<html>
  <head>
    <meta content="text/html; charset=utf-8" http-equiv="Content-Type">
  </head>
  <body bgcolor="#FFFFFF" text="#000000">
    According to the ncurses 6.0 release notes (the site is at
    <a class="moz-txt-link-rfc2396E" href="https://www.gnu.org/software/ncurses/">"https://www.gnu.org/software/ncurses/"</a>), the intent of the release
    is to provide a new stable ABI, <br>
    which is binary-incompatible with ncurses 5, <font color="#3333ff">but
      it can still be configured to support the ncurses5</font>. 
    Because this release introduces <font color="#3333ff">symbol-versioning</font>.
    <br>
    The <font color="#3333ff">symbol-versioning </font>allows the
    developers of a library tomark each public sysmbol(both data and
    functions) with an identifier denoting the library name and <br>
    the version for which it was built. By doing this , users ofthe
    library have a way to help ensure that applications do not
    accidentally load an imcompatible library. <br>
    In addition, private sysmbols can be hidden entirely.<br>
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------<br>
    When I first release the patch 5.9 -&gt; 6.0  with the option
    --with-abi-version=5, I want to be compatible with the existing
    system, so I do the following test:<br>
    <br>
        $bitbake world core-image-sato-sdk<br>
    <br>
    tested it on the qemuarm, qemux86, qemux86-64 with multilib
    (oe-core), and no error happened .<br>
    <br>
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------<br>
    When I second release the patch "drop the compatible option
    --with-abi-version=5", because I want to introduces the feature <font
      color="#3333ff">symbol-versioning<font color="#000000">, so I
        droped the <br>
        option --with-abi-version=5 , and I do the following test:<br>
        <br>
            $bitbake world core-image-sato-sdk<br>
        <br>
        tested it on the qemuarm, qemux86, qemux86-64 with multilib
        (oe-core), and no error happened.<br>
        <br>
        Today, Robert let me test it on the following layers:<br>
        <br>
            /meta-openembedded/meta-oe <br>
            /meta-openembedded/meta-networking<br>
      </font></font><font color="#3333ff"><font color="#000000"><font
          color="#3333ff"><font color="#000000">   
            /meta-openembedded/meta-meta-webserver</font></font><font
          color="#3333ff"><font color="#000000"><i class="moz-txt-slash"><span
                class="moz-txt-tag"><br>
              </span></i></font></font><br>
        so I add the layers to my build and run the following commands:<br>
        <br>
            $bitbake world<br>
        <br>
        It reported the following error:<br>
        <font color="#3333ff"><br>
              Summary: 15 tasks failed:<br>
           
          /buildarea/raid0/ddu/poky/meta-openembedded/meta-oe/recipes-graphics/takao-fonts/takao-fonts_003.03.01.bb,
          do_fetch<br>
           
          /buildarea/raid0/ddu/poky/meta-openembedded/meta-oe/recipes-support/portaudio/portaudio-v19_svn.bb,
          do_fetch<br>
           
          /buildarea/raid0/ddu/poky/meta-openembedded/meta-oe/recipes-gnome/gtkhtml2/gtkhtml2_svn.bb,
          do_fetch<br>
           
          /buildarea/raid0/ddu/poky/meta-openembedded/meta-python/recipes-devtools/python/python-cryptography-vectors_1.1.bb,
          do_fetch<br>
           
          /buildarea/raid0/ddu/poky/meta-openembedded/meta-networking/recipes-support/geoip/geoip_1.6.6.bb,
          do_fetch<br>
           
          /buildarea/raid0/ddu/poky/meta-openembedded/meta-networking/recipes-connectivity/ufw/ufw_0.33.bb,
          do_fetch<br>
           
          /buildarea/raid0/ddu/poky/meta-openembedded/meta-oe/recipes-support/ctapi-common/ctapi-common_1.1-11.bb,
          do_fetch<br>
           
          /buildarea/raid0/ddu/poky/meta-openembedded/meta-python/recipes-devtools/python/python-decorator_4.0.2.bb,
          do_fetch<br>
           
          /buildarea/raid0/ddu/poky/meta-openembedded/meta-python/recipes-connectivity/python-thrift/python-thrift_0.9.2.bb,
          do_fetch<br>
           
          /buildarea/raid0/ddu/poky/meta-openembedded/meta-python/recipes-devtools/python/python-sqlalchemy_0.7.9.bb,
          do_fetch<br>
           
          /buildarea/raid0/ddu/poky/meta-openembedded/meta-webserver/recipes-httpd/cherokee/cherokee_1.2.98.bb,
          do_fetch<br>
           
          /buildarea/raid0/ddu/poky/meta-openembedded/meta-oe/recipes-graphics/ttf-fonts/ttf-mplus_027.bb,
          do_fetch<br>
           
          virtual:native:/buildarea/raid0/ddu/poky/meta-openembedded/meta-oe/recipes-support/uim/uim_1.8.6.bb,
          do_install<br>
          <font color="#cc0000"> 
            /buildarea/raid0/ddu/poky/meta-openembedded/meta-oe/recipes-connectivity/umip/umip_1.0.bb,
            do_compile</font><br>
           
          /buildarea/raid0/ddu/poky/meta-openembedded/meta-oe/recipes-graphics/x11vnc/x11vnc_0.9.13.bb,
          do_package_qa</font><br>
        <br>
        It seems the umip was affect by the ncurses version change. So I
        went to the log.do_compile, the log file say:<br>
        <br>
        <font color="#cc0000">fatal error: gram.h : No such file or
          directory<br>
          compilation terminated</font><br>
        <br>
        So I think the ncurses version change didn't affect the whole
        system. Then I release the patch "ncurses: drop the compatible
        option"</font></font><font color="#3333ff"><font color="#000000"><i
          class="moz-txt-slash"><span class="moz-txt-tag"><br>
          </span><span class="moz-txt-tag"></span></i>-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------<br>
      </font></font>
    <div class="moz-cite-prefix">On 2016年01月21日 10:17, Khem Raj wrote:<br>
    </div>
    <blockquote
cite="mid:CAMKF1soSrq1P9ztt5_Uavc7e_jtHtR-PNzYH7PERUE3dJfgeaQ@mail.gmail.com"
      type="cite">
      <pre wrap="">On Wed, Jan 20, 2016 at 5:30 PM, Robert Yang <a class="moz-txt-link-rfc2396E" href="mailto:liezhi.yang@windriver.com">&lt;liezhi.yang@windriver.com&gt;</a> wrote:
</pre>
      <blockquote type="cite">
        <pre wrap="">
I had talked with Dengke, we have upgraded ncurses to 6, so remove
--with-abi-version=5 seems better. He had tested world and sdk
build based on oe-core. He will test it based on meta-openembedded.
</pre>
      </blockquote>
      <pre wrap="">
we need to ensure that all apps using ncurses are able to work with
new APIs if we drop compatibility.
may be have ncurses5 package still around to provide backward compatibility.
</pre>
    </blockquote>
    <br>
  </body>
</html>