Re: [Clirr-devel] detect source compatibility problems (was: comments on recent commits)

[Simon K. <si...@ec...>]

On Wed, 2004-06-23 at 06:33, Lars K=FChne wrote:
> Simon Kitching wrote:
>=20
> >>On a related note, I have recently found out that messing around with=
 a=20
> >>method's declared exceptions will not result in binary compatibility=20
> >>problems (surprising to me). Should we extend clirr's scope to check =
for=20
> >>source compatibility as well?
> >>   =20
> >>
> >
> >I'm +1 on reporting compile-time incompatibilities too.=20
> >
> >I agree that changing the declared exceptions should be reported; a
> >WARNING seems most appropriate to me. What do you think?
> >
> > =20
> >
>=20
> My idea currently is to introduce two severity levels in ApiDeffierence=
,=20
> one for binary compatibility and one for source compatibility. For most=
=20
> differences we detect today they would have the same value.
>=20
> For added checked exceptions binary compatibility severity would be=20
> WARNING (because strange things will happen if the exception is actuall=
y=20
> thrown and the client does not handle it), for removed checked=20
> exceptions it would be INFO. In both cases source compatibility would b=
e=20
> ERROR.
>=20
> This would have some effect on the Ant task, e.g. it should be possible=
=20
> to break the build only on binary errors. I'm currently working on that.
>=20
> You like the idea?

That sounds good to me.