maximilln's Journal: Happy LFS

I made it. I made my goal. By the end of Sunday night, before I went to bed, I had successfully checked Hotmail (ensuring PSM) in Mozilla, was happily connected to my favorite IRC networks, and was playing CDs and streaming audio from the network. UDE handles my windows and Qiv handles the background. Shaded and faded aterms make for attractive looking workspace. /etc/X11/app-default/XTerm needs some tweaking, though. Xterms aren't accepting select and paste using gpm. There's also the issue with the delete key. There's an LFS hint about keymapping and interpreting.

Libmpeg required an explicit endian #define on the K6-3. Firefox won't start on either system (the P2 or the K6-3), complaining about an invalid function on line 451 of run-mozilla.sh. I built mozilla three times. The first time I built it and forgot --enable-crypto. You'd think that'd be a default. The second time I built it verbatim by the bLFS book. Before I went to bed last night I rebuilt mozilla again trying to pare down the amount of excess configure flags. This morning I did "make install" and skirted all of the other stuff in the bLFS book just to see if it works. It does.

Funny thing is that after checking my Hotmail, doing some other stuff, and coming back to check Hotmail again, mozilla crashed claiming a segmentation fault at line 451 of run-mozilla.sh. That is the same line that causes Firefox to refuse to start on my machines. I thought about submitting a bug report but don't feel like taking the time to rebuild mozilla _again_ with the debugging and test output. It takes four hours to build that beast on the P2 400 and the K6-3 400.

Today I formally mapped out the process of my LFS installation scripts. Tonight I'll install emacs, tidy up the scripts, and maybe submit them to the lfs mailing lists. Now that I've mapped out the structure it's becoming easier to identify the functions (subroutines) that I will need to write if I want to code them in C. I'd like to do that.

Maybe I should write them in x86 asm. I'd probably never have the time in a lifetime to write all of the subroutines necessary to pull that off. Then again, it's really nothing more than a little disk I/O and getting a few inputs from the user. The learning curve to figuring out the environment of x86 asm would be harder than actually writing all of the loops.