Gentoo Archives: gentoo-dev

From: William Hubbs <williamh@g.o>
To: gentoo-dev@l.g.o
Cc: mgorny@g.o
Subject: Re: [gentoo-dev] [PATCH 1/3] go-module.eclass: introduce new eclass to handle go modules
Date: Wed, 11 Sep 2019 23:11:24
Message-Id: 20190911231115.GA20687@whubbs1.dev.av1.gaikai.org
In Reply to: Re: [gentoo-dev] [PATCH 1/3] go-module.eclass: introduce new eclass to handle go modules by "Michał Górny"
1 On Wed, Sep 11, 2019 at 07:47:04PM +0000, Michał Górny wrote:
2 > Dnia September 11, 2019 7:40:41 PM UTC, William Hubbs <williamh@g.o> napisał(a):
3 > >On Wed, Sep 11, 2019 at 08:31:16PM +0200, Michał Górny wrote:
4 > >> On Wed, 2019-09-11 at 13:22 -0500, William Hubbs wrote:
5 > >> > On Wed, Sep 11, 2019 at 07:38:17PM +0200, Michał Górny wrote:
6 > >> > > On Wed, 2019-09-11 at 12:21 -0500, William Hubbs wrote:
7 > >> > > > Copyright: Sony Interactive Entertainment Inc.
8 > >> > > > Signed-off-by: William Hubbs <williamh@g.o>
9 > >> > > > ---
10 > >> > > > eclass/go-module.eclass | 76
11 > >+++++++++++++++++++++++++++++++++++++++++
12 > >> > > > 1 file changed, 76 insertions(+)
13 > >> > > > create mode 100644 eclass/go-module.eclass
14 > >> > > >
15 > >> > > > diff --git a/eclass/go-module.eclass b/eclass/go-module.eclass
16 > >> > > > new file mode 100644
17 > >> > > > index 00000000000..7009fcd3beb
18 > >> > > > --- /dev/null
19 > >> > > > +++ b/eclass/go-module.eclass
20 > >> > > > @@ -0,0 +1,76 @@
21 > >> > > > +# Copyright 1999-2015 Gentoo Foundation
22 > >> > >
23 > >> > > You need to replace your calendar. And copyright holder.
24 > >> >
25 > >> > Sure, I thought I ffixed that.
26 > >> >
27 > >> > > > +# Distributed under the terms of the GNU General Public
28 > >License v2
29 > >> > > > +
30 > >> > > > +# @ECLASS: go-module.eclass
31 > >> > >
32 > >> > > Any reason to change naming from golang-* to go-* now?
33 > >> >
34 > >> > Well, "lang" is sort of redundant, and there will be only one
35 > >eclass, so
36 > >> > I thought I would make things a bit more simple.
37 > >> >
38 > >> > > > +# @MAINTAINER:
39 > >> > > > +# William Hubbs <williamh@g.o>
40 > >> > > > +# @SUPPORTED_EAPIS: 7
41 > >> > > > +# @BLURB: basic eclass for building software written in the go
42 > >> > > > +# programming language that uses go modules.
43 > >> > > > +# @DESCRIPTION:
44 > >> > > > +# This eclass provides a convenience src_prepare() phase and
45 > >some basic
46 > >> > > > +# settings needed for all software written in the go
47 > >programming
48 > >> > > > +# language that uses go modules.
49 > >> > > > +#
50 > >> > > > +# You will know the software you are packaging uses modules
51 > >because
52 > >> > > > +# it will have files named go.sum and go.mod in its top-level
53 > >source
54 > >> > > > +# directory. If it does not have these files, use the golang-*
55 > >eclasses.
56 > >> > > > +#
57 > >> > > > +# If the software you are packaging uses modules, the next
58 > >question is
59 > >> > > > +# whether it has a directory named "vendor" at the top-level
60 > >of the source tree.
61 > >> > > > +#
62 > >> > > > +# If it doesn't, you need to create a tarball of what would be
63 > >in the
64 > >> > > > +# vendor directory and mirror it locally. This is done with
65 > >the
66 > >> > > > +# following commands if upstream is using a git repository:
67 > >> > > > +#
68 > >> > > > +# @CODE:
69 > >> > > > +#
70 > >> > > > +# $ cd /my/clone/of/upstream
71 > >> > > > +# $ git checkout <release>
72 > >> > > > +# $ go mod vendor
73 > >> > > > +# $ tar cvf project-version-vendor.tar.gz vendor
74 > >> > > > +#
75 > >> > > > +# @CODE:
76 > >> > > > +#
77 > >> > > > +# Other than this, all you need to do is inherit this eclass
78 > >then
79 > >> > > > +# make sure the exported src_prepare function is run.
80 > >> > > > +
81 > >> > > > +case ${EAPI:-0} in
82 > >> > > > + 7) ;;
83 > >> > > > + *) die "${ECLASS} API in EAPI ${EAPI} not yet established."
84 > >> > > > +esac
85 > >> > > > +
86 > >> > > > +if [[ -z ${_GO_MODULE} ]]; then
87 > >> > > > +
88 > >> > > > +_GO_MODULE=1
89 > >> > > > +
90 > >> > > > +BDEPEND=">=dev-lang/go-1.12"
91 > >> > > > +
92 > >> > > > +# Do not download dependencies from the internet
93 > >> > > > +# make build output verbose by default
94 > >> > > > +export GOFLAGS="-mod=vendor -v -x"
95 > >> > > > +
96 > >> > > > +# Do not complain about CFLAGS etc since go projects do not
97 > >use them.
98 > >> > > > +QA_FLAGS_IGNORED='.*'
99 > >> > > > +
100 > >> > > > +# Upstream does not support stripping go packages
101 > >> > > > +RESTRICT="strip"
102 > >> > > > +
103 > >> > > > +EXPORT_FUNCTIONS src_prepare
104 > >> > >
105 > >> > > Don't you need to inherit some other eclass to make it build?
106 > >> >
107 > >> > The primary reason for all of the golang-* eclasses was the GOPATH
108 > >> > variable, which is not relevant when you are using modules.
109 > >> >
110 > >> > I can look at adding a src_compile to this eclass, but I haven't
111 > >thought
112 > >> > about what it would contain yet.
113 > >> >
114 > >> > > > +
115 > >> > > > +# @FUNCTION: go-module_src_prepare
116 > >> > > > +# @DESCRIPTION:
117 > >> > > > +# Run a default src_prepare then move our provided vendor
118 > >directory to
119 > >> > > > +# the appropriate spot if upstream doesn't provide a vendor
120 > >directory.
121 > >> > > > +go-module_src_prepare() {
122 > >> > > > + default
123 > >> > > > + # Use the upstream provided vendor directory if it exists.
124 > >> > > > + [[ -d vendor ]] && return
125 > >> > > > + # If we are not providing a mirror of a vendor directory we
126 > >created
127 > >> > > > + # manually, return since there may be nothing to vendor.
128 > >> > > > + [[ ! -d ../vendor ]] && return
129 > >> > > > + # At this point, we know we are providing a vendor mirror.
130 > >> > > > + mv ../vendor . || die "Unable to move ../vendor directory"
131 > >> > >
132 > >> > > Wouldn't it be much simpler to create appropriate directory
133 > >structure
134 > >> > > in the tarball? Then you wouldn't need a new eclass at all.
135 > >> >
136 > >> > You would definitely need an eclass (see the settings and
137 > >dependencies).
138 > >> >
139 > >> > Take a look at the differences in the spire and hub ebuilds in this
140 > >> > series. I'm not sure what you mean by adding the directory
141 > >structure to
142 > >> > the tarball? I guess you could add something to the vendor tarball
143 > >when
144 > >> > you create it.
145 > >>
146 > >> I mean packing it as 'spire-1.2.3/vendor' or whatever the package
147 > >> directory is, so that it extracts correctly instead of making a
148 > >tarball
149 > >> that needs to be moved afterwards.
150 > >
151 > >That would clobber the upstream provided vendor directory and that's
152 > >what I want to avoid with the first test in src_prepare.
153 >
154 > If upstream already includes vendored modules, why would you create your own tarball in the first place?
155
156 You are right, and currently I quietly ignore your vendor tarball if upstream
157 vendors the dependencies also. I could change this to generate a warning
158 or die and force you to fix the ebuild, but that would not be possible
159 if I follow your suggestion because I would not be able to tell whether
160 the vendored dependencies came from us or upstream.
161
162 Also, another concern about your suggestion is the --transform switch
163 that would have to be added to the tar command people use to create the
164 vendor tarball, something like:
165
166 tar -acvf package-version-vendor.tar.gz --transform='s#^vendor#package-version-vendor#' vendor
167
168 You suggested that a maintainer could create a new tarball and build on
169 top of it. I guess you mean don't use upstream's tarball if they don't
170 vendor and create my own tarball and add the vendor directory to it. I'm
171 against that option because I don't feel that we should manually tinker
172 with upstream tarballs. That opens a pretty big can of worms imo.
173
174 William

Attachments

File name MIME type
signature.asc application/pgp-signature

Replies