Date: Fri, 26 Oct 2001 16:23:09 -0700 From: Bill Fenner <fenner@research.att.com> To: jhay@icomtek.csir.co.za Cc: sobomax@freebsd.org, jhb@freebsd.org, cvs-all@freebsd.org, cvs-committers@freebsd.org, obrien@freebsd.org Subject: Re: cvs commit: ports/devel/automake Makefile distinfo pkg-plist Message-ID: <200110262323.QAA14607@windsor.research.att.com>
next in thread | raw e-mail | index | archive | help
>How much of this "breakage" is truth and how much FUD? I ran "make configure" on all 220 ports with USE_AUTOCONF or USE_AUTOMAKE using autoconf 2.52, 2.13-000227 (today's autoconf port) and 2.13 (last week's autoconf port). I used NO_DEPENDS, so the data is somewhat noisy, however the baseline with 2.13 helps to put an upper bound on the noise. The raw data: 2.52 failure rate: 81 + 45 / 220 2.13-000227 failure rate: 14 + 74 / 220 2.13 baseline failure rate: 14 + 74 / 220 These are "autoconf failures" + "./configure failures" / "total ports". At best, there are 38 ports (81+45 - (14+74)), 17% that fail with autoconf 2.52 . At worst, there are 67 (81 - 14), 30%. I think the docs overstate 2.52's backwards compatability. A port-by-port list is at http://people.freebsd.org/~fenner/autoconf.txt . The output of the 2.52 and 2.13 runs are http://people.freebsd.org/~fenner/autoconfcheck-2.{52,13}.txt . The output of the 2.13-000227 run is different from the 2.13 run in unimportant ways, but I guess I could publish that too if anyone cared. Bill To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200110262323.QAA14607>