On Nov 26, 2011, at 15:44 , Ryan Davis wrote:

>=20
> On Nov 26, 2011, at 06:30 , Andrew Grimm wrote:
>=20
>> I tried compiling the trunk version on OS X Snow Leopard, and it
>> didn't work, and it didn't give a meaningful error message:
>>=20
>> https://gist.github.com/1395715 has the attempt, with the failure
>> during the ./configure step. At the end, there is also uname
>> information, and the version of ruby-head used.
>>=20
>> I had previously tried compiling it under rvm, and it didn't work.
>> User mpapis of the #rvm channel also tried compiling it on OS X using
>> rvm, and couldn't get it to work either. He didn't try compiling
>> directly (ie not using rvm), however. mpapis also tried compiling on
>> Linux, and it worked there (see the gist for OS details).
>>=20
>> I'm hesitant in claiming that this is a bug, because I've never tried
>> compiling Ruby without RVM before. Can someone who has experience =
with
>> directly compiling ruby on OS X check whether it's working for them?
>=20
> I've currently got r33782 (2011-11-17) checked out and it configures =
on osx just fine.
>=20
> If you can build that revision and not the absolute latest, then it is =
definitely a bug.
>=20
> I think it is a bug regardless because it looks like a syntax error in =
configure.in.

Yeah. It's a bug. I'll report it.