Date: Sat, 17 Sep 2005 15:13:32 -0700 From: Colin Percival <cperciva@freebsd.org> To: Ceri Davies <ceri@submonkey.net> Cc: cvs-ports@FreeBSD.org, cvs-all@FreeBSD.org, ports-committers@FreeBSD.org Subject: Re: cvs commit: ports/sysutils/portsnap Makefile distinfo Message-ID: <432C950C.5050301@freebsd.org> In-Reply-To: <20050917220320.GA36715@submonkey.net> References: <200509171609.j8HG9tS9035379@repoman.freebsd.org> <20050917220320.GA36715@submonkey.net>
next in thread | previous in thread | raw e-mail | index | archive | help
Ceri Davies wrote: > On Sat, Sep 17, 2005 at 04:09:55PM +0000, Colin Percival wrote: >> Update to version 0.9.5. This is identical to 0.9.4 except for a fix >> which makes make_index not coredump when it encounters a circular >> dependency in the ports tree. > > That's wonderful, except it leaves me with a zero-length INDEX. The > core dump got me halfway through the audio/ category at least, and was > actually better in that respect. > > Any fix in the works? A few ports people are looking at this and trying to decide how best to fix the port in question. As a more long-term fix, I'm going to add extra checks on my portsnap build server to stop new INDEX files from being distributed until they are correct. Right now people get an INDEX built from the most recent tree on which "make describes" succeeded; once the current breakage ends, I'm going to change that to giving people an INDEX from the most recent tree where "make describes" succeeded _and_ produced a file which portsnap's make_index manages to turn into an INDEX without any errors. Colin Percival
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?432C950C.5050301>