Re: Serialized Changeset

Tez Kamihira <tez@kamihira.com> writes:
> Here's my second subject to this community.
>
> I tried to define a new changeset format:
>
> http://scm.bluegate.org/scs.txt

Looks like a worthwhile effort!

You don't explicitly say that an SCS file should be acceptable to the
'patch' program -- i.e., today's 'patch' would just ignore anything it
didn't know about. I assume this is a goal?

Obviously patching programs can be extended; my question is about the
current install base of today's patch program.

> and have some questions.
>
> a.) Does your system have any sort of changeset ?
> (according to the book, it'd be "yes".)

Yes. In Subversion, a "revision" or a "commit" is a changeset. It is
the difference from one versioned tree to the next. A commit may
include:

   - Textual modifications to files, symbolic links,
   - Copies of files, symbolic links, and directories
   - Additions of files, symbolic links, and directories
   - Deletions of files, symbolic links, and directories
   - Renames of files, symbolic links, and directories
   - Changes to medatada (properties)
   - Deletions of metadata
   - Additions of metadata

We typically refers to a changeset by its revision number, e.g., "r14861".

> b.) Do you think this specification has enough ability to describe
> your system's changeset completely ?

It looks like it can, yes.

If there were a shorter, informal summary of the format, it would be
easier to know for sure. Your document jumps straight to the formal
specification. The formal specification is very important, as a
reference for implementation, but it's a bit unwieldy for someone who
just wants an overview.

Best,
-Karl Fogel

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Fri May 27 19:38:29 2005