Summary: | version bump ebuild for nvu 0.6 | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | basic <basic> |
Component: | New packages | Assignee: | Mozilla Gentoo Team <mozilla> |
Status: | RESOLVED FIXED | ||
Severity: | enhancement | CC: | chriswhite |
Priority: | High | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Attachments: |
nvu-0.60.ebuild
nvu-0.60.ebuild diff diff v2 mozconfig.patch patch v3 patch v3-r1 |
Description
basic
2004-12-06 01:03:35 UTC
Created attachment 45369 [details]
nvu-0.60.ebuild
basic: please add a comment, if you provide an modified ebuild. also the header in invalid -> http://www.gentoo.org/proj/en/devrel/handbook/handbook.xml?part=2&chap=3 Created attachment 45385 [details]
nvu-0.60.ebuild
second try, fixed the header. Sorry, forgot to mention this:
This is a slightly modified ebuild from the 0.50 version.
I changed (aside from the header), the KEYWORDS definition, the 2 lines
starting with 'epatch' to use the patches for 0.50 (I checked that they apply
cleanly).
Created attachment 45386 [details, diff]
diff
oops, uploaded the wrong file! Here's a diff. with the header fixed (finally).
Created attachment 45488 [details, diff]
diff v2
with the previous ebuild, it seems to default to for --enable-optimize=-O .
That causes nvu to crash on startup on my system (gcc 3.3.4-r1 x86). In this
diff, I moved the patching and mozconfig stuff to src_unpack. Additionally, I
copied the logic for --enable-optimize from mozilla.eclass .
Created attachment 45489 [details, diff]
mozconfig.patch
Here's a patch to remove --enable-strip from mozconfig and let portage handle
stripping for easier debugging
Created attachment 45500 [details, diff]
patch v3
forgot that src_unpack does not default to the source dir (got part of it right
and part of it wrong :P). This adds a cd ${S} at the top.
Created attachment 45501 [details, diff]
patch v3-r1
oops uploaded the wrong file. This one should work (famous last words...)
Everything looked pretty much ok. Only comment I have to make is that when bumping ebuilds, if the previous ebuild (0.50 say) had: ~foo ~bar ~foobar 0.60 keeps the same ~arch keywords. However, if it had: foo bar ~foobar Stable would get downgraded to unstable in the next version, so: ~foo ~bar ~foobar and unstable would stay unstable. The only time that keyword dropping really occurs is: a) The next version uses some special library that some arches haven't marked b) The next version is INSANELY rebuilt c) It's a new ebuild, and someone submits it with a huge number of arches that they haven't tested These are really the only reasons we'd ever drop keywords. Otherwise, the basic policy applies. |