Date: Sat, 6 Oct 2018 02:06:52 GMT From: Kathe <ka...@sdf.org> Message-ID: <201810060206.w9626qtd012...@sdf.org>
| From: Greg Troxel <g...@lexort.com> | Often, features date from Seventh Edition, Sixth, or even earlier. The | ed(1) man page says it appeared in First Edition; I can only say it was | there in Sixth. -x was added in 7th edition, if you think you saw it in 6th you are either mis-remembering, or using sources closer to 7th edition (as the names suggest, the manuals are the definitions, the sources were in a constant state of flux, but the commonly distributed 6th edition had no -x in ed). | The BSD way is to respect tradition, absent a compelling reason. Aside from that (and of course, assuming ed is your editor of choice), -x is the only way to safely edit an encrypted file. The alternative would be to unencrypt fiirst, then edit, then encrypt again - which leaves the plaintext version in the filesystem (consider what happens when the drug lord is editing his customer file, and there's a raid ... he can pull the plug on the computer, but does not have tiime to remove the unencrypted file, the ed /tmp copy of it, or all the blocks that contain all that evidence...) | isn't having a tightly built ed a compelling reason enough? No. | btw, why is /bin/ed dynamically linked? Everything in /bin is dynamically linked. It means that they all get the benefit of libc (etc) bug fixes. | wouldn't it need to be usable even under extreme conditions? That's /rescue/ed kre