Re: bzr repository ready?

[Stephen J. Turnbull]

Óscar Fuentes writes:

 > Stephen missed a word here.

If you like.  As I understand it, however, the *idea* of a checkout is
to depend on the parent branch for history and metadata, and the
"heavyweight" checkout was actually sort of an implementation
accident: it was simplest to start by implementing checkouts as a
command which automatically bound the new branch to its parent, and
the "ideal" (lightweight) checkout was implemented later.

 > What he says is correct for a *lightweight* checkout. A normal
 > checkout contains all the metadata, and in essence

Please refer to the recent address@hidden archives.  "Checkout"
in discussion there now means "lightweight checkout".  Ian Clatworthy
is ready to flip to lightweight by default at any time, and he even
wants bound branches to disappear entirely AFAICT.  Aaron Bentley and
Barry Warsaw talk about the "checkouts" used to implement the pipeline
plugin, etc, which are definitely lightweight.  And so on.

It is true that as of v2.0.1, 'bzr help' still refers to "normal
checkouts" (== bound branches) and --lightweight is required to get a
lightweight checkout, but it is clear that the trend is bipolar:
(unbound) branches for decentralized workflows, and (lightweight)
checkouts for centralized workflows and special applications (like
"build-only").  I think it's better to follow the modern terminology
here on emacs-devel.