Date: Thu, 21 Jan 2010 18:49:05 +0100 From: "Julian H. Stacey" <jhs@berklix.com> To: gary.jennejohn@freenet.de Cc: ports@FreeBSD.org, Doug Barton <dougb@FreeBSD.org> Subject: Re: Suggestion: A new variable for a few Makefiles: IS_BINARY Message-ID: <201001211749.o0LHn5gh033380@fire.js.berklix.net> In-Reply-To: Your message "Thu, 21 Jan 2010 17:01:46 %2B0100." <20100121170146.672acbbd@ernst.jennejohn.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Hi, Reference: > From: Gary Jennejohn <gary.jennejohn@freenet.de> > Reply-to: gary.jennejohn@freenet.de > Date: Thu, 21 Jan 2010 17:01:46 +0100 > Message-id: <20100121170146.672acbbd@ernst.jennejohn.org> Gary Jennejohn wrote: > On Thu, 21 Jan 2010 07:34:02 -0800 > Doug Barton <dougb@FreeBSD.org> wrote: > > > On 1/20/2010 6:47 PM, Julian H. Stacey wrote: > > > Some may not don't mind > > > installing binaries from elsewhere, but FreeSBD could protect more, > > > > What is it that you're trying to protect people from? In other words, > > what bad thing do you think is going to happen if someone installs a > > binary, and why do you think that we would allow something dangerous > > into the ports collection in the first place? > > > > I know Julian very well and he's, umm, very cautious. > > I assume he wants a knob he can turn on to avoid installing binaries while > doing unattended installations, i.e. with BATCH set to yes. Thanks, Yes, sort of, Mechanism would be similar to implemenation methoid of BATCH, (Except we could improve on it by not having different strings eg Makefile: IS_INTERACTIVE=YES Mk/*: BATCH= but perhaps Makefile: IS_BINARY=YES Mk/*: SKIP_BINARY_INSTALL=YES # (or NO) ) It could be use with BATCH (but doesnt have to be), eg personally I do a multi pass build on my ports, cd /usr/ports ; make BERKLIX_CLIENTS=YES BATCH=YES # overnight cd /usr/ports ; make BERKLIX_CLIENTS=YES # later build of interactives next day Usage of SKIP_BINARY_INSTALL would be personaly choosable I'd specify SKIP_BINARY_INSTALL=YES (if not default in Mk/*.mk) but others could equally choose Not to assert it. Example: I just did an upgrade (to 8.0) & compiled & installed 694 ports (inc. dependencies), from my set of */Makefile.local Manually scanning Makefiles each remake (inc changing dependencies), looking for binary rogue software sneaking through, is pointless, when binaries without source could so easily be marked in Makefiles. The mechanism would help people & companies that have a policy: Install No software without matching sources. Others could ignore occasional declarative markers in a few ports like www/opera/Makefile IS_BINARY=YES One could debate whether default make install behaviour from Mk/ should be to continue to install (as now), or exit 1, on the few ports needing IS_BINARY=YES; Default would be over-ridable by an env var to make command. There is a spectrum of usage, no one preference fits all, No point trying to convince each other 1 policy is best for all :-) Some run binaries ie BLOBs, drivers, opera, mega ports (openoffice), & flash binaries under linux emulators etc, & maybe havent source, enthusiasm, time, space or interest to build their own. Some (by temperament, employer, or market sector, eg security companies, firewall manufacturers, military development support, government security etc), may want to No software tools installed without matching sources. Not everyone needs the hook, but that's no reason not implement it; it is simple, helps some, & will not harm others who can ignore it. Cheers, Julian -- Julian Stacey: BSD Unix Linux C Sys Eng Consultants Munich http://berklix.com Mail plain text not quoted-printable, HTML or Base64 http://www.asciiribbon.org
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201001211749.o0LHn5gh033380>