it would be nice to have the constraints sorted by type and name in the table scripts (this one that are checked in to SVN and are used to show me the differences)

Example: I have a table with 11 FKs which I will see as
ALTER TABLE [pld].[tvertraege] ADD CONSTRAINT [tblfk4] FOREIGN KEY ...
ALTER TABLE [pld].[tvertraege] ADD CONSTRAINT [tblfk3] FOREIGN KEY ...
ALTER TABLE [pld].[tvertraege] ADD CONSTRAINT [tblfk6] FOREIGN KEY ...
ALTER TABLE [pld].[tvertraege] ADD CONSTRAINT [tblfk7] FOREIGN KEY ...
ALTER TABLE [pld].[tvertraege] ADD CONSTRAINT [tblfk5] FOREIGN KEY ...
ALTER TABLE [pld].[tvertraege] ADD CONSTRAINT [tblfk] FOREIGN KEY ...
ALTER TABLE [pld].[tvertraege] ADD CONSTRAINT [tblfk2] FOREIGN KEY ...
ALTER TABLE [pld].[tvertraege] ADD CONSTRAINT [tblfk9] FOREIGN KEY ...
ALTER TABLE [pld].[tvertraege] ADD CONSTRAINT [tblfk8] FOREIGN KEY ...
ALTER TABLE [pld].[tvertraege] ADD CONSTRAINT [tblfk11] FOREIGN KEY ...
ALTER TABLE [pld].[tvertraege] ADD CONSTRAINT [tblfk10] FOREIGN KEY ...

in the compare window. If (as in this example) the constraints are nummered it will just about possible to "reorder" them (if their number is limited) but some indexes / fk's have other table / column names (which may sometimes only slightly different) as suffix which makes the (unneccessary) reordering much harder.

The same applies to indexes.