On 2/1/23 16:26, enh via Toybox wrote: > i think there's a catch-22 there though --- the trouble is that until > an editor meets the "good enough that i'm using it to develop itself" > bar, it really struggles to make much forward progress.
In 1991, Linus Torvalds accidentally fried his minix partition by pointing his terminal program at /dev/hda instead of /dev/sda, and not noticing until it had overwritten several kilobytes with "ATDT university-number\n+++ATH" repetitions. Luckily his /home with the Linux source code and the compiler he'd made run under Linux was on the second partition, and his early Linux versions all booted from floppy. He then had a choice of not just reinstalling Minix but patching it back into a usable 32 bit system with the multiple third party patch stacks that Tanenbaum refused to accept upstream (which took him a month to do the first time), or dogfood Linux. Aboriginal Linux drove my busybox development. I'm trying to get mkroot to drive my toybox development but it's not to the point I can build it under itself yet, or build LFS (let alone AOSP) under it. That's why I keep circling back to the shell: I cheated and built several gnu tools in aboriginal to complete the circuit until the toybox bits were ready, and never DID properly tackle the shell there. I've ALMOST got toysh where it needs to be, but it's a big lift. And the darn compiler builds are a moving target... Rob _______________________________________________ Toybox mailing list Toybox@lists.landley.net http://lists.landley.net/listinfo.cgi/toybox-landley.net
