Re: libgo patch committed: Change build procedure to use build tags

[Ian Lance Taylor <iant at golang dot org>]

On Fri, Aug 12, 2016 at 6:56 AM, Rainer Orth
<ro@cebitec.uni-bielefeld.de> wrote:
> Hi Ian,
>
>>>> this patch broke i386-pc-solaris2.12 and sparc-sun-solaris2.12
>>>> bootstrap, however: in both cases, the 64-bit build of os.lo fails like this:
>>>>
>>>> /vol/gcc/src/hg/trunk/local/libgo/go/os/dir.go:82:8: error: reference to undefined name 'libc_readdir_r'
>>>>    i := libc_readdir_r(file.dirinfo.dir, entryDirent, pr)
>>>>         ^
>>>>
>>>> Neither dir_largefile.go (which is correctly omitted, being 32-bit only)
>>>> nor dir_regfile.go (which is needed here) is included in the
>>>> compilation.
>>>
>>> Sorry, I don't know what I messed up in my testing.  I committed the
>>> appended patch, which should fix the problem.
>>
>> I had found a different one to the same effect, but running match.sh
>> with yours looked right, too.
>
> here's another issue found during wider testing: on Solaris 10 with
> /bin/ksh as CONFIG_SHELL, the build aborts like this:
>
> objcopy: 'bufio.o': No such file
> Makefile:5072: recipe for target 'bufio.gox' failed
> make[4]: *** [bufio.gox] Error 1
>
> It turns out that no bufio.o is created indeed:
>
> /vol/gcc/src/hg/trunk/local/libgo/../install-sh -c -d .; files=`echo  | sed -e 's/[^ ]*\.gox//g' -e 's/[^ ]*\.dep//'`; /bin/ksh ./libtool --tag GO --mode=compile /var/gcc/regression/trunk/10-gcc/build/./gcc/gccgo -B/var/gcc/regression/trunk/10-gcc/build/./gcc/ -B/vol/gcc/i386-pc-solaris2.10/bin/ -B/vol/gcc/i386-pc-solaris2.10/lib/ -isystem /vol/gcc/i386-pc-solaris2.10/include -isystem /vol/gcc/i386-pc-solaris2.10/sys-include     -minline-all-stringops  -g -O2 -I . -c -fgo-pkgpath=`echo bufio.lo | sed -e 's/.lo$//' -e 's/-go$//'` -o bufio.lo $files
>
> The list of input files, output from the likes of
>
> /bin/ksh /vol/gcc/src/hg/trunk/local/libgo/match.sh --goarch=386 --goos=solaris --srcdir=/vol/gcc/src/hg/trunk/local/libgo/go/bufio --extrafiles=""
>
> is empty.  Running match.sh with /bin/ksh -x reveals
>
> + set -e
> + unset LANG
>
> where the script aborts.  It turns out that /bin/ksh errors out on an
> unset of a variable that isn't set.
>
> Maybe one could explicitly set LANG=C here, or use the `portable unset'
> used in configure?

That is a shell portability problem I was not aware of.  I took your
LANG=C suggestion.  Bootstrapped and ran Go testsuite on
x86_64-pc-linux-gnu.  Committed to mainline.

Ian

Attachment: patch.txt
Description: Text document