Subject: shoelace: fsck lockup
Date: Sat, 9 Nov 91 01:19:04 PST
From: pmacdona@sol.UVic.CA (Peter MacDonald)
It seems that the failure of shoelace to fsck on my hard drive
is due to the number of inodes. An alloc of memory is done which
wipes out the stack! Ouch. Bitten by the 16bit bug again (the
very reason why most of us are using minix). Anyways, if we
can find a way to reduce the memory requirements of shoelace,
then fsck and mkfs should work within shoelace. I suspect that
removal of this excess funtionallity is required. For example,
commenting out the single call to fsck in shoelace resulted in the
code size dropping from 49K to 30K!!!. Add stack and data to this
and join the over 64K club for a voyage into lala land.
Either way, the code to perform these functions already exists and can be
extracted from shoelace for standalone products. If no one objects
shortly I will begin this project.
BTW: Earl Chew is the author of shoelace and must be given credit for
its existance. Please speak up if you know if the code for fsck in
nic.funet.fi is PD. It appears to be suspiciously like Earls shoelace code.