Re: linux-ppc harfbuzz GUB build error

[Hans Aikema]

> On 14 Dec 2016, at 14:56, Hans Aikema <address@hidden> wrote:
> 
> 
>> On 14 Dec 2016, at 14:30, Masamichi Hosoda <address@hidden> wrote:
>> 
>>>>> New build from Masamichi’s branch passed the linux ppc harfbuzz, so the 
>>>>> ICU patch worked. Now looking into a breaking freebsd-x86 lilypond error: 
>>>> 
>>>> Would you show me the whole log file?
>>> Went into a bit of experimentation afterwards, still failed attempts only, 
>>> after discovering that the message is basically a side-effect of 
>>> set -u
>>> combined with autoconf apparently trying to use unbound variable 
>>> CONFIG_SHELL
>>> 
>>> On a cleanly booted docker-image recreated the issue with 
>>> bin/gub freebsd-x86::lilypond 
>> 
>> In my environments, autoconf does not raise such error.
>> Do you set the "set -u" or "set -o nounset" somewhere?
>> If so, would you remove the setting?
> The set -u (or to be complete set -ux) I discovered inside
> 
> /home/gub/gub/target/freebsd-x86/src/lilypond-git.sv.gnu.org--lilypond.git-master/smart-configure.sh
> 
> which was the command that was called when the error was signalled
> 
> 
> 
> _______________________________________________
> lilypond-devel mailing list
> address@hidden
> https://lists.gnu.org/mailman/listinfo/lilypond-devel

My suspicion was that Ubuntu and CentOS shells treat  'set -u’ different when 
used in an
if test "x$CONFIG_SHELL" = x; then 
construct 

so as a varification of that assumption on the prompt I did:
set -u
if test "x$CONFIG_SHELL" = x; then echo "Configshell not set"; else echo 
"Configshell set"; fi

for both a CentOS 7 and an Ubuntu Xenial container.
Unfortunately both yield a 
bash: CONFIG_SHELL: unbound variable

will do some more digging inside autoconf to see if I can spot any reason why 
CONFIG_SHELL would not be set in my case