Re: ensure that generated files are read-only

[Paul Eggert]

Jim Meyering <address@hidden> writes:

> I'm surprised that the compromise of adding advisory comments rubs
> you (Bruno) so hard the wrong way.  Does anyone else object to
> adding both lines?

I'm afraid I'm mildly annoyed by them too.  I use Emacs with (setq
enable-local-variables 0), so I get this prompt:

  Set local variables as specified in -*- line of foo.c (y or n)

when I visit a file, and this means my typehead can get
misinterpreted.  If the file name is sufficiently long and my window
sufficiently narrow, the screen hops around a bit too, which is
visually distracting.  I'd rather have a solution that didn't involve
that question.  (I realize I can shut it off for all files, but that
has other negative consequences.)

Also, these changes mean I can't easily copy files from coreutils back
to gnulib; I have to remember to remove the DO NOT EDIT notice.
Again, no big deal, but it's one other minor thing to do (or forget to
do :-).  Currently, 346 files shared between gnulib and coreutils have
the DO NOT EDIT notice, so it's quite likely I'll run into the
problem.

Since the gnulib and coreutils copies are now identical in most cases
except for the DO NOT EDIT notice, I was thinking of having
"bootstrap" use symbolic links instead of copies, for source files
whose contents are identical if you omit the DO NOT EDIT notice.  This
would mean I'd edit my gnulib copy if I visit my coreutils copy, which
is almost always what I want to do.  It might also mean another
annoying question about whether I want to follow a symbolic link to a
CVS controlled file, but I have fewer qualms about shutting that
message off.