This is the mail archive of the
libstdc++@sources.redhat.com
mailing list for the libstdc++ project.
Re: libstdc++-v3 2.90.9
- To: Phil Edwards <pedwards at disaster dot jaj dot com>
- Subject: Re: libstdc++-v3 2.90.9
- From: Benjamin Kosnik <bkoz at redhat dot com>
- Date: Mon, 13 Nov 2000 12:37:39 -0800 (PST)
- cc: libstdc++ at sources dot redhat dot com, pfeifer at dbai dot tuwien dot ac dot at
> 1) Copy the docs repository into the wwwdocs repository, thus
> preserving history.
> 2) Do not create any links into the new htdocs/libstdc++/ tree
> until everything works.
> 3) Change structure of new libstdc++ docs as necessary to work with
> the wwwdocs preprocessing.
> 4) Remove the contents of the docs dir, replace with a pointer
> to the new pages. Explain that there will be docs there only
> in snapshots and releases.
> 5) Start a new ChangeLog in wwwdocs/htdocs/libstdc++ to keep track
> (it still is official documentation, after all).
Ok. I'm a bit confused. I take it libstdc++-v3/docs is now going to be
replaced by wwwdocs/htdocs/libstdc++?
I'm assuming there's a good reason why we can't just keep the docs as
part of the libstdc++-v3 source distribution, as is currently the case? I
kind of liked the current setup, but I understand if it somehow isn't
going to work.
> I'm not entirely thrilled with (4), since it would mean that users checking
> out CVS won't get the local copies of the build instructions. They'd still
> be on the web, but that's an extra step, and much slower besides. Hmm.
Indeed.
> We tried to think up some way of getting help from CVS modules, like having
> "libstdc++-v3" actually check out the real sources *and* the web pages,
> and silently do a merge on the local user's end. That ain't gonna happen;
> the directory trees are too far apart for the modules trick to work, and
> we can't have the CVS client execute arbitrary shell commands to mv stuff
> around (for good reason).
Hmm. If these are now going to be two separate modules, then it seems
ill-advised to try and force the old "unified" behavior onto two disjoint
entities. Let's try to keep this simple.....
-benjamin