[PREVIEW] line-endings support

[Aaron Bentley]

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

Stephen J. Turnbull wrote:
> Aaron Bentley writes:
>  > Stefan Monnier wrote:
>  > >> Given source tree foo with branch bar, I can easily imagine foo and bar
>  > >> wanting different submission addresses and also wanting to merge one
>  > >> another.  That would make the submission address change unexpectedly and
>  > > 
>  > > That's a more general problem: how to keep 2 branches mutually sync'd
>  > > while preserving some differences between them.
>  > 
>  > I view branch configuration data as different from tree content, and see
>  > configuration as something that I never want to propagate when I merge.
> 
> I doubt that your view is anything like universal, though.  For
> example, I imagine that most people would consider .bzrignore to be
> "configuration", but also want changes to it to propagate.

You are conflating "branch configuration", which I was talking about,
with versioned tree metadata.  You can call that "configuration", but
- - It configures an entirely different object (the working tree rather
  than the branch.)
- - Its values must be versioned because they refer to versioned data.
- - Its values do not tend to vary from one site to another

> Stefan's view *is* more general

Do you have data to back this statement up?

> and the facility would be valuable

Which facility?

- - Partial synchronization of diverged branches à la tla update?
  Yes, I agree that it would be valuable.

- - EOL conversion?
  I prefer to use editors that don't trample on line ending conventions,
  But I agree that it is useful for many people.

- - Propagation of branch configuration data via working tree?
  I have examined our existing branch configuration settings and
  determined that all but one are definitely not suited to this means of
  propagation.  The remaning one is an open question.  I oppose it on
  this basis.

> so
> I hope you will consider addressing the problem on that basis rather
> than "content is content and config is config and ne'er the twain
> shall meet."

I think my argument is more nuanced than that.

Aaron
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFIBmJ50F+nu1YWqI0RAtYPAJ9hPp9MtQiCj3AGe7DU5fHN0MLg+QCghpmf
40nDg3NlA+kXUobE7YT8YHE=
=NypW
-----END PGP SIGNATURE-----