Copyright notices

Copyright notices

[Eli Zaretskii]

Since it took me by surprise, I thought that maybe other maintainers
aren't aware of this text from maintain.texi, which says how to update
copyright years in the files' copyright notices:

    To update the list of year numbers, add each year in which you have
    made nontrivial changes to the package.  (Here we assume you're using
    a publicly accessible revision control server, so that every revision
    installed is also immediately and automatically published.)  When you
    add the new year, it is not required to keep track which files have
    seen significant changes in the new year and which have not.  It is
    recommended and simpler to add the new year to all files in the
    package, and be done with it for the rest of the year.

The last sentence is what took me by surprise, when Richard Stallman
asked people who work on Emacs to add 2006 to _all_ the files in the
distribution, regardless of whether they were modified in that year or
not.

Perhaps we should have a script to do the same once a year, and stop
worrying about updating the individual notices when we modify each
file.

Re: Copyright notices

[Robert Dewar]

Eli Zaretskii wrote:

> The last sentence is what took me by surprise, when Richard Stallman
> asked people who work on Emacs to add 2006 to _all_ the files in the
> distribution, regardless of whether they were modified in that year or
> not.

Well copyright notices have little if any legal significance these
days, so I guess you can claim whatever you want without affecting
the copyright status, but it sure seems strange to me to claim
copyright for a new year if no substantive changes are made (I
don't think you can use a recursive argument that says that you
have updated the copyright because you *have* modified the file,
namely you have updated the copyright year :-)
> 
> Perhaps we should have a script to do the same once a year, and stop
> worrying about updating the individual notices when we modify each
> file.

If it is going to be done by algorithm, then it seems a bit of a waste
of time anyway. Given that the copyright lasts 90 years, I often wonder
why we bother at all with this copyright year update stuff.

Re: Copyright notices

[Eli Zaretskii]

> Date: Sat, 11 Feb 2006 06:13:43 -0500
> From: Robert Dewar <dewar@adacore.com>
> CC:  gdb@sources.redhat.com
> 
> Well copyright notices have little if any legal significance these
> days, so I guess you can claim whatever you want without affecting
> the copyright status, but it sure seems strange to me to claim
> copyright for a new year if no substantive changes are made

IANAL, but I'm guessing that erring on the other side, i.e. forgetting
to add a year when changes _were_ made, is being considered a greater
evil.

Re: Copyright notices

[Robert Dewar]

Eli Zaretskii wrote:

> IANAL, but I'm guessing that erring on the other side, i.e. forgetting
> to add a year when changes _were_ made, is being considered a greater
> evil.

Probably so, although since copyright notices since 1986 have
no real legal significance there is no danger of evil whatever
is done.

Re: Copyright notices

[Daniel Jacobowitz]

On Sat, Feb 11, 2006 at 01:07:46PM +0200, Eli Zaretskii wrote:
> Since it took me by surprise, I thought that maybe other maintainers
> aren't aware of this text from maintain.texi, which says how to update
> copyright years in the files' copyright notices:

If you think this was surprising, you should see the first version that
was posted to gnu-prog; it read as if doing so was required rather than
optional.  I understand the FSF's reasoning here (even if it seems
somewhat strange to me).

> Perhaps we should have a script to do the same once a year, and stop
> worrying about updating the individual notices when we modify each
> file.

I think this is not a bad idea.  The current set of notices are fairly
accurate - I won't say 100%, but maybe 95% - so I think that being
able to add a year by script would meet our needs.

Emacs already has some widgetry to add copyright years, doesn't it?
Can that easily work on a list of files?

-- 
Daniel Jacobowitz
CodeSourcery

Re: Copyright notices

[Eli Zaretskii]

> Date: Sat, 11 Feb 2006 09:36:22 -0500
> From: Daniel Jacobowitz <drow@false.org>
> 
> Emacs already has some widgetry to add copyright years, doesn't it?

Yes.  The command copyright-update updates the copyright years in the
current buffer by adding the current year to the existing list.

> Can that easily work on a list of files?

Not at this time, but I can write a command to do that, if we are
going to use Emacs for this.

Re: Copyright notices

[Daniel Jacobowitz]

On Sat, Feb 11, 2006 at 06:30:22PM +0200, Eli Zaretskii wrote:
> > Date: Sat, 11 Feb 2006 09:36:22 -0500
> > From: Daniel Jacobowitz <drow@false.org>
> > 
> > Emacs already has some widgetry to add copyright years, doesn't it?
> 
> Yes.  The command copyright-update updates the copyright years in the
> current buffer by adding the current year to the existing list.
> 
> > Can that easily work on a list of files?
> 
> Not at this time, but I can write a command to do that, if we are
> going to use Emacs for this.

Well, I did not find a good non-Emacs tool for doing this; I'd rather
using something that already exists, than write our own.  It should
just be a matter of finding all the files we consider "owned" by GDB
(all in the gdb directory, or sim, or include/gdb/ ?) and excluding
the generated ones.

-- 
Daniel Jacobowitz
CodeSourcery

Re: Copyright notices

[Eli Zaretskii]

> Date: Sat, 11 Feb 2006 13:27:27 -0500
> From: Daniel Jacobowitz <drow@false.org>
> 
> Well, I did not find a good non-Emacs tool for doing this; I'd rather
> using something that already exists, than write our own.  It should
> just be a matter of finding all the files we consider "owned" by GDB
> (all in the gdb directory, or sim, or include/gdb/ ?) and excluding
> the generated ones.

The above specification of which files to include/exclude sounds very
vague.  Is there some existing script or algorithm that would allow to
find the relevant files programmatically?  If not, then perhaps the
first good approximation would be a Dired-style command: mark the
files you want to update (or mark all of them, then unmark the ones
you don't want), and invoke a command that will update all the marked
files.  WDYT?

Re: Copyright notices

[Daniel Jacobowitz]

On Sat, Feb 11, 2006 at 09:05:00PM +0200, Eli Zaretskii wrote:
> > Date: Sat, 11 Feb 2006 13:27:27 -0500
> > From: Daniel Jacobowitz <drow@false.org>
> > 
> > Well, I did not find a good non-Emacs tool for doing this; I'd rather
> > using something that already exists, than write our own.  It should
> > just be a matter of finding all the files we consider "owned" by GDB
> > (all in the gdb directory, or sim, or include/gdb/ ?) and excluding
> > the generated ones.
> 
> The above specification of which files to include/exclude sounds very
> vague.  Is there some existing script or algorithm that would allow to
> find the relevant files programmatically?  If not, then perhaps the
> first good approximation would be a Dired-style command: mark the
> files you want to update (or mark all of them, then unmark the ones
> you don't want), and invoke a command that will update all the marked
> files.  WDYT?

Well, what I'd do would be shell-script the above: have a script print
out a list of files, and remove certain others, based on a manually
defined list.

(That could just be one big invocation of GNU "find").

I think the exceptions are going to be pretty small: generated
documentation (e.g. man pages, info files), gdbarch.c / gdbarch.h,
configure scripts, et cetera.  Of course, there will still be a list
of places that need to be updated by hand (including gdbarch.sh and
the texinfo manuals, probably).

-- 
Daniel Jacobowitz
CodeSourcery

Re: Copyright notices

[Eli Zaretskii]

> Date: Sat, 11 Feb 2006 14:07:44 -0500
> From: Daniel Jacobowitz <drow@false.org>
> 
> Well, what I'd do would be shell-script the above: have a script print
> out a list of files, and remove certain others, based on a manually
> defined list.

If the exclude list is manually defined, Emacs could use it directly.

> (That could just be one big invocation of GNU "find").

Emacs already has a command to run `find' and present the results in a
Dired-style buffer.

> Of course, there will still be a list of places that need to be
> updated by hand (including gdbarch.sh and the texinfo manuals,
> probably).

What is special about Texinfo files?  The fact that Copyright appears
there in several different places? or did you mean something else?

Re: Copyright notices

[Daniel Jacobowitz]

On Sat, Feb 11, 2006 at 09:18:03PM +0200, Eli Zaretskii wrote:
> > Date: Sat, 11 Feb 2006 14:07:44 -0500
> > From: Daniel Jacobowitz <drow@false.org>
> > 
> > Well, what I'd do would be shell-script the above: have a script print
> > out a list of files, and remove certain others, based on a manually
> > defined list.
> 
> If the exclude list is manually defined, Emacs could use it directly.

OK.  I don't know the first thing about Emacs scripting :-)  So I think
I'll defer to you on this...

> > (That could just be one big invocation of GNU "find").
> 
> Emacs already has a command to run `find' and present the results in a
> Dired-style buffer.

Gotcha.

> > Of course, there will still be a list of places that need to be
> > updated by hand (including gdbarch.sh and the texinfo manuals,
> > probably).
> 
> What is special about Texinfo files?  The fact that Copyright appears
> there in several different places? or did you mean something else?

Right; I just figured that the in-text copyrights would have to be
updated by hand even if we auto-updated the one at the top of the
file.  Maybe that's not right.

On Sat, Feb 11, 2006 at 02:20:51PM -0500, Bob Rossi wrote:
> Of course, instead, you could use grep to look for files that 'find'
> produced and search for "Copyright (C)" to know what files should be
> updated. That way, you don't have to store a list of files that needs to
> be maintained.

No, you want to have an exclusion list rather than inclusion - you
don't want to manually futz with the copyright header in an
autoconf-generated script, for instance.

-- 
Daniel Jacobowitz
CodeSourcery

Re: Copyright notices

[Bob Rossi]

On Sat, Feb 11, 2006 at 02:07:44PM -0500, Daniel Jacobowitz wrote:
> On Sat, Feb 11, 2006 at 09:05:00PM +0200, Eli Zaretskii wrote:
> > > Date: Sat, 11 Feb 2006 13:27:27 -0500
> > > From: Daniel Jacobowitz <drow@false.org>
> > > 
> > > Well, I did not find a good non-Emacs tool for doing this; I'd rather
> > > using something that already exists, than write our own.  It should
> > > just be a matter of finding all the files we consider "owned" by GDB
> > > (all in the gdb directory, or sim, or include/gdb/ ?) and excluding
> > > the generated ones.
> > 
> > The above specification of which files to include/exclude sounds very
> > vague.  Is there some existing script or algorithm that would allow to
> > find the relevant files programmatically?  If not, then perhaps the
> > first good approximation would be a Dired-style command: mark the
> > files you want to update (or mark all of them, then unmark the ones
> > you don't want), and invoke a command that will update all the marked
> > files.  WDYT?
> 
> Well, what I'd do would be shell-script the above: have a script print
> out a list of files, and remove certain others, based on a manually
> defined list.
> 
> (That could just be one big invocation of GNU "find").
> 
> I think the exceptions are going to be pretty small: generated
> documentation (e.g. man pages, info files), gdbarch.c / gdbarch.h,
> configure scripts, et cetera.  Of course, there will still be a list
> of places that need to be updated by hand (including gdbarch.sh and
> the texinfo manuals, probably).

Of course, instead, you could use grep to look for files that 'find'
produced and search for "Copyright (C)" to know what files should be
updated. That way, you don't have to store a list of files that needs to
be maintained.

Bob Rossi

Re: Copyright notices

[Eli Zaretskii]

> Date: Sat, 11 Feb 2006 14:20:51 -0500
> From: Bob Rossi <bob@brasko.net>
> 
> Of course, instead, you could use grep to look for files that 'find'
> produced and search for "Copyright (C)" to know what files should be
> updated. That way, you don't have to store a list of files that needs to
> be maintained.

The relevant Emacs command already does that--it searches for the
telltale Copyright strings and updates them, it doesn't just blindly
add a Copyright notice where none is present.

But that doesn't solve the problem that Daniel pointed out, since the
generated files also have the Copyright notice.  See the various
configure scripts, for example.  We don't want to update those, but
instead should regenerate them whenever their source files get updated
with another year's copyright.

Re: Copyright notices

[Jim Blandy]

On 2/11/06, Daniel Jacobowitz <drow@false.org> wrote:
> Emacs already has some widgetry to add copyright years, doesn't it?
> Can that easily work on a list of files?

Sure, it'd just take a little bit of lisp.

The way I've got it hooked up is to have it check automatically for a
Copyright notice to update when I save a file:

;;;; Offer to update copyright years.
(add-hook 'write-file-hook 'copyright-update)

I hit C-x C-s, and it offers to update the year for me if it's not current.

There's disagreement on what the right thing is, legally.  Certainly
the FSF has lawyers on hand to give them opinions.  But if I'm
recalling this right, Dan Berlin (who is a lawyer), argues that to
slap a copyright year on a file when that file has not been
significantly changed in that year amounts to a false assertion. 
That's the year that determines when the copyright expires, so you
would appear to be trying to exaggerate the amount of protection the
file should receive.  And false assertions like that weaken your case,
in the court's eyes.

But don't worry: Disney will make sure that nothing ever goes out of
copyright again, regardless of when it was last changed, so this is
all moot.