On Tue, Nov 11, 2008 at 04:36:27PM +0100, Lele Gaifax <[EMAIL PROTECTED]> wrote: > > With fast-import, this is not an issue, as all the darcs backend does > > is to say "hey, a new patch is applied, here is the new result, take > > this as a commit", and this is a copletely valid behaviour. The easy > > part is that the importer will 1) just ignore the "file modification" > > if the file is unchanged and 2) will automatically detect renames / > > copies, so you don't have to care about it too much. > > Interesting but... does that cover corner cases like > > $ darcs mv a b > $ darcs mv b c > > where older darcses (or other VCS didn't collapse) that into a single > "mv a c"?
That will be collapsed to a single rename, because git/fast-import will just store the new state (it's called a 'tree') then generate the diff / detect renames on the fly if you request so. > Or > > $ svn mv a b > $ touch a > $ svn add a > > that is a "replacement" in svn parlance? I think git/fast-import detects that as well, but it isn't interesting for the conversion, either, since it's all about "here is the new state, store it". More here: http://wincent.com/a/about/wincent/weblog/archives/2007/11/one_of_the_thin.php
pgpZ12lxn4HzJ.pgp
Description: PGP signature
_______________________________________________ darcs-users mailing list [email protected] http://lists.osuosl.org/mailman/listinfo/darcs-users
