<html>
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
  </head>
  <body>
    <div class="moz-cite-prefix">Le 29/09/2021 à 13:41, Sveinar Søpler a
      écrit :<br>
    </div>
    <blockquote type="cite"
      cite="mid:520524313.280.1632915700043.JavaMail.zimbra@dexter.no">
      <meta http-equiv="content-type" content="text/html; charset=UTF-8">
      <div style="font-family: arial, helvetica, sans-serif; font-size:
        12pt; color: #000000">
        <div>The logfile from the WineHQ Ubuntu 21.04 build on OBS is
          here:<br>
        </div>
        <div><a
href="https://build.opensuse.org/build/Emulators:Wine:Debian/xUbuntu_21.04/x86_64/wine-devel/_log"
            moz-do-not-send="true" class="moz-txt-link-freetext">https://build.opensuse.org/build/Emulators:Wine:Debian/xUbuntu_21.04/x86_64/wine-devel/_log</a><br
            data-mce-bogus="1">
        </div>
        <div><br data-mce-bogus="1">
        </div>
        <div>I could not immediately find warnings about what you are
          talking about here, but there sure is enough warnings about
          "format argument" (-Wformat and -Wstrict-aliasing) stuff, that
          i don't get when i compile wine myself locally. Might not be
          related to what you ask here, but since you were talking about
          "gcc 11 becomes mainstream", i thought it was worth to point
          to the official OBS buildlog for Ubuntu 21.04 - that uses
          gcc-11 (although gcc-mingw-w64 is still at 10.3).<br
            data-mce-bogus="1">
        </div>
        <div><br data-mce-bogus="1">
        </div>
        <div>I guess debian/ubuntu "debuild" system sets these -Wformat=
          and -Wstrict-aliasing options when building as default, and
          that seem to throw a lot of warnings. I did not browse through
          the entire log, but maybe it is just some overly eager checks
          that some distro's enable by default, that it is not really
          critical?<br data-mce-bogus="1">
        </div>
        <div><br data-mce-bogus="1">
        </div>
        <div>Do you have a "sample warning" for me to search for? Eg.
          <!--StartFragment-->
          <pre>dlls/ncrypt/main.c:106:11: warning: format '%lx' expects argument of type 'long unsigned int', but argument 5 has type 'NCRYPT_PROV_HANDLE' {aka 'long long unsigned int'} [-Wformat=]</pre>
          <!--EndFragment--> </div>
      </div>
    </blockquote>
    <p>to correct myself:</p>
    <p>- I thought the trigger was "simply" the upgrade to GCC version
      11... after some quick search, Fedora claims moving from 11.0
      (F34) to 11.2 (F35), but looks like F34 is pushing gcc 11.2 in
      updates since July... so the root cause of those new warnings need
      to be clarified</p>
    <p>- the warnings I'm seeing are triggered by
      -Wmisleading-indentation, -Warray-bounds, -Wsizeof-array-div,
      -Wmaybe-uninitialized, which are all active when -Wall is used</p>
    <p>it's strange from the log link above that non of the options work
      with the cross compiler</p>
    <p>[  569s] checking whether the cross-compiler supports -target
      x86_64-w64-mingw32 -fuse-ld=lld... no<br>
      [  569s] checking whether the cross-compiler supports
      -fno-strict-aliasing... no<br>
      [  569s] checking whether the cross-compiler supports
      -Werror=unknown-warning-option... no<br>
      [  569s] checking whether the cross-compiler supports
      -Werror=ignored-optimization-argument... no<br>
      [  569s] checking whether the cross-compiler supports
      -Wdeclaration-after-statement... no<br>
      [  570s] checking whether the cross-compiler supports
      -Wempty-body... no<br>
      [  570s] checking whether the cross-compiler supports
      -Wignored-qualifiers... no<br>
      [  570s] checking whether the cross-compiler supports
      -Winit-self... no<br>
      [  570s] checking whether the cross-compiler supports
      -Wpacked-not-aligned... no<br>
      [  570s] checking whether the cross-compiler supports
      -Wpragma-pack... no<br>
      [  570s] checking whether the cross-compiler supports
      -Wshift-overflow=2... no<br>
      [  570s] checking whether the cross-compiler supports
      -Wstrict-prototypes... no<br>
      [  570s] checking whether the cross-compiler supports
      -Wtype-limits... no<br>
      [  570s] checking whether the cross-compiler supports
      -Wunused-but-set-parameter... no<br>
      [  570s] checking whether the cross-compiler supports -Wvla... no<br>
      [  570s] checking whether the cross-compiler supports
      -Wwrite-strings... no<br>
      [  570s] checking whether the cross-compiler supports
      -Wpointer-arith... no<br>
      [  570s] checking whether the cross-compiler supports
      -Wlogical-op... no<br>
      [  570s] checking whether the cross-compiler supports
      -Wabsolute-value... no<br>
      [  570s] checking whether the cross-compiler supports
      -Wno-format... no<br>
      [  570s] checking whether the cross-compiler supports
      -Wformat-overflow... no<br>
      [  570s] checking whether the cross-compiler supports -Wnonnull...
      no<br>
      [  570s] checking whether the cross-compiler supports -mcx16... no<br>
      [  570s] checking whether the cross-compiler supports -gdwarf-2...
      no<br>
      [  571s] checking whether the cross-compiler supports
      -gstrict-dwarf... no<br>
      [  571s] checking whether the cross-compiler supports
      -fexcess-precision=standard... no<br>
      I'd expect that a couple of them are supported<br>
    </p>
  </body>
</html>