Re: [PATCH v3 2/5] selinux-policy-2.eclass: respect the sanctity of -rN for EAPI=8
"Rahul Sandhu" <[email protected]>
| Newsgroups | gmane.linux.gentoo.devel |
|---|---|
| Message-ID | <[email protected]> |
On Wed Apr 8, 2026 at 6:08 AM BST, Sam James wrote: > Ulrich Müller <[email protected]> writes: > >>>>>>> On Tue, 07 Apr 2026, Rahul Sandhu wrote: >> >>> +if [[ ${EAPI} = 7 ]]; then >>> + # @ECLASS_VARIABLE: BASEPOL >>> + # @DESCRIPTION: >>> + # This variable contains the version string of the selinux-base-policy package >>> + # that this module build depends on. It is used to patch with the appropriate >>> + # patch bundle(s) that are part of selinux-base-policy. >>> + : "${BASEPOL:="${PVR}"}" >>> +else >>> + # @ECLASS_VARIABLE: BASEPOL >>> + # @DESCRIPTION: >>> + # This variable contains the version string of the selinux-base-policy package >>> + # that this module build depends on. It is used to patch with the appropriate >>> + # patch bundle(s) that are part of selinux-base-policy. >>> + # _pN corresponds to -rN in the patch bundle(s). >>> + : "${BASEPOL:="${PV/_p/-r}"}" >>> +fi >> >> This won't work. eclass-to-manpage expects eclass documentation to be >> left-aligned. Anything that is indented will be ignored. >> >> You also cannot have a conditional there, i.e. if both comments were >> without indentation then eclass-to-manpage would see a duplicate >> definition and would terminate with an error. > > Yes. Just change it so that it's one description where the text mentions > different behaviour for EAPI 7 vs other values. > > That's better anyway even if you ignore alignment or processing > problems -- what are the semantics supposed to be for a duplicate > eclassdoc? :) Hopefully should be sorted in the v4 series!