Rocky Bernstein wrote: > On Mon, Feb 13, 2012 at 4:19 AM, Jim Meyering <[email protected]> wrote: > > Rocky Bernstein wrote: > > It's been a couple of months since I first sent this was sent without > nary > > an ack. Comments? > > Hi Rocky, > > su is barely on life support in coreutils. > Meaning that it's no longer really maintained. > We stopped installing it (by default) back in the 2007-2008 > time frame (6.9.90, 7.0). We nearly removed it altogether. > > Do you build/install coreutils yourself, or use it via a distribution? > If the latter, which distribution? > > CenOS 6.2 which seems to use coreutils 8.4 > > I don't remember looking carefully at your patch before, > but glanced through it just now. Here are some suggestions > if you'd like to pursue it: > > - use "error (0, 0, ...", not fprintf ( > - indent with spaces, not TABs > - use gnu indentation/formatting style (esp. wrt braces) > - mention the change in NEWS > - for a behavior change like this, we would like a test case > > The above seems all reasonable and easily doable. But are you suggesting is > even > after getting this into the next coreutils, no OS is likely to use it? > > The basic idea is that a colleague and I were confused by the fact that -l and > -p conflict although this was not apparent in the documentation. I wound up > downloading the source code to understand fully. > > So something even as simple as mentioning this in doc/coreutils.texi would > have > been helpful (which is the first part of that patch). > > Going further I realized that it is probably erroneous to supply both -l -p > and > that can be easily warned. > > But if there is a better place such to report such problems or get code > improved, let me know! Thanks.
I see that Fedora still uses su from coreutils, too, so this is a worthwhile change.
