1 |
Dnia September 11, 2019 11:11:15 PM UTC, William Hubbs <williamh@g.o> napisał(a): |
2 |
>On Wed, Sep 11, 2019 at 07:47:04PM +0000, Michał Górny wrote: |
3 |
>> Dnia September 11, 2019 7:40:41 PM UTC, William Hubbs |
4 |
><williamh@g.o> napisał(a): |
5 |
>> >On Wed, Sep 11, 2019 at 08:31:16PM +0200, Michał Górny wrote: |
6 |
>> >> On Wed, 2019-09-11 at 13:22 -0500, William Hubbs wrote: |
7 |
>> >> > On Wed, Sep 11, 2019 at 07:38:17PM +0200, Michał Górny wrote: |
8 |
>> >> > > On Wed, 2019-09-11 at 12:21 -0500, William Hubbs wrote: |
9 |
>> >> > > > Copyright: Sony Interactive Entertainment Inc. |
10 |
>> >> > > > Signed-off-by: William Hubbs <williamh@g.o> |
11 |
>> >> > > > --- |
12 |
>> >> > > > eclass/go-module.eclass | 76 |
13 |
>> >+++++++++++++++++++++++++++++++++++++++++ |
14 |
>> >> > > > 1 file changed, 76 insertions(+) |
15 |
>> >> > > > create mode 100644 eclass/go-module.eclass |
16 |
>> >> > > > |
17 |
>> >> > > > diff --git a/eclass/go-module.eclass |
18 |
>b/eclass/go-module.eclass |
19 |
>> >> > > > new file mode 100644 |
20 |
>> >> > > > index 00000000000..7009fcd3beb |
21 |
>> >> > > > --- /dev/null |
22 |
>> >> > > > +++ b/eclass/go-module.eclass |
23 |
>> >> > > > @@ -0,0 +1,76 @@ |
24 |
>> >> > > > +# Copyright 1999-2015 Gentoo Foundation |
25 |
>> >> > > |
26 |
>> >> > > You need to replace your calendar. And copyright holder. |
27 |
>> >> > |
28 |
>> >> > Sure, I thought I ffixed that. |
29 |
>> >> > |
30 |
>> >> > > > +# Distributed under the terms of the GNU General Public |
31 |
>> >License v2 |
32 |
>> >> > > > + |
33 |
>> >> > > > +# @ECLASS: go-module.eclass |
34 |
>> >> > > |
35 |
>> >> > > Any reason to change naming from golang-* to go-* now? |
36 |
>> >> > |
37 |
>> >> > Well, "lang" is sort of redundant, and there will be only one |
38 |
>> >eclass, so |
39 |
>> >> > I thought I would make things a bit more simple. |
40 |
>> >> > |
41 |
>> >> > > > +# @MAINTAINER: |
42 |
>> >> > > > +# William Hubbs <williamh@g.o> |
43 |
>> >> > > > +# @SUPPORTED_EAPIS: 7 |
44 |
>> >> > > > +# @BLURB: basic eclass for building software written in the |
45 |
>go |
46 |
>> >> > > > +# programming language that uses go modules. |
47 |
>> >> > > > +# @DESCRIPTION: |
48 |
>> >> > > > +# This eclass provides a convenience src_prepare() phase |
49 |
>and |
50 |
>> >some basic |
51 |
>> >> > > > +# settings needed for all software written in the go |
52 |
>> >programming |
53 |
>> >> > > > +# language that uses go modules. |
54 |
>> >> > > > +# |
55 |
>> >> > > > +# You will know the software you are packaging uses modules |
56 |
>> >because |
57 |
>> >> > > > +# it will have files named go.sum and go.mod in its |
58 |
>top-level |
59 |
>> >source |
60 |
>> >> > > > +# directory. If it does not have these files, use the |
61 |
>golang-* |
62 |
>> >eclasses. |
63 |
>> >> > > > +# |
64 |
>> >> > > > +# If the software you are packaging uses modules, the next |
65 |
>> >question is |
66 |
>> >> > > > +# whether it has a directory named "vendor" at the |
67 |
>top-level |
68 |
>> >of the source tree. |
69 |
>> >> > > > +# |
70 |
>> >> > > > +# If it doesn't, you need to create a tarball of what would |
71 |
>be |
72 |
>> >in the |
73 |
>> >> > > > +# vendor directory and mirror it locally. This is done with |
74 |
>> >the |
75 |
>> >> > > > +# following commands if upstream is using a git repository: |
76 |
>> >> > > > +# |
77 |
>> >> > > > +# @CODE: |
78 |
>> >> > > > +# |
79 |
>> >> > > > +# $ cd /my/clone/of/upstream |
80 |
>> >> > > > +# $ git checkout <release> |
81 |
>> >> > > > +# $ go mod vendor |
82 |
>> >> > > > +# $ tar cvf project-version-vendor.tar.gz vendor |
83 |
>> >> > > > +# |
84 |
>> >> > > > +# @CODE: |
85 |
>> >> > > > +# |
86 |
>> >> > > > +# Other than this, all you need to do is inherit this |
87 |
>eclass |
88 |
>> >then |
89 |
>> >> > > > +# make sure the exported src_prepare function is run. |
90 |
>> >> > > > + |
91 |
>> >> > > > +case ${EAPI:-0} in |
92 |
>> >> > > > + 7) ;; |
93 |
>> >> > > > + *) die "${ECLASS} API in EAPI ${EAPI} not yet |
94 |
>established." |
95 |
>> >> > > > +esac |
96 |
>> >> > > > + |
97 |
>> >> > > > +if [[ -z ${_GO_MODULE} ]]; then |
98 |
>> >> > > > + |
99 |
>> >> > > > +_GO_MODULE=1 |
100 |
>> >> > > > + |
101 |
>> >> > > > +BDEPEND=">=dev-lang/go-1.12" |
102 |
>> >> > > > + |
103 |
>> >> > > > +# Do not download dependencies from the internet |
104 |
>> >> > > > +# make build output verbose by default |
105 |
>> >> > > > +export GOFLAGS="-mod=vendor -v -x" |
106 |
>> >> > > > + |
107 |
>> >> > > > +# Do not complain about CFLAGS etc since go projects do not |
108 |
>> >use them. |
109 |
>> >> > > > +QA_FLAGS_IGNORED='.*' |
110 |
>> >> > > > + |
111 |
>> >> > > > +# Upstream does not support stripping go packages |
112 |
>> >> > > > +RESTRICT="strip" |
113 |
>> >> > > > + |
114 |
>> >> > > > +EXPORT_FUNCTIONS src_prepare |
115 |
>> >> > > |
116 |
>> >> > > Don't you need to inherit some other eclass to make it build? |
117 |
>> >> > |
118 |
>> >> > The primary reason for all of the golang-* eclasses was the |
119 |
>GOPATH |
120 |
>> >> > variable, which is not relevant when you are using modules. |
121 |
>> >> > |
122 |
>> >> > I can look at adding a src_compile to this eclass, but I haven't |
123 |
>> >thought |
124 |
>> >> > about what it would contain yet. |
125 |
>> >> > |
126 |
>> >> > > > + |
127 |
>> >> > > > +# @FUNCTION: go-module_src_prepare |
128 |
>> >> > > > +# @DESCRIPTION: |
129 |
>> >> > > > +# Run a default src_prepare then move our provided vendor |
130 |
>> >directory to |
131 |
>> >> > > > +# the appropriate spot if upstream doesn't provide a vendor |
132 |
>> >directory. |
133 |
>> >> > > > +go-module_src_prepare() { |
134 |
>> >> > > > + default |
135 |
>> >> > > > + # Use the upstream provided vendor directory if it exists. |
136 |
>> >> > > > + [[ -d vendor ]] && return |
137 |
>> >> > > > + # If we are not providing a mirror of a vendor directory |
138 |
>we |
139 |
>> >created |
140 |
>> >> > > > + # manually, return since there may be nothing to vendor. |
141 |
>> >> > > > + [[ ! -d ../vendor ]] && return |
142 |
>> >> > > > + # At this point, we know we are providing a vendor mirror. |
143 |
>> >> > > > + mv ../vendor . || die "Unable to move ../vendor directory" |
144 |
>> >> > > |
145 |
>> >> > > Wouldn't it be much simpler to create appropriate directory |
146 |
>> >structure |
147 |
>> >> > > in the tarball? Then you wouldn't need a new eclass at all. |
148 |
>> >> > |
149 |
>> >> > You would definitely need an eclass (see the settings and |
150 |
>> >dependencies). |
151 |
>> >> > |
152 |
>> >> > Take a look at the differences in the spire and hub ebuilds in |
153 |
>this |
154 |
>> >> > series. I'm not sure what you mean by adding the directory |
155 |
>> >structure to |
156 |
>> >> > the tarball? I guess you could add something to the vendor |
157 |
>tarball |
158 |
>> >when |
159 |
>> >> > you create it. |
160 |
>> >> |
161 |
>> >> I mean packing it as 'spire-1.2.3/vendor' or whatever the package |
162 |
>> >> directory is, so that it extracts correctly instead of making a |
163 |
>> >tarball |
164 |
>> >> that needs to be moved afterwards. |
165 |
>> > |
166 |
>> >That would clobber the upstream provided vendor directory and that's |
167 |
>> >what I want to avoid with the first test in src_prepare. |
168 |
>> |
169 |
>> If upstream already includes vendored modules, why would you create |
170 |
>your own tarball in the first place? |
171 |
> |
172 |
>You are right, and currently I quietly ignore your vendor tarball if |
173 |
>upstream |
174 |
>vendors the dependencies also. I could change this to generate a |
175 |
>warning |
176 |
>or die and force you to fix the ebuild, but that would not be possible |
177 |
>if I follow your suggestion because I would not be able to tell whether |
178 |
>the vendored dependencies came from us or upstream. |
179 |
|
180 |
Why would anyone create a vendor tarball if things work without it? That makes no sense. Also adding unused archives to SRC_URI is a QA violation. |
181 |
|
182 |
> |
183 |
>Also, another concern about your suggestion is the --transform switch |
184 |
>that would have to be added to the tar command people use to create |
185 |
>the |
186 |
>vendor tarball, something like: |
187 |
> |
188 |
>tar -acvf package-version-vendor.tar.gz |
189 |
>--transform='s#^vendor#package-version-vendor#' vendor |
190 |
> |
191 |
>You suggested that a maintainer could create a new tarball and build on |
192 |
>top of it. I guess you mean don't use upstream's tarball if they don't |
193 |
>vendor and create my own tarball and add the vendor directory to it. |
194 |
>I'm |
195 |
>against that option because I don't feel that we should manually |
196 |
>tinker |
197 |
>with upstream tarballs. That opens a pretty big can of worms imo. |
198 |
|
199 |
No. I suggested that rather than adding another git clone and checking out a tag (which sooner or later would mean someone forgetting and using master instead), you could unpack the same archive you're going to use in the ebuild. |
200 |
> |
201 |
>William |
202 |
|
203 |
|
204 |
-- |
205 |
Best regards, |
206 |
Michał Górny |