Bug 438860 – Clarify slot matching behavior for EAPI 5 ebuilds that have sub-slots

• Product: Gentoo Hosted Projects
• Status: RESOLVED NEEDINFO

[Description Zac Medico 2012-10-18 21:47:36 UTC]

This question came up on irc:

<lxnay> will matching foo/bar:0 always work when foo/bar has SLOT="0/1.0" ?
<ferringb> should
<lxnay> is this by spec or a gray area?
<ferringb> gray area
<lxnay> :(
<ferringb> ebuilds shouldn't be able to use that
<zmedico> maybe the spec needs clarification. it should work for EAPI 5 and earlier though
* ferringb conceded nothing of that sort
<zmedico> I thought obvious that foo/bar:0 should match, at least for EAPI 5 and earlier.

[Comment 1 Ciaran McCreesh 2012-10-18 22:09:19 UTC]

EAPIs that know nothing about subslots should be ignoring the subslot part of SLOT.

[Comment 2 Zac Medico 2012-10-18 22:12:41 UTC]

That foo/bar:0 atom should work the same in EAPI 5 as it does for earlier EAPIs "that know nothing about subslots", right?

[Comment 3 Ciaran McCreesh 2012-10-18 22:18:45 UTC]

That's the way I interpreted it, yeah. If a dep spec has a :slot without a subslot, it's matched against the slot (excluding the subslot, if it has one) of the package.

If a dep spec has a :slot/subslot, it's matched against the slot/subslot of the package. If a package has no explicit subslot (either because it's left out or because it's an older EAPI), its subslot is the same as its slot.

[Comment 4 Ulrich Müller 2018-08-08 14:18:45 UTC]

Anything to be done here? Feel free to reopen with a patch attached.