[MERGE] Refactor diffing

Aaron Bentley aaron.bentley at utoronto.ca
Tue Nov 27 06:23:58 GMT 2007


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

Martin Pool wrote:
>> Martin, do you want this in 1.0?
> 
> Yes, please merge this.  It looks much cleaner.

Will do.

> +    # The type or contents of the file were unsuitable for diffing
> +    CANNOT_DIFF = object()
> +    # The file has changed in a semantic way
> +    CHANGED = object()
> +    # The file content has changed, but there is no semantic change
> +    UNCHANGED = object()
> 
> (comment) I get the point of them, but it seems suboptimal they'll
> just be printed as <object at 1234123> or some such.  If we made them eg
> = intern('CANNOT_DIFF') do you think people would be too tempted to
> use them as strings not symbols?

I think we've been pretty good about that with e.g. TREE_ROOT and
NULL_REVISION.  I'm happy to make them plain strings, without interning
even.  I just thought perhaps this was a better way of ensuring the
constants had unique values, so I thought I'd try it.

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

iD8DBQFHS7f+0F+nu1YWqI0RApeZAJ0cisomvg6Gk79kJr9GdjZLVOAcGQCeOtBJ
+IDMh0zph3GRP5V0X6Qq1b0=
=RGq8
-----END PGP SIGNATURE-----



More information about the bazaar mailing list