Re: Bug#3146: 1.1 installation: ae doesn't work
Bill Mitchell writes ("Re: Bug#3146: 1.1 installation: ae doesn't work"):
> I'm not sure you're reading that snippet from the FSSTND correctly.
> I'm not sure you're reading it incorrectly either, but I read it
I *wrote* that snippet in the FSSTND ! The description of e2fsck is
only an *example*. It says quite clearly:
Note that an implementation does not need to contain all the files and
directories specified in this standard to be compliant or compatible.
It is merely necessary for those files that it does contain to be
I don't see how you can think that the FSSTND mandates that a
distribution must contain /bin/ed after reading those two sentences.
I'm not saying that a distribution shouldn't contain /bin/ed, but that
the FSSTND is irrelevant and that some common sense would be a good
idea. Just because ed isn't on the base disks, honestly ...
(And yes, I prefer ed as a recovery editor, but this is my personal
> The clarifying example given said that e2fsck would not be required
> if there were no ext2 filesystems to be checked. It did not say
> whether or not e2fsck would be required if there were ext2 filesystems
> but there were ten other tools to check them with. That's the
> situation with regard to ed. There are text files to be edited, and
> there are other editors available to edit them.
The FSSTND doesn't specify what tools people must provide. You can
have a perfectly FSSTND-compliant system that uses ext2 on all of its
filesystems but doesn't provide e2fsck or any other way to check
filesystems. This system is broken, true, but its brokenness is (a)
obvious and (b) out of the FSSTND's scope.
> Rather than debating about what we each think the FSSTND might
> be trying to say, how about getting a ruling on whether /bin/ed
> is or is not required by the FSSTND? Dan Quinlan, are you out
Please, can't you just take my word for it ?