Dan McGhee wrote: > On 07/28/2010 08:19 PM, Jeremy Huntwork wrote: >> Originally, the CD came about because of the general sentiment that >> installing a host distro to get started building LFS was annoying. Doing >> so usually meant you couldn't use the whole hard drive for LFS without >> doing some crazy acrobatics. A secondary purpose was to have a nice >> packaged way to download the book, the sources you needed and a host OS >> all in one shot. If you were forced to work offline, you had everything >> you needed. >> > And this is a wonderful concept. I used it more than once. >> I think those reasons are still essentially the main ones. Merge them >> together and generalize it a bit, and what you get is: >> >> 'Provide a packaged system that equips a user with the tools they need >> to build LFS and obtain online support while building.' >> >> What do you think? Can it be improved? Does it miss any purpose that the >> LiveCD should try to fill? >> > This is adequate and simple. Let me, though, throw something in that may > add to the complexity and point the way to a heavy "maintaining" effort. > I have no idea what makes the current scripts oppressive, and it seems > to me that this may have been the main reason the project ended in its > current state. > > As stated, what would a LiveCD like this give a user that the user > couldn't get by using a livecd from any distro? Yeah, installing another > distro as a host system created all the negatives you mentioned--I've > walked that path. But one does not need to "install the distro." All > that's required is a cd similar to Ubuntu that boots into a fully > functional version of Ubuntu. I'm sure the other distros have something > similar.
Yes, but they all have the same fatal flaw, the needed development tools to build an LFS system aren't in the livecds. They don't include any tool that isn't needed on them, the size constraints. Bison, gettext-devel, are just two examples of missing tools. > The question then becomes, "Do we want the LFS-LiveCD to be just like > any other livecd, or do we want it to be uniquely LFS?" The follow-on > question becomes, "If we want a unique livecd, how do we do it?" It would have to be uniquely lfs. > One of the attractions LFS has for me is that everything is built > uniquely on my machine using my hardware. I don't have to worry about > any "optimizations" or package configuration that may not quite suit my > system. I think that this is the main reason for the LFS-toolchain. So > to make this LiveCD uniquely LFS, we could "build" an iso image that > someone could burn and use as their host system. I think that this is > what happened the first time I used the LiveCD years ago and I really > liked it. That is still what it provides. > If others don't want to go down that road, then I have one wording > recommendation for the purpose: > > 'Provide a packaged system that equips a user with the tools they need > to build [a basic] LFS [system] and obtain online support while building.' > > I think "basic," or other words to relay the same concept, lets people > know that the end result will be a bootable system as it exists at the > and of Chapter 8. > > Dan Jaqui -- http://linuxfromscratch.org/mailman/listinfo/livecd FAQ: http://www.linuxfromscratch.org/faq/ Unsubscribe: See the above information page
