On 09/29/2015 05:31 PM, Alexis Ballier wrote: > On Sat, 19 Sep 2015 23:04:14 +0200 > hasufell <hasuf...@gentoo.org> wrote: > >> 2. slowly start migrating those ~550 packages with "libressl" USE flag >> which is similar to gnutls USE flag. >> There will be no virtual, because those don't give sufficient control >> (libressl and openssl are not ABI compatible). > > If API compatibility is guaranteed, a virtual makes more sense than a > useflag. However, even with this assumption, it's not possible these > days, because we need to: > - Fix || ( a:= b:= ) deps handling in package managers (hey dynamic > deps :) ) > - Create transitive := deps (useful for other things than virtuals, hey > ocaml & haskell). > > > So yeah, because of lack of better solution this seems to be the best > method so far. >
Even if we fix the problem of SUBSLOTs in virtuals, we still have the situation that makes for example virtual/ffmpeg insufficient for a lot of use cases: even the API is not always strictly compatible, so ebuild writers sometimes need more fine-grained control over the version strings. This problem is very likely to arise with libressl as well. They want to be API compatible, but: * sometimes drop functions that they think have security flaws that can't be reasonably fixed * have their own set of functions on top of the openssl API Because of the last point, it is also possible that packages start to rely on libressl-specific API and not just on the common part. So a virtual doesn't seem to make sense, no matter how you look at it, IMO.
