Date: Sun, 2 Jun 2002 23:26:39 -0700 From: Jordan K Hubbard <jkh@queasyweasel.com> To: The Anarcat <anarcat@anarcat.ath.cx> Cc: libh@FreeBSD.ORG, Alexander Langer <alex@big.endian.de> Subject: Re: SYSTEM package contradictions Message-ID: <DCDDB676-76BA-11D6-BE49-0003938C7B7E@queasyweasel.com> In-Reply-To: <20020603020043.GF288@lenny.anarcat.ath.cx>
next in thread | previous in thread | raw e-mail | index | archive | help
The package documentation describes the correct behavior. The implementation only describes what got implemented, and not even correctly. :) On Sunday, June 2, 2002, at 07:00PM, The Anarcat wrote: > The SYSTEM package documentation is contradictory to its > implementation. The doc in sysinstall pretends that the > > "SYSTEM package provides the following features: > OSNAME (FreeBSD for us) > OSVERSION: `uname -s` version `uname -r` > PLATFORM (i386, alpha, sparc64, IA64, ...) > CPU (386, 486, Pentium, Athlon, ...; MMX, 3DNow, SSE, ...) > libh version (comparable to /var/db/port.mkversion) > " > > The current implementation gives me: > > Features: > i386 3.0 386 0 > FreeBSD 4.5-STABLE 0 0 > > on my machine. A few questions: > > 1- shouldn't the feature names be (eg) "OSNAME" instead of "FreeBSD"? > In the case of the "CPU" feature, the feature names could become > arbitrarly esoteric and obstruct the feature namespace. > > I suggest using (eg, again) "OSNAME" instead. > > 2- there are features missing: the OSVERSION, CPU and libh version. > > I could probably implement dummy feature entries in the DB along with > #warnings. > > [cf. Database.cc:733 > features->push_back( new Feature( architecture_feature, > architecture_feature_version, architecture_feature_serial, 0 ) ); > features->push_back( new Feature( buf.sysname, buf.release, 0, 0 > ) ); > ] > > 3- I don't understand what use is OSVERSION. From the doc, my > OSVERSION string would be something like: > FreeBSD lenny.anarcat.ath.cx 4.5-STABLE FreeBSD 4.5-STABLE #6: Sun May > 26 18:16:45 EDT 2002 > anarcat@lenny.anarcat.ath.cx:/usr/obj/usr/src/sys/LENNY i386 version > 4.5-STABLE > > I don't think it's necessary. > > 4- what is the format of the "standard revision" feature element? x.y? > x.y,z? And in any case, how should be interpreted the version of the > FreeBSD (or OSNAME) feature? > > 5- libh version? That's 0.2.2 right now? Should this be harmonized > with the zip archive attributes? > > The way I see it, the SYSTEM package should provide the features: > > Name Version Serial Snapshot > OSVERSION 4.5 [1] [1] > OSNAME FreeBSD 0 0 > PLATFORM i386 0 0 > CPU i586 0 0 > LIBH_VERS 0.2.2 0 200206022150 [2] > > [1] there's been discussions on how to handle versionning of the > FreeBSD source tree, but this is a tricky situation. I guess that with > tree-based releases, only the snapshot date could be used. I'm not > sure how to use those 2 columns. > [2] this is not really formatted this way, but gives an idea > > I'm not sure this is appropriate since "Version" maybe shouldn't > support arbitrary strings... > > A. > > -- > Imagination is more important than knowledge > - Albert Einstein > <mime-attachment> -- Jordan K. Hubbard Engineering Manager, BSD technology group Apple Computer To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-libh" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?DCDDB676-76BA-11D6-BE49-0003938C7B7E>