Kevin Kofler <kevin.kofler@xxxxxxxxx> wrote: > I do not see why we need to patch each and every application for this. Since > the library is at least API-compatible (not sure about the ABI), libidn2- > devel should have Obsoletes and Provides for libidn-devel and symlinks > libidn.so → libidn2.so and idn.h → idn2.h. Then all that is needed in > application packages is a rebuild. Changing one package scales much better > than changing all other packages. Obviously Simon Josefsson must have had a compelling reason to make a new and parallel-installable library instead of just releasing a new version of LibIDN. The first and most obvious reason that comes to mind is that applications should be able to rely on the libraries they use. A version upgrade of a library shouldn't change its functionality in incompatible ways. IDNA2008 isn't fully compatible with IDNA2003, and therefore the switch should be done only after application authors have assessed the impact on their particular application. Parallel-installable libraries allow them to do this and switch over one application at a time. What you're proposing would enforce the change, causing applications to suddenly map some domains differently before upstream authors are prepared, as if LibIDN2 had been released as simply a new version of LibIDN. I know it's very popular to break functionality left and right, but there are some of us weirdos who actually like software that works consistently and reliably. The existence of LibIDN2 indicates that Simon Josefsson is a rare responsible programmer who wants application authors to be able to rely on his libraries. I ask the breakage-loving majority to please respect his decision, and not force their own preferences on his work. Björn Persson
Attachment:
pgpKwv5jUwXdM.pgp
Description: OpenPGP digital signatur
_______________________________________________ devel mailing list -- devel@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe send an email to devel-leave@xxxxxxxxxxxxxxxxxxxxxxx
