Date: Mon, 13 Feb 2006 09:47:34 +0100 From: Pav Lucistnik <pav@FreeBSD.org> To: Kris Kennaway <kris@obsecurity.org> Cc: Hajimu UMEMOTO <ume@FreeBSD.org>, nobutaka@FreeBSD.org, Kris Kennaway <kris@FreeBSD.org>, cvs-all@FreeBSD.org, ports-committers@FreeBSD.org, ports@FreeBSD.org, cvs-ports@FreeBSD.org Subject: Re: cvs commit: ports/japanese/migemo Makefile Message-ID: <1139820454.86374.9.camel@pav.hide.vol.cz> In-Reply-To: <20060213035409.GA9141@xor.obsecurity.org> References: <200602122224.k1CMOEAY019403@repoman.freebsd.org> <ygeu0b3ly44.wl%ume@mahoroba.org> <20060213035409.GA9141@xor.obsecurity.org>
next in thread | previous in thread | raw e-mail | index | archive | help
--=-ITxszJTxWkOx6vs6wwfE
Content-Type: text/plain; charset=ISO8859-2
Content-Transfer-Encoding: quoted-printable
Kris Kennaway p=ED=B9e v ne 12. 02. 2006 v 22:54 -0500:
> On Mon, Feb 13, 2006 at 12:46:19PM +0900, Hajimu UMEMOTO wrote:
> > Hi,
> >=20
> > >>>>> On Sun, 12 Feb 2006 22:24:14 +0000 (UTC)
> > >>>>> Kris Kennaway <kris@FreeBSD.org> said:
> >=20
> > kris> kris 2006-02-12 22:24:14 UTC
> >=20
> > kris> FreeBSD ports repository
> >=20
> > kris> Modified files:
> > kris> japanese/migemo Makefile=20
> > kris> Log:
> > kris> BROKEN: Coredump during build
> >=20
> > I examined this issue bit, then found that it occurs because of
> > shortage of an initial thread's stack. The japanese/migemo still can
> > be built fine with non-pthread Ruby.
> > 2MB which is the recent value of THR_STACK32_INITIAL is still short
> > for building japanese/migemo. I doubled THR_STACK32_INITIAL, then
> > japanese/migemo could be built fine with pthread Ruby.
> > So, I think it is hard to fix this breakage by japanese/migemo port.
> > It seems there are not a few problems with pthread Ruby including this
> > issue. So, it may better to disable pthread support of Ruby by
> > default at least until our next release is out, IMHO.
>=20
> Thanks for investigating, you should talk to pav who did the ruby
> upgrade.
Bumping default thread stack size would be a good solution, there were a
lot of reports about that value being absurdly small on FreeBSD compared
to other systems. Can we perhaps toss some pthread_attr_setstacksize()
into the ruby port?
See /usr/ports/devel/glib20/files/extra-patch-gthread_gthread-posix.c
for inspiration.
--=20
Pav Lucistnik <pav@oook.cz>
<pav@FreeBSD.org>
Sun couldn't care less about Linux other than it now being necessary
in order to be buzzword compliant.
-- Al Dente
--=-ITxszJTxWkOx6vs6wwfE
Content-Type: application/pgp-signature; name=signature.asc
Content-Description: Toto je =?iso-8859-2?Q?digit=E1ln=EC?=
=?ISO-8859-1?Q?_podepsan=E1?= =?iso-8859-2?Q?_=E8=E1st?=
=?ISO-8859-1?Q?_zpr=E1vy?=
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (FreeBSD)
iD8DBQBD8EemntdYP8FOsoIRAp0fAJ9JgjYXtjsT/Przm4Hs3yYXuKk2FQCgnk0P
E2+tykdk4iLJx4JZclS7o/M=
=Obpb
-----END PGP SIGNATURE-----
--=-ITxszJTxWkOx6vs6wwfE--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1139820454.86374.9.camel>