<div dir="ltr"><div>I've been successful in getting a fully functional build on FreeBSD/PowerPC, including menu icons, toolbars and graphics.<br><br></div>The issue, as I discovered, was in the code used to define endianness, for FreeBSD/PowerPC. This also resolved the issue with the Base application. <br><br>I have developed patches and would like to contribute this work for inclusion in the code base.<br></div><div class="gmail_extra"><br><div class="gmail_quote">On Mon, Dec 5, 2016 at 1:58 PM, Curtis Hamilton <span dir="ltr"><<a href="mailto:clhamilto@gmail.com" target="_blank">clhamilto@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr"><div>I've successfully built Libreoffice-4.3.7 on FreeBSD 10.2/PowerPC64, that has menu icons, toolbars and graphics. With the exception of the Base application, everything seems to be working correctly.</div><div><br></div><div>To get it to build, I had to take the "hatchet" approach, due to a flaw in the big/little endian logic in the code. The code does not seem to be defining "OSL_BIGENDIAN" properly, thus testing fails and defaults to little endian. I removed and/or commented the code tests for "OSL_BIGENDIAN" in most modules where it appeared. Leaving the software hard coded for "big endian". I believe the issue with the Base application is due to the "hatchet" that was taken to the code.</div><div><br></div><div>Armed with this knowledge, I working to find and correct the flawed endian logic and develop patches. I'm sure the results will be easily repeatable for LO5.<br><br></div></div><div class="HOEnZb"><div class="h5"><div class="gmail_extra"><br><div class="gmail_quote">On Mon, Nov 21, 2016 at 7:32 AM, Michael Stahl <span dir="ltr"><<a href="mailto:mstahl@redhat.com" target="_blank">mstahl@redhat.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><span>On 19.11.2016 00:19, Curtis Hamilton wrote:<br>
><br>
><br>
> On 11/18/2016 05:26, Michael Stahl wrote:<br>
>> On 17.11.2016 17:54, Curtis Hamilton wrote:<br>
<br>
</span><span>>> My build environment:<br>
>><br>
>> OS: FreeBSD 10.2-STABLE powerpc (64-bit)<br>
>> Compiler: GCC version 4.8.5<br>
>> Java Version: OpenJDK Runtime Environment (build 1.8.0_102-b14)<br>
>> OpenJDK 64-Bit Server VM (build 25.102-b14, mixed mode)<br>
>><br>
>> I’m hoping that someone with knowledge of the workings of the user<br>
>> interface can provide some insight on what to look at. I’m not sure if<br>
>> this is a LibreOffice issue or FreeBSD on PowerPC issue.<br>
>> it could be either, in particular if it's big-endian PPC there might be<br>
>> an endian issue in some image handling code, of which there is quite a<br>
>> lot for various different formats.<br>
>><br>
> that was my initial assumption, but could not figure out how<br>
> OSL_BIGENDIAN was being defined, since "osl/endian.h" is not being<br>
> included in all files that tests for OSL_BIGENDIAN.<br>
<br>
</span>urgh, really? ... yes that is clearly the case ... i hope it should be<br>
fixed with these commits<br>
<br>
<a href="https://gerrit.libreoffice.org/#/c/31039/" rel="noreferrer" target="_blank">https://gerrit.libreoffice.org<wbr>/#/c/31039/</a><br>
<a href="https://gerrit.libreoffice.org/#/c/31040/" rel="noreferrer" target="_blank">https://gerrit.libreoffice.org<wbr>/#/c/31040/</a><br>
<br>
<br>
</blockquote></div><br></div>
</div></div></blockquote></div><br></div>