Hi, I thought I'd try my hand at lfs. I have gone for the ALFS install using nALFS-6.1. I have sorted the configuration side of things, then I ran "runit.sh". all goes well. I started the process ("s" then "c"), but the installation stops when it starts to unpack binutils saying "wrong md5sum". Sure enough, the md5sum is different to that stated in the nALFS install script and is wrong for other packages too. Has anyone else had this problem?
I found a thread which suggested I do a "chmod 644" on all the destination sources (contradiction there!), I did this but it made no difference, what else could cause bad md5sum's?

edit:
binutils-2.15.94.0.2.2.tar.bz2
expected md5sum= e9bcf86d9e6b5a2d840e76f22a3fce5f
actual md5sum= 996ae4af987cbef7c4b20e7b79db3b97 (checked on this computer too)

edit2:
both the md5 and sha sums are correct for the iso image I used to burn the lfs livecd
lfslivecd-x86-6.1-3.iso

edit3:
I think I'll start from scratch anyway and do it all manually. After all, that's the whole point of lfs isn't it!

I took the time to build LFS 6.1 both ways, using Slack as a host and using the LFS liveCD.........both systems are pretty solid/stable........try it if you have the time.


I have not done the hardened OR automated versions of LFS......because there is really no learning curve with either of these at all........

I also ran into this using the LFS-6.1 version using the LiveCD. In my investigations, IIRC the
nALFS profile for 6.1 was still using md5sums from the 6.0 version in the packages.ent file (in
the config subdirectory) for binutils as well as a few other packages. In the end, since I new the
packages were valid [you can get actual packages and md5sums from whatever website
normally hosts the package], I just editted the packages.ent file and substituted the proper md5sum
value for the old one. First attempt at getting LFS up and running took a couple of days to finally
get past the initial build stage as I would end up starting from scratch each time I had to perform
this type of alteration. Eventually I did end up with a running LFS system, though, so I say keep at
it.

Second note, I had problems with compiling flex. To get by this, I went to a second terminal
(virtual of course) and compiled and installed the package by hand, then went back to the nALFS
window, unmarked the finished steps, and continued with the process.

Side note: with my AMD Athlon XP2000+ based system and 512M RAM, it took a good 5.5 hours of
run time to complete, including the "flex" compilation above. I still haven't done the BLFS stuff
yet.

Rick

Thanks for your replies. I don't think I'll be changing anything, I'll just use the old fashioned manual method, though I'm sure anyone else who has tried ALFS 6.1 will find this thread useful.
freakyg, you said you did the compile under slack, I didn't know this was possible, I think I'll try it. Anything I should bear in mind when compiling under slack? It is going to be on a HP XE3 notebook, I thought a notebook would be ideal to use as the hardware is fixed and will never change.