Hello. My situation has got better and more stable. I have been thinking about hlfs, and I have an hour a day, or so, that I can put back into it.
I have been brainstorming, and have a lot of ideas. But first things first. I'd like to use the LFS-6.7 book to get HLFS going again. This was discussed a bit before, but I don't remember very well how the discussion turned out. Using the LFS stable book saves the time needed to stabilize the package versions. This work is already done. Time would be spent more directly on project goals, instead of duplicating the efforts of LFS. I have lost the idea that a reboot is needed. New versions of udev need a recent kernel version, and almost every distribution includes file system posix capabilities, so these issues can be added to the host system requirements instead of a reboot. So that's it for the moment. Just add what worked before to the LFS book, and go from there. I have some new ideas, like having different configurations for a development system, a server, and a client, mainly with networking, but this can be brought up at a later time. I also thought about a problem with using user based package/file management. Make files don't necessarily fail when there is a permission denied during install. This is a bit of a problem when installing packages that are outside of the HLFS book, because the admin would need to read the make output to check, and this would lead to human error. I'm preferring the Trip unionfs method, because it would resolve this issue, and still allow non-root to install packages with no risk of damaging the existing system. It also allows file permissions to be normalized before they're actually installed, instead of after. Just a thought. Do any of you have opinions about anything here? robert
signature.asc
Description: This is a digitally signed message part.
-- http://linuxfromscratch.org/mailman/listinfo/hlfs-dev FAQ: http://www.linuxfromscratch.org/faq/ Unsubscribe: See the above information page
