| 1 |
GLEP: 3
|
| 2 |
Title: Ebuild maintainter extension GLEP
|
| 3 |
Version: $Revision: 1.1 $
|
| 4 |
Last-Modified: $Date: 2003/06/10 17:31:01 $
|
| 5 |
Author: Caleb Tennis <caleb@gentoo.org>
|
| 6 |
Status: Deferred
|
| 7 |
Type: Standards Track
|
| 8 |
Content-Type: text/x-rst
|
| 9 |
Created: 09-Jun-2003
|
| 10 |
Post-History: 10-Jun-2003
|
| 11 |
|
| 12 |
|
| 13 |
Abstract
|
| 14 |
========
|
| 15 |
|
| 16 |
Gentoo's portage tree attempts to provide a self contained, easy to use, and
|
| 17 |
automatic installation procedure for as many packages as can be maintained by
|
| 18 |
developers.
|
| 19 |
|
| 20 |
This GLEP proposes allowing non-core Gentoo developers to be considered as
|
| 21 |
ebuild maintainers sponsored via a core Gentoo developer. This system will
|
| 22 |
allow more ebuilds to be available in portage with active maintainers for
|
| 23 |
those ebuilds.
|
| 24 |
|
| 25 |
This GLEP only applies to EBUILD based bugs that contain a request for a
|
| 26 |
package to be committed or version bumped within portage.
|
| 27 |
|
| 28 |
Motivation
|
| 29 |
==========
|
| 30 |
|
| 31 |
As of the first draft of this GLEP, there are 1387 EBUILD bug requests in
|
| 32 |
Gentoo's bugzilla database. Many of these requests contain ebuilds that
|
| 33 |
have been submitted by the bug reporter and are simply awaiting a Gentoo
|
| 34 |
developer to sponsor the submittal of the ebuild.
|
| 35 |
|
| 36 |
|
| 37 |
|
| 38 |
Rationale
|
| 39 |
=========
|
| 40 |
|
| 41 |
Gentoo's portage tree already contains the most popular ebuilds for packages
|
| 42 |
available today. Many teams exist that are responsible for maintaining these
|
| 43 |
core ebuilds in the portage tree. But, for ebuilds that are not as commonly
|
| 44 |
used, there is no good focal point upon which to rest these ebuilds.
|
| 45 |
|
| 46 |
For example, any submitted ebuild that is a KDE application gets routed to the
|
| 47 |
KDE team. However, the KDE team may be unfamiliar with the submitted ebuild.
|
| 48 |
A new graphical MySQL editor may be submitted to the MYSQL team, but none of
|
| 49 |
the members of that team may be familiar or have the desire to learn a new
|
| 50 |
program to submit it to portage.
|
| 51 |
|
| 52 |
We want to be able to provide for as many ebuilds in portage as feasible and
|
| 53 |
make sure that all ebuilds have some person who is responsible for
|
| 54 |
maintenance.
|
| 55 |
|
| 56 |
|
| 57 |
Backwards Compatibility
|
| 58 |
=======================
|
| 59 |
|
| 60 |
No current policies exist that interfere with this document.
|
| 61 |
|
| 62 |
|
| 63 |
Implementation
|
| 64 |
==============
|
| 65 |
|
| 66 |
Incoming ebuild bug reports should continue to be processed as normal.
|
| 67 |
|
| 68 |
Bug reports that *do not* contain an attached ebuild should be marked as
|
| 69 |
NEEDINFO. A message asking the user to create and submit an ebuild should be
|
| 70 |
attached to the bug.
|
| 71 |
|
| 72 |
Bug reports that *do* have an attached ebuild should be responded to with
|
| 73 |
a message asking if the reporter agrees to provide maintence and support for
|
| 74 |
the ebuild and package.
|
| 75 |
|
| 76 |
If a reporter *does not* agree to provide package maintence, the bug report
|
| 77 |
should be marked WONTFIX.
|
| 78 |
|
| 79 |
If a reporter *does* agree to provide package support, the ebuild should
|
| 80 |
be added to portage with a note in the ChangeLog that the reporter is
|
| 81 |
considered the maintainer of that particular ebuild.
|
| 82 |
|
| 83 |
Any incoming bug reports that are related to this ebuild should continue to
|
| 84 |
get processed as normal. The team that the ebuild goes to should then CC the
|
| 85 |
author of the ebuild. Optionally, if a docs-team member has prior knowledge
|
| 86 |
that the ebuild is externally maintained, he/she can add that person to the CC
|
| 87 |
list.
|
| 88 |
|
| 89 |
Security
|
| 90 |
========
|
| 91 |
|
| 92 |
**At the very least**, all ebuilds must be looked over by the developer
|
| 93 |
handling the commit.
|
| 94 |
|
| 95 |
In no case should a submitted digest file be used. The developer is
|
| 96 |
responsible for creating the digest file based on an actual download of the
|
| 97 |
source code.
|
| 98 |
|
| 99 |
Potential breaches in security can still exist, however. The developer
|
| 100 |
handling the installation should take every step to ensure that no ebuild,
|
| 101 |
package, or other files have been compromised.
|
| 102 |
|
| 103 |
|
| 104 |
Future
|
| 105 |
======
|
| 106 |
|
| 107 |
Current proposals to rethink Gentoo portage and bug handling (a.k.a Herds) are
|
| 108 |
still in negotiation. It is the intention of the author of this GLEP to evolve
|
| 109 |
the concept of this GLEP as the Herds concept matures and stabilizes.
|
| 110 |
|
| 111 |
|
| 112 |
References
|
| 113 |
==========
|
| 114 |
|
| 115 |
.. [#GLEP2] GLEP 2, Sample ReStructuredText GLEP Template, Goodyear,
|
| 116 |
(http://glep.gentoo.org/glep-0002.html)
|
| 117 |
|
| 118 |
|
| 119 |
Copyright
|
| 120 |
=========
|
| 121 |
|
| 122 |
This document has been placed in the public domain.
|
| 123 |
|