History before 1999

classic Classic list List threaded Threaded
7 messages Options
Reply | Threaded
Open this post in threaded view
|

History before 1999

Eli Zaretskii
"cvs log" shows that the oldest version of GDB sources for which
there's information in the CVS is from April 1999.  Is there any
reasonably uncomplicated way to do the equivalent of "cvs diff" and
"cvs annotate" for older versions?

Reply | Threaded
Open this post in threaded view
|

Re: History before 1999

jimb (Bugzilla)
On 1/7/06, Eli Zaretskii <[hidden email]> wrote:
> "cvs log" shows that the oldest version of GDB sources for which
> there's information in the CVS is from April 1999.  Is there any
> reasonably uncomplicated way to do the equivalent of "cvs diff" and
> "cvs annotate" for older versions?

Was it really 1999?  Anyway, before sourceware.org got started, the
only revision controlled history was in Cygnus's private CVS
repository.  That was mixed in with various NDA'd stuff, including
some in the core files of GDB, so we couldn't publish the whole
history without a lot of work sanitizing it.

So, I don't think there's any way for you to get history farther back.

Reply | Threaded
Open this post in threaded view
|

Re: History before 1999

Eli Zaretskii
> Date: Sat, 7 Jan 2006 09:13:59 -0800
> From: Jim Blandy <[hidden email]>
> Cc: [hidden email]
>
> So, I don't think there's any way for you to get history farther back.

Too bad.  Fortunately, there are old GDB releases in the old-releases
directory of the sourceware server, so I guess I could use them as
snapshots of the old repository.

Thanks for the feedback.

Reply | Threaded
Open this post in threaded view
|

Re: History before 1999

Stan Shebs
Eli Zaretskii wrote:

>>Date: Sat, 7 Jan 2006 09:13:59 -0800
>>From: Jim Blandy <[hidden email]>
>>Cc: [hidden email]
>>
>>So, I don't think there's any way for you to get history farther back.
>>
>
>Too bad.  Fortunately, there are old GDB releases in the old-releases
>directory of the sourceware server, so I guess I could use them as
>snapshots of the old repository.
>
>Thanks for the feedback.
>
In theory, an energetic person could graft old public releases into
the repository, making a sort of synthetic history. Easier in SVN
perhaps?

Didn't we do regular snapshots in between releases? If anybody
still has those around, they could supplement the release history.

Stan


Reply | Threaded
Open this post in threaded view
|

Re: History before 1999

jimb (Bugzilla)
On 1/9/06, Stan Shebs <[hidden email]> wrote:
> In theory, an energetic person could graft old public releases into
> the repository, making a sort of synthetic history. Easier in SVN
> perhaps?

Insertion of revisions into an existing history, even at the front,
isn't something SVN has any special support for.  I think you'd have
to dump, edit, and re-load the repository.

Reply | Threaded
Open this post in threaded view
|

Re: History before 1999

Daniel Jacobowitz-2
On Mon, Jan 09, 2006 at 12:33:20PM -0800, Jim Blandy wrote:
> On 1/9/06, Stan Shebs <[hidden email]> wrote:
> > In theory, an energetic person could graft old public releases into
> > the repository, making a sort of synthetic history. Easier in SVN
> > perhaps?
>
> Insertion of revisions into an existing history, even at the front,
> isn't something SVN has any special support for.  I think you'd have
> to dump, edit, and re-load the repository.

However, Ian has some nice hack for stitching CVS repositories together
during a CVS->SVN transition - he wrote them to add the old-gcc
repository to GCC.  I bet we could add the additional history in the
process of switching to SVN, if there's interest in doing that :-)

I've been thinking on and off about the "modules" problem that I
mentioned the last time the idea came up; I think it's not infeasible.

--
Daniel Jacobowitz
CodeSourcery

Reply | Threaded
Open this post in threaded view
|

Re: History before 1999

Ian Lance Taylor
Daniel Jacobowitz <[hidden email]> writes:

> On Mon, Jan 09, 2006 at 12:33:20PM -0800, Jim Blandy wrote:
> > On 1/9/06, Stan Shebs <[hidden email]> wrote:
> > > In theory, an energetic person could graft old public releases into
> > > the repository, making a sort of synthetic history. Easier in SVN
> > > perhaps?
> >
> > Insertion of revisions into an existing history, even at the front,
> > isn't something SVN has any special support for.  I think you'd have
> > to dump, edit, and re-load the repository.
>
> However, Ian has some nice hack for stitching CVS repositories together
> during a CVS->SVN transition - he wrote them to add the old-gcc
> repository to GCC.  I bet we could add the additional history in the
> process of switching to SVN, if there's interest in doing that :-)

Sure, that would work.

Ideally, somebody would go through the Cygnus repository and strip out
the redacted stuff.  Unfortunately that would have to be done by
somebody currently at Red Hat.  The simplest approach would be to just
check out each revision one by one, run redact and/or sanitize, and
check the redacted revision into a new repository.  Unfortunately that
would not be reliable, as some code which should have been redacted
was checked in without appropriate markers.  So somebody would have to
actually look at the revisions and make sure that everything was
properly sanitized.

It is perhaps worth noting that the old Cygnus contracts promised to
keep things secret until permitted, or until three years had passed.
If Red Hat kept that up, then in fact everything in the Cygnus
repository up to 1999 is now safe to reveal.

Ian