Re: Strategy for tracking 3rd party source.

[Mark D. Baushke]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

address@hidden writes:

> I need to track some 3rd party source code, in this case the Linux
> kernel and then manage our local modifications to the kernel (don't
> worry, GPL is being respected).  I've run into an issue where an
> import updates files on the HEAD and I'm not sure if I want this to
> happen.
> 
> For example, I import a kernel with the vendor tag of KERNEL_ORG and a
> release tag of LINUX_2_6_12.  I add my local changes and commit them.
> All is well.
> 
> There are 2 projects that pull the HEAD from this module into their
> projects, this works pretty well.  The problem arises when I import a
> new version of the kernel, say LINUX_2_6_13.  Its possible that it
> could be a while (couple of days) before HEAD is working again - not
> an ideal situation for the projects that depend on this module.
> 
> What I am thinking of doing is reserving HEAD for the 3rd party source
> and then making a branch, MYBRANCH.  The other projects can then pull
> from MYBRANCH (our internal HEAD so to speak), I can then do the
> import, merge the changes into MYBRANCH and commit the working code
> into MYBRANCH after its working.  That way MYBRANCH is always usable.
> 
> Is this a reasonable approach?  

Your outlined method should work.

> Is there a better way?

I believe you want the 'cvs import -X' feature provided by Chris
Demetriou or you want to use his 'cvsimport_killnew' script.

Read http://lists.gnu.org/archive/html/bug-cvs/2004-06/msg00173.html

This feature was added to cvs 1.12.10. The latest CVS release is cvs
1.12.12.

        -- Mark
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (FreeBSD)

iD8DBQFDFVRfCg7APGsDnFERAvslAKD2HF+LS1G1+omohoRCKZ64qLvUCgCg0CxB
bdph92h+d0GvKYHEOPxWqgo=
=RXEz
-----END PGP SIGNATURE-----