Date: Wed, 13 Nov 2002 20:01:57 -0500 From: The Anarcat <anarcat@anarcat.ath.cx> To: Jordan K Hubbard <jkh@queasyweasel.com> Cc: Alexander Langer <alex@big.endian.de>, libh@freebsd.org Subject: Re: Problem confirmed (?) and death to lib[h]disk (!) (Re: serious libh linking problems) Message-ID: <20021114010157.GA307@lenny.anarcat.ath.cx> In-Reply-To: <53135641-F76A-11D6-9957-000393BB9222@queasyweasel.com> References: <20021113234308.GJ9829@xtanbul.studio.espresso-com.com> <53135641-F76A-11D6-9957-000393BB9222@queasyweasel.com>
next in thread | previous in thread | raw e-mail | index | archive | help
--k+w/mQv8wyuph6w0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable Much clearer, thank you. I was confused over the use of my name in the last sentence. ;) That said, I totally agree. Getting basic package stuff working is #1 priority, and I even consider the UI stuff secondary, at the point we're at now (that is, no tvision backend, and a barely working Qt). But we're getting there. Once I get that linking problem solved, I think a lot of problems will also be easier to solve, mainly because I won't have to recompile tclh.static (and dependant modules) each time I make a change. :) A. On Wed Nov 13, 2002 at 04:45:08PM -0800, Jordan K Hubbard wrote: > I'll restate my last sentence in different terms: Unless the libh=20 > project produces the kind of results that people can see, touch and=20 > actually use to install FreeBSD and some set of additional packages,=20 > you won't get the kind of help you need to do all of the additional=20 > rock-polishing work required, some of which could very well be=20 > "enhanced scriptability" and any number of other features that doubling= =20 > or even tripling the number of bodies available to do the work could=20 > bring. But you're never get there if you don't first release a working= =20 > prototype, hence doing that is essentially more important than ANYTHING= =20 > else of architectural or academic value. >=20 > Does that make it clearer? :-) >=20 > - Jordan >=20 > On Wednesday, November 13, 2002, at 03:43 PM, The Anarcat wrote: >=20 > >Exactly. However, scripting is not necessarly for 'the average user', > >but more for the power user, actually. > > > >One might wish to script the disk editor to automagically create > >fixed-sized partitions or something... > > > >That said, I'm really confused by your last sentence, could you > >make it clearer? :) > > > >Antoine <-- yes, antoine > > > >On Wed Nov 13, 2002 at 03:34:06PM -0800, Jordan K Hubbard wrote: > >>I think you may have taken this as a wider mandate for "what's > >>scriptable" than I intended. I also see no reason why the disk editor > >>would be changed by an average user, though I'm sure both of you would > >>also agree that being able to localize it is pretty important. :-) > >>That said, getting out a working prototype should probably be given a > >>higher degree of importance than anything else for all the reasons=20 > >>that > >>Antoine states. > >> > >>- Jordan > >> > >>On Wednesday, November 13, 2002, at 02:09 PM, The Anarcat wrote: > >> > >>>On Wed Nov 13, 2002 at 01:41:01PM -0800, Jordan K Hubbard wrote: > >>>>I think that perhaps the "core" of sysinstall can be compiled but > >>>>everything to do with the user interface, the details of which > >>>>distributions are selected, and so on - just about everything that's > >>>>"policy level" should be scripted. Why? Because it will make=20 > >>>>things > >>>>100X easier for the universities and large ISPs and whatnot of the > >>>>world to completely change syinstall's behavior to fit their own > >>>>unique > >>>>needs, say with different default package sets, menus and UIs in > >>>>different languages or different layouts, you name it. I would only > >>>>expect those parts of sysinstall which are so "core" and essential=20 > >>>>and > >>>>nature that nobody would ever want to customize them to be compiled. > >>> > >>>That is all well and nice in words, but I think there are more > >>>pressing matter for now. > >>> > >>>Of course everthing *can* be scripted. But why script the disk=20 > >>>editor? > >>>Or if we script it, why would it even be part of libh's core? > >>> > >>>I think a disk editor is outside libh's scope. It can be pretty easy, > >>>once we get dynamic linking back online, to make a script load a > >>>(third party?) disk library and script from there. But the disk=20 > >>>library > >>>is too much for libh for handle, I think, especially with the GEOM > >>>changes. > >>> > >>>So, yes, I agree that libh must provide a UI-indendant scripting > >>>language but it doesn't mean it must provide every damn feature > >>>scripts might need. > >>> > >>>2 things: > >>> > >>>- UI library > >>>- package system > >>> > >>>rest is third party loadable modules. heck, if we can't make it third > >>>party, how can we possibly pretend to extend libh in any way?? > >>> > >>>That's what I'm willing to maintain. If anything else breaks, I think > >>>it shouldn't hinder libh development, which is hard enough as it is > >>>now. > >>> > >>>Sorry for the ranting, but things are getting pretty hard now. I've > >>>been struggling for a pretty good while and now that we're almost > >>>getting to have a semi-working package system, I'm stopped by yet > >>>another thing. It's really annoying. > >>> > >>>Cheers, > >>> > >>>A. > >>> > >>>To Unsubscribe: send mail to majordomo@FreeBSD.org > >>>with "unsubscribe freebsd-libh" in the body of the message > >>> > >>-- > >>Jordan K. Hubbard > >>Engineering Manager, BSD technology group > >>Apple Computer > >> > >> > -- > Jordan K. Hubbard > Engineering Manager, BSD technology group > Apple Computer >=20 --=20 Imagination is more important than knowledge - Albert Einstein --k+w/mQv8wyuph6w0 Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.1 (FreeBSD) iD8DBQE90vYEttcWHAnWiGcRAlO8AJ94t/7NpSuuzC1GxKpH6N3wV94wxQCcDm3k nttMgImlMj/m6kF7HX+YQMc= =wmVx -----END PGP SIGNATURE----- --k+w/mQv8wyuph6w0-- 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?20021114010157.GA307>