Date: Thu, 31 Jul 2014 09:08:07 +0000 From: Alexey Dokuchaev <danfe@FreeBSD.org> To: marino@freebsd.org Cc: svn-ports-head@freebsd.org, svn-ports-all@freebsd.org, Roman Bogorodskiy <novel@FreeBSD.org>, ports-committers@freebsd.org Subject: Re: svn commit: r363524 - in head/emulators/pearpc: . files Message-ID: <20140731090807.GA81522@FreeBSD.org> In-Reply-To: <53DA0037.1060809@marino.st> References: <201407302302.s6UN2GCn094189@svn.freebsd.org> <20140731060809.GA20983@FreeBSD.org> <53D9E30D.3030902@marino.st> <20140731065333.GA39915@FreeBSD.org> <53DA0037.1060809@marino.st>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, Jul 31, 2014 at 10:37:11AM +0200, John Marino wrote: > Well, sure, everyone would agree that thorough run-time testing would be > ideal but it's simply unrealistic due to: > > 1) The time required > 2) more important, the expertise required. The vast majority of the 25K > ports in the tree, I've never used. I have no idea how to do basic Much of this burden has to be (and usually is, for real maintainers) on submitter's good conscience (which, of course, does not mean committers should not review and check every submission). It becomes worse when someone sees a port she wants to update just because there is new version available, without questioning if it's actually any good to update, and not being actual software user. Mass updaters like Kato's sweatshop slaves fall in this category, I presume. > runtime testing on them much less a thorough review (All I could give is > to see if it starts up) Start-up test probably makes sane minimal requirement for a commit (esp. if ported software does not have test suit with it). But it's often not hard to see if it can do something simple yet useful (e.g. build and run "hello world" program in case of languages, read/write a file for a text processing application, connect to a network server or fetch a page from the net for a client). > In this case, to run-test this is even more obnoxious, you need to > download an image from apple, presumably a large image. I have a 20G > data limit *per month* (unlimited at night) because I have > satellite-based internet. I can't afford data use like this. Well of course in this case you're fine to skip this step. What's not as good, however, is that submitter probably did not do this test as well. :( > And --- we have have a huge backlog on the PRs. the time is better > spent getting through PRs 3, 6 even 12 months old rather than burning it > all on one 10-year unmaintained port. Let's use what limited volunteer > time we have wisely, and let users submit new PRs if there's a run issue > (which indicates the port is actually used as well). Fair enough. I just hope that in strive for closed PR numbers we won't lower the quality of what we commit (as opposed to what's usually being submitted). ./danfe
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20140731090807.GA81522>