Date: Wed, 29 Aug 2012 22:39:44 -1000 From: Doug Barton <dougb@FreeBSD.org> To: Stanislav Sedov <stas@FreeBSD.org> Cc: Steve Wills <swills@FreeBSD.org>, freebsd-rc@FreeBSD.org, ruby@FreeBSD.org Subject: Re: procname when ruby is used Message-ID: <503F26D0.1050109@FreeBSD.org> In-Reply-To: <20120829203614.5f51db26061ea094f122379f@FreeBSD.org> References: <503E6D62.3000101@FreeBSD.org> <503EC42B.6000302@FreeBSD.org> <503ED8C5.2010203@FreeBSD.org> <20120829203614.5f51db26061ea094f122379f@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On 08/29/2012 05:36 PM, Stanislav Sedov wrote: > On Wed, 29 Aug 2012 23:06:45 -0400 > Steve Wills <swills@FreeBSD.org> mentioned: > >> On 08/29/12 21:38, Doug Barton wrote: >>> >>> I'm pretty sure you actually want to use command_interpreter instead of >>> procname. It should actually be very rare to use procname directly in an >>> rc.d script. >> >> Got it, although that means picking the value at build time, but that >> seems OK. >> > > We actually already have a practice of doing that with RUBY_SHEBANG, so it > seems reasonable. GMTA. :) I'm sensitive to the issue of this being build time reliant which means that if the user upgrades their ruby version the rc.d script could become outdated. I think that this could be ameliorated by forcing the shebang line to be just /usr/local/bin/ruby, but I'm not sure how ruby handles that. One way we could improve the situation would be to support a glob pattern for command_interpreter. Haven't thought through the implications of that though. Doug
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?503F26D0.1050109>