On 1/11/06, Ken Moffat <[EMAIL PROTECTED]> wrote: > Certainly farce shows perl, vim, and nscd were compiled at different > times - open farce-substitiutions in view to see what got replaced by > tokens (and yes, 2.3.6 in nscd is probably a library version although it > got matched to regex KV02).
OK, I looked at farce-substitutions. These look fine now. > My gut feeling is that cleaning the toolchain might remove a lot of > differences (ar archives, and programs linked statically) which (at > least in lfs-svn) show up between the first and second builds, > particularly much of the 'differs, but same code'. I changed the toolchain completely to the DIY way, and it didn't change much on my newly ordered system. It probably won't change much since glibc is built the same way in both chapters. Ken, why doesn't farce do stripping on statically linked binaries? Greg's ICA does this, and it seems safe to me. Here's an unsolicited suggestion for farce, too. Declare variables as local inside functions unless the variable is needed in the global namespace. Just makes things simpler because you don't have to worry about erroneous results if a variable already has a value assigned. Again, thanks for farce. It's helping tremendously in the ICA. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/lfs-dev FAQ: http://www.linuxfromscratch.org/faq/ Unsubscribe: See the above information page
