Gentoo boostrap woes

[afterthefall]

After accidentally hosing my Slackbox in an upgrade, I've decided to give Gentoo a try, but unfortunately, I am running into some problems with the bootstrap process.

I have configured my make.conf to taste, and have performed all necessary partitioning and so on. The bootstrap performs flawlessly until it reaches ncurses. The actual error is as follows:

!!! ERROR: sys-libs/ncurses-5.3-r1 failed
!!! function econf, line 303, exitcode 1
!!! econf failed

Other info:
- using 1.4rc1 livecd to boot
- using 1.4rc4 stage1 tarball to install
[edit] x86 arch on an AMD k6

This rc difference shouldn't matter, since the entire system is brought up to date with the emerge sync - although there was a small complaint about portage not being the latest release, but since everything runs smoothly up to the ncurses incident, this should really be no problem.

Anyone have any suggestions?

[beat_feet]

Did you emerge sycn or emerge rsync after changing root to the /mnt/gentoo?

[endoalpha]

What are your USE flags and what are your CFLAGS/CXXFLAGS settings? What is your CHOST setting? Also, do not bootstrap with ACCEPT_KEYWORDS="~x86" option.

[afterthefall]

I did emerge sync immediately after mounting on /mnt/gentoo and /mnt/gentoo/boot. After the sync, I got a message regarding an out of date portage. and ironically, emerge portage went off without a hitch.

Currently, I've gotten the 586 stage 2 install going right now - unfortunately I need a comp by Monday, so this will have to do (unless the error was because of the ACCEPT_KEYWORD flag, which means I should have enough time to get this puppy rolling on Saturday and Sunday).

Off the top of my head....

CFLAGS/CXXFLAGS was set to:
-march=k6 -03 -pipe (not terribly agressive)

CHOST was set to the i586 pc gnu linux string

USE flags are both various and sundry

BUT.....I did set the ACCEPT_KEYWORD="~x86"


- atf

[hard candy]

One more thing to check on, make sure you have ncurses set as one of your use flags.

[endoalpha]

Unless you are prepared for lots of trouble, don't use the ACCEPT_KEYWORDS="~x86".

[afterthefall]

ncurses, was definately in the USE flags list, so it sounds like problem is the ACCEPT_KEYWORDS="x86". Time to halt this one, reboot and try from stage 1 again *crosses his fingers*.

-atf

[afterthefall]

The bootstrap went off without a hitch. Took 17 or so hours (2am Friday to 9pm Saturday), but it made it through flawlessly.

Unfortunately, I've run into yet another problem - actually booting the bastard up. My configuration is basically by the book:

/dev/hda1 (ext3) is MBR
/dev/hda2 is swap
/dev/hda3 (reiserfs) is root

I've compiled reiser support into the kernel, as well as setting/unsetting necessary Gentoo kernel features (/proc, /dev, etc.) and I have lilo configured to boot hda1, swap hda2 and root hda3. Metalog and vcron are happily residing as their respective defaults.

Everything seems just fine, until I unmount and reboot, at which time I get the perky little lilo window and try to boot into Gentoo. The "loading gentoo..." message pops up and then for no sane reason, the computer reboots itself and returns to lilo, merrily looping the process into oblivion.

Now, I'm sure that linux would be quite pleased with itself if it could continue this, but unfortunately, I have very different plans for this machine (involving functionality and so forth....) I'm breaking my cherry on kernal compilation with this install, so I really am not exactly sure whether or not I've compiled an unbootable kernel or what.

fstab looks fine, and there seems to be no problem with lilo. For completeness, I tested and got the same results under GRUB, so I have the sneaking suspicion that either I have an unbootable/errant kernel, or I am facing some mounting (no pun intended) problems.

-atf


PS - I'll follow up with a settings post after I get a few opinions from the more seasoned users.