On 7/17/07, Ken Moffat <[EMAIL PROTECTED]> wrote: > > What do your bootscript patches do, and how do you view their risk > ?
Should be backwards compatible. A couple I have queued up but keep forgetting to commit them. Add an early script to quiet the kernel log level, then kill the level handling in init.d/modules. http://linuxfromscratch.org/pipermail/lfs-dev/2007-June/059501.html Second one would kill this bug: http://wiki.linuxfromscratch.org/lfs/ticket/2026 Two other things I haven't done, but would like to (meaning, I may punt for now since I haven't tested). Support for IP aliasing/labelling: http://linuxfromscratch.org/pipermail/lfs-dev/2007-June/059442.html Fixing pidofproc to actually do the right thing w.r.t pid files. Dennis Perkins actually put that part together. The other change he does there is make the pid list local and echo it back. This is slightly different and possibly not backwards compatible since the current behavior has the other *proc() functions just using the global pidlist variable created in pidofproc. I prefer Dennis' way and can fix the functions to catch the result in it's own variable, but it could break existing scripts which call pidofproc directly. In most cases, though, pidofproc is being called indirectly through loadproc or statusproc. The difference is this: current: pidofproc -s $1 # creates pidlist for pid in $pidlist; do ... new: pids=$(pidofproc -s $1) # pidlist is local to pidofproc for pid in $pids; do ... http://linuxfromscratch.org/pipermail/lfs-support/2007-July/033304.html That may be able to be split into two parts since the pid file handling is a bug and should be safe to fix. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/lfs-dev FAQ: http://www.linuxfromscratch.org/faq/ Unsubscribe: See the above information page
