I am inputting it OK on the Line in and Line out. I have the
                            OSS simulation installed for my ALSA audio.

                            Thanks for all the help,

                            73,

                            Rick, KV9U
                          • Rick
                            If I look in the /dev directory, I can not begin to understand what I would need to look at to determine the ports that are in use. Even using the KDE Info
                            Message 13 of 16 , Jun 5, 2007
                            • 0 Attachment
                              If I look in the /dev directory, I can not begin to understand what I
                              would need to look at to determine the ports that are in use. Even using
                              the KDE Info Center was of minimal help as it does not specify what is
                              active. It does find the specific name of the USB manufacturers product.

                              I am not clear on whether or not I need to have the exact correct COM
                              port specified on the left side if I am using hamlib. I have the hamlib
                              checked on the right side and the highlighted red activation on the left
                              side.

                              In other words, if I have the DTR +12 v checked, is the program going to
                              provide the required voltage or does it need to know the COM port
                              specifications?

                              I discovered on the hamlib web site that you can go into terminal mode
                              and run the rigctl commands. I tried this out and using the /dev/ttyUSB0
                              was able to get it to work instantly. Similarly, when I use the GUI
                              Hamlib commands, it responds quickly from the Hamlib side, but it is
                              sluggish from the rig side back to Hamlib.

                              The fldigi program then, does not respond well to frequency changes in
                              the rig, taking a long time or sometimes never updating the frequency.
                              Also, it can trigger PTT after a fairly long wait, but seems unable to
                              unkey the PTT once it is activated.

                              I am getting a consistent error on the bottom of the fldigi program
                              indicating "Hamlib loop Get Mode: Hamlib getMode error" so something is
                              not quite right yet.

                              After getting the Audacity package installed I was finally able to track
                              down the executable and run the program and it works OK picking up the
                              Line Input. As mentioned earlier, I am not getting any audio to display
                              on the fldigi program, so I must be missing something.

                              73,

                              Rick, KV9U


                              w1hkj wrote:
                              > 1. I found that I had entered /dev/tty/USB0 instead of /dev/ttyUSB0.
                              >
                              > Ans. Linux treats every serail port in a uniform way. All device
                              > drivers are located in the /dev directory. The tty part of the name
                              > indicates a terminal device that obeys certain character oriented
                              > rules for the data stream. The Sx (S0, S1 etc) indicate that the
                              > hardware is a serial port ergo ttyS0, ttyS1 etc. If the device name
                              > is appended with the USBx it indicates the the hardware is a USB type
                              > which give the names ttyUSB0, ttyUSB1 etc. The determination of which
                              > physical device is designated by 0, 1, 2 etc is made during boot when
                              > the OS discovers what hardware is on the machine. If you replace a
                              > USB device the OS can find most on a plug and play basis. If you
                              > added a COM type of serial device that discovery must be made at boot
                              > time.
                              >
                              > 2. Is this /dev/ttyUSBx the USB side of the port only or is this
                              > associated with the COM port?
                              >
                              > Ans. You can learn more about which physical device is associated
                              > with which logical device on an Ubuntu based system using the Device
                              > Manager gui application. My mother board sound system, an AC'97 audio
                              > controller from Intel, shows up as /dev/dsp1 and /dev/mixer1 for the
                              > sound and mixer. The SignaLink USB interface is a USB Audio
                              > interface, Texas Insturments Japan, and is logical device(s) /dev/dsp
                              > and /dev/mixer
                              >
                              > 3. If I am using Hamlib, do I need to enter a */dev/tty/Sx?* Or just
                              > check on the highlight box for Hamlib on the left side?
                              >
                              > Ans. Remember that the device name is */dev/ttySx*. If you are
                              > using hamlib and not the rig.xml (rigcat) access method to gain access
                              > to the rig then you will need to specify the port associated with the
                              > serial device that hamlib will find the rig. That is entered on the
                              > right hand side of the RigCtl panel. A caveat ... the hamlib
                              > interface code for the Icom rigs was written with the expectation that
                              > you would have a real CI-V interface and not one which steals power
                              > from the DTR or RTS pin. rigcat i/o recognizes the need to retain DTR
                              > at +12V when the rig.xml file so specifies. rigcat can therefore be
                              > used with your homebrew interface and not hamlib.
                              >
                              > 4. What is normal response time for the updating of the rig display
                              > frequency? It often takes a very long time, >5 to 10 seconds to update
                              > but sometimes very quickly.
                              > 5. Is this possibly due to the fact that I am not providing the right
                              > power to the transistor circuitry because I don't have the */dev/tty/Sx*
                              > on the right setting?
                              >
                              > Ans. The rig is queried every 50 msec for new data. I have tested
                              > this on an IC-756PRO-II and have seen no problems associated with
                              > latency. I am controlling the rig via /dev/ttyS0 which is physically
                              > connected to a homebrew CI-V interface which steals power from DTR.
                              > If you are using the rig.xml file which I provided on a private email
                              > I suggest leaving the left hand side of the RigCtl panel blank with
                              > the exception of the selection of RigCAT for push to talk.
                              >
                              > 6. I still can not get the audio to connect up to the program. I am
                              > pretty sure I am inputting it OK on the Line in and Line out. I have
                              > the OSS simulation installed for my ALSA audio.
                              >
                              > Ans. Try the audio editor Audacity ( http://audacity.sourceforge.net
                              > ) to see if you're sound system is working OK with the OS.
                              >
                              > Dave, W1HKJ
                              >
                              >
                              > ------------------------------------------------------------------------
                              >
                              > No virus found in this incoming message.
                              > Checked by AVG Free Edition.
                              > Version: 7.5.472 / Virus Database: 269.8.7/830 - Release Date: 6/3/2007 12:47 PM
                              >
                            Your message has been successfully submitted and would be delivered to recipients shortly.