I know prepalldocs is supposed to be "banned"... But no replacement exists. So repoman should not fail on existing ebuilds that use this functionality. It would be more justifiable to emit a warning saying that prepalldocs is deprecated, and to be replaced. Until such a replacement is in place however, failing is not the right solution. It is annoying to have to use --force on such occassions...
Created attachment 184911 [details, diff] reduce prepalldocs usage to a warning
I'd say that this warning does no good at the present time. What are people supposed to do if they get it? IMO the right thing is to wait until a replacement is available (in EAPI 3?). However, I've noticed that prepalldocs was removed from several ebuilds. These should be changed to the new EAPI later, but it is likely that people will forget. Could the warning be disabled, until we have a replacement for prepalldocs?
When I originally included the prepalldocs check, I was under the impression that the plan was to eliminate prepalldocs usage from the tree. If that's not the case then the check seems pretty pointless.
I've removed the prepalldocs check in svn r13118.
(In reply to comment #3) > When I originally included the prepalldocs check, I was under the impression > that the plan was to eliminate prepalldocs usage from the tree. If that's not > the case then the check seems pretty pointless. Petteri had clarified it in <http://archives.gentoo.org/gentoo-dev-announce/msg_6a64b163b72227c919ef390dba789991.xml>: ,---- | To me it seems that based on summaries and other factors some developers | seem to have understood that prepalldocs should immediately be removed | from all ebuilds using it. When I voted on the issue it was my intention | to put the issue on the table so that a proper technical solution can be | achieved. If we just leave it there, it's most likely that nothing will | happen. So until we have a decision on what the replacement will be I don't | see a need to remove current prepalldocs usage but any new usage must be | avoided. `---- (In reply to comment #4) > I've removed the prepalldocs check in svn r13118. Thanks.
(In reply to comment #3) > When I originally included the prepalldocs check, I was under the impression > that the plan was to eliminate prepalldocs usage from the tree. If that's not > the case then the check seems pretty pointless. > All my checks before you have put in as just warnings so I assumed (perhaps falsely) that the same would happen to this one too.
The prepalldocs check is removed in 2.1.6.9 and 2.2_rc26. We can add it back again after there's an alternative available.
(In reply to comment #7) > The prepalldocs check is removed in 2.1.6.9 and 2.2_rc26. We can add it back > again after there's an alternative available. EAPI 4 has docompress, so I think it's about time to add the warning back. The commit that had removed it is here: <http://git.overlays.gentoo.org/gitweb/?p=proj/portage.git;a=commit;h=9da1ab2d6d894d784f9020688302743ce641fa30>
Created attachment 260796 [details, diff] Patch to re-add prealldocs warning There were some conflicts upon reverting the removal commit, which are resolved in attached patch.
We should probably wait a little while, at least until the council approves EAPI 4 in the tree, and maybe even until portage with EAPI 4 is marked stable.
This is fixed in 2.1.9.42 and 2.2.0_alpha26. It now checks for ecompress, ecompressdir, prepall, prepalldocs, and preplib: http://git.overlays.gentoo.org/gitweb/?p=proj/portage.git;a=commit;h=5aae42eb1cd319c19e92bce890ea7eb3b7d184d7 http://git.overlays.gentoo.org/gitweb/?p=proj/portage.git;a=commit;h=1ba916f6aa4c82b4e0f74955b9146b8cbf65b859