This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Manual patch: terminology

"Joseph S. Myers" <> writes:

> On Mon, 2 Feb 2004, Gerald Pfeifer wrote:
> > Are you saying that you are "reverse engineering" those patches for
> > the GCC 3.3 manuals that were made on money from the FSF but never
> > contributed back to GCC?
> Yes.  Chris Devers vanished so all we have to work from is the 15000-line
> diff attached to bug 13573 (full of gratuitous paragraph reformattings).  

 This doesn't look fun.  I was hoping to avoid having gratuitous paragraph
reformatting and rewording while the book was being prepared.  Oh well.

> At some point we'll need to consider that patch's general removal of all
> uses of the first person (far beyond just the obviously dubious RMS-isms
> such as "Coming as I do from a Lisp background, I balk ..."), and of most
> parentheses, without the benefit of the original rationale from when these
> changes were made.  I think that for future editions we should insist on
> only co-operating at all with GNU Press if they work directly from the
> relevant CVS release branch, with the printed manual having sources that
> are exactly the branch sources as of some date (tagged to identify them as
> the sources of that manual edition) and no changes going in to the manual
> without passing the usual review processes and first going on mainline.  
> (There is an FSFPRINT conditional I added specifically to support
> differences in the ordinary manuals and the FSF-printed versions.  They
> didn't use it.  Certain changes, such as the different form in which the
> internals manual is referenced from the published edition, clearly should
> be handled by FSFPRINT-conditional macros.)
> -- 
> Joseph S. Myers

 This plan may generate more traffic than you want.  However, what happened
with this version is last year at the FSF meeting Lisa, opus, went to dinner 
with a bunch of us and got a bunch of people started on the using gcc
manual.  We created a project on savannah, gnupress, and used the bug-gnupress
mailing list to co-ordinate ourselves.  Once we had were ready, we got many
of the people subscribed to to subscribe to bug-gnupress
and to take a chunk of the gcc manual to proof.  It turned out pretty well,
except we had a couple proof readers who seemed determined to rewrite the
entire thing.  After some time, Lisa hired someone to work at the FSF and
finish the book.  Unfortunatly, whoever was working at the FSF didn't even
coordinate with the bug-gnupress list.

 Anyway, I'm not too sure if the above story has any meaning.  However, I 
think what I would like to see is to have a branch off of the release branch
for the gnupress manual which could be regularily merged one or both ways.
Obviously merging back into mainline would require the normal procedure,
but I don't think a gnupress branch would require anything too special.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]