Re: [Mingw-cross-env-list] [patch] fix for building Qt on Mac with "-arch windows"

[Tony Theodore]

On 21 May 2010 23:16, Mark Brand <address@hidden> wrote:
>
>>> It's possible that the solution will expose new issues, so please try it
>>> before applying! :)
>>
>> With those changes, CFG_MAC_ARCHS remains unset, so the build fails
>> later on with some byte order/endian error messages (log attached).
>>
>> I'll try to have another look at it tonight, I'm curious to see what
>> "-arch windows" does on other systems. If it gets silently ignored, we
>> may just be able to remove the configure check for MAC and let it be
>> passed along.
>
> Thanks for trying that. I've looked at your log and it looks like you tried
> out the Qt 4.7.0-beta1 version. We seem to be making progress.
>
> The block at line 5586 in the configure script is where the logic is wrong
> for PLATFORM_MAC:
>
> if [ "$CFG_ENDIAN" = "auto" ]; then
>    if [ "$PLATFORM_MAC" = "yes" ]; then
>        true #leave as auto
>    else
>        "$unixtests/endian.test" "$XQMAKESPEC" $OPT_VERBOSE "$relpath"
> "$outpath"
>        F="$?"
>        if [ "$F" -eq 0 ]; then
>            CFG_ENDIAN="Q_LITTLE_ENDIAN"
>        elif [ "$F" -eq 1 ]; then
>            CFG_ENDIAN="Q_BIG_ENDIAN"
>        else
>            echo
>            echo "The target system byte order could not be detected!"
>            echo "Turn on verbose messaging (-v) to see the final report."
>            echo "You can use the -little-endian or -big-endian switch to"
>            echo "$0 to continue."
>            exit 101
>        fi
>    fi
> fi
>
> The workaround is probably to add "-little-endian" to the configure options?
> Would you mind trying that? If it works, I'll patch the configure script
> too.

Adding -little-endian does indeed work and allows the build to
complete and the test program to run.

Cheers,

Tony