On Wed, Nov 28, 2018 at 03:56:29PM +0100, Ævar Arnfjörð Bjarmason wrote:
>
> On Thu, Nov 22 2018, Jeff King wrote:
>
> > On Thu, Nov 22, 2018 at 02:17:01AM -0800, Carlo Arenas wrote:
> >> PS. upstreaming the PERL_PATH fix is likely to be good to do soonish
> >> as I presume at least all BSD might be affected, let me know if you
> >> would rather me do that instead as I suspect we might be deadlocked
> >> otherwise ;)
> >
> > Yeah, the $PERL_PATH thing is totally orthogonal, and should graduate
> > separately.
>
> On the subject of orthagonal things: This test fails on AIX with /bin/sh
> (but not /bin/bash) due to some interaction of ssize_b100dots and the
> build_option function. On that system:
>
> $ ./git version --build-options
> git version 2.20.0-rc1
> cpu: 00FA74164C00
> no commit associated with this build
> sizeof-long: 4
> sizeof-size_t: 4
>
> But it somehow ends up in the 'die' condition in that case statement. I
> dug around briefly but couldn't find the cause, probably some limitation
> in the shell constructs it supports. Just leaving a note about this...
That's weird. The functions involved are pretty vanilla. I'd suspect
something funny with the sed invocation:
build_option () {
git version --build-options |
sed -ne "s/^$1: //p"
}
but that's the one thing that shouldn't be dependent on the shell in
use.
Can you manually replicate the shell commands to see where it goes
wrong?
-Peff
