On 05/08/13 18:06, Samuli Suominen wrote:
On 05/08/13 18:00, Michael Palimaka wrote:
On 5/08/2013 21:58, Samuli Suominen wrote:
On 05/08/13 13:56, Michael Palimaka wrote:
On 5/08/2013 19:33, Samuli Suominen wrote:
This is a friendly reminder.
I've found the tree again to have dependencies like:
dev-libs/openssl:=
virtual/jpeg:=
Is there any reason for the subslot operator being specified at all? I
don't see those packages defining any subslots.
You don't need to see it, because portage sets implicit subslot /0 in
EAPI="5" so it's there, even if you don't see it.
Then when it's changed (= set), the effect is same.
- Samuli
Even though the subslot is implicit, is that any reason to still use the
operator? We don't know what the maintainer's future intentions for the
subslot will be.
For example, we caused many useless rebuilds with poppler because
dependants added the subslot operator without consideration (many
packages link only against one of the stable interface libraries, rather
than the main library for which the subslot was intended).
The plan is to change SLOT of virtual/jpeg from "0" to eg. "0/1" after
next SONAME change in the default of the virtual, so it's useful to have
everything depend on virtual/jpeg:0= ready, to get the benefits of the
subslot.
This is different than poppler, this is just single shared library
package, like eg. libpng, tiff, lcms, ... where as poppler has many
different backends and own libraries for them.
[ .. ]
okay, maybe this plan sucks as some have suggested in later posts in
this thread.
however the main point from first post stands, don't at least do
virtual/jpeg:= deps, use at least virtual/jpeg:0 or virtual/jpeg:0= whatever
