| 1 |
-----BEGIN PGP SIGNED MESSAGE----- |
| 2 |
Hash: SHA1 |
| 3 |
|
| 4 |
> The new eclass is supposedly used by 732 in-tree packages [1], |
| 5 |
> and possibly a few dozens out-of-the-tree. Some parts of the eclass API |
| 6 |
> are barely used. I would really prefer to avoid yet another eclass |
| 7 |
> migration. |
| 8 |
|
| 9 |
that's a shame, I don't think it is avoidable with the changes you are |
| 10 |
suggesting |
| 11 |
|
| 12 |
> However, I would like to do a few breaking changes to simplify |
| 13 |
> the eclass and its maintenance: |
| 14 |
> |
| 15 |
> 1. Make EGIT_SOURCEDIR default to ${WORKDIR}/${P} rather than ${S} |
| 16 |
> (to support setting S to subdirectory of git repo), |
| 17 |
> |
| 18 |
> 2. Kill EGIT_HAS_SUBMODULES and autodetect submodules, |
| 19 |
> |
| 20 |
> 3. Kill EGIT_OPTIONS since it limits the possibility of changing eclass |
| 21 |
> code, |
| 22 |
> |
| 23 |
> 4. Kill EGIT_MASTER (it's more trouble than benefit), |
| 24 |
> |
| 25 |
> 5. Possibly kill EGIT_PROJECT -- it won't be good enough anymore, |
| 26 |
> |
| 27 |
> 6. Kill EGIT_NONBARE and support bare checkouts only. Supporting two |
| 28 |
> different layouts introduces a lot redundant code. |
| 29 |
> |
| 30 |
> 7. Kill EGIT_NOUNPACK -- possibly replace it with proper fetching |
| 31 |
> function, or just src_fetch(), |
| 32 |
> |
| 33 |
> 8. Kill EGIT_DIR -- it supposedly should not even be overriden. |
| 34 |
> |
| 35 |
These changes will cause significant breakage. That is fine for |
| 36 |
migration, but not for in place eclass changes. |
| 37 |
> |
| 38 |
> But it's not all removing. There are also a few things I would like to |
| 39 |
> change/add: |
| 40 |
> |
| 41 |
> 1. Replace 'git clone' with 'git init' + 'git fetch' that would be |
| 42 |
> a bit simpler, |
| 43 |
> |
| 44 |
> 2. Add complete & working support for shallow clones, and make it the |
| 45 |
> default. This means a lot of space-saving for people who only use |
| 46 |
> the repos for ebuilds. |
| 47 |
> |
| 48 |
> 3. Add complete & proper support for submodules. Currently, submodules |
| 49 |
> enforce non-bare clones and are fetched during unpack. After |
| 50 |
> the change, they will be fetched and unpacked like normal repos. |
| 51 |
> |
| 52 |
> |
| 53 |
> The use of API features in *.ebuild maps like the following; |
| 54 |
> |
| 55 |
> EGIT_REPO_URI 521 |
| 56 |
> EGIT_BRANCH 66 |
| 57 |
> EGIT_SOURCEDIR 21 |
| 58 |
> EGIT_PROJECT 17 |
| 59 |
> EGIT_HAS_SUBMODULES 15 |
| 60 |
> EGIT_COMMIT 12 |
| 61 |
> EGIT_BOOTSTRAP 12 |
| 62 |
> EGIT_MASTER 7 |
| 63 |
> EGIT_NOUNPACK 2 |
| 64 |
> EGIT_STORE_DIR 1 |
| 65 |
> EGIT_NONBARE 1 |
| 66 |
> EGIT_DIR 1 |
| 67 |
> EVCS_OFFLINE 0 // these are for make.conf |
| 68 |
> EGIT_REPACK 0 |
| 69 |
> EGIT_PRUNE 0 |
| 70 |
> EGIT_OPTIONS 0 |
| 71 |
> |
| 72 |
> I will need to take a look which of those cases can be replaced easily. |
| 73 |
> |
| 74 |
> |
| 75 |
> How should I proceed? Assuming that git-2.eclass is used by live |
| 76 |
> ebuilds only, and those ebuilds can be subject to random breakage, |
| 77 |
> I could supposedly just start changing API of the eclass. |
| 78 |
> |
| 79 |
negative, breaking ebuilds is different than upstream breaking things |
| 80 |
and needing ebuild fixes. Randomly breaking 700+ ebuilds isn't cool. |
| 81 |
|
| 82 |
> On the other hand, I could also go for beautiful git-r1.eclass, |
| 83 |
> and cleanly switch the packages. |
| 84 |
> |
| 85 |
> Then, I could go for something involving the two -- create a new |
| 86 |
> git-r1.eclass that has API fully stripped, and start deprecating |
| 87 |
> features from git-2.eclass. We would be able to switch to git-r1 to |
| 88 |
> test offending packages safely, then do a big switch of remaining |
| 89 |
> packages and make the two eclasses temporarily equivalent. |
| 90 |
> |
| 91 |
> What are your thoughts? |
| 92 |
> |
| 93 |
I have a VCS ebuild for nearly everything I officially maintain and |
| 94 |
almost everything in the pentoo overlay. If you change the eclass in |
| 95 |
place that will break dozens of ebuilds just for me. Please don't. |
| 96 |
|
| 97 |
- -Zero |
| 98 |
-----BEGIN PGP SIGNATURE----- |
| 99 |
Version: GnuPG v2.0.20 (GNU/Linux) |
| 100 |
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ |
| 101 |
|
| 102 |
iQIcBAEBAgAGBQJSIYESAAoJEKXdFCfdEflKdXUQAI3Rwqyku4FCgO/3uivB/ltN |
| 103 |
+pDlw0DCHt9IlBK1Jh3t+ohXVdXhi6iZyQ7ly+t+5phA3zhR79yTJWkxmXq8E7br |
| 104 |
TM99Fsfd3Cqmdc54F4uA6MHI9MQj7efCkE3mas9bks8SPBnNTVwWmjVYUxBXZXfm |
| 105 |
J94+tcCEdesVqz2/iivm0iQ6W7EraCTG+umz/wz1urqIfDQBO8mDLHoM+jiovnUb |
| 106 |
tArOZ3GIhS/Rj+S/ZtH4VpvarFrH5ZzfO1GpNAvaFS8kGLRdEoUnMYk6K+WNdbkZ |
| 107 |
SYldaL8M/gNKWfmRU+j8OGK7bsNJx43Bqei7oUyMqkUVsTBpjmkPvw59aFKVlb7J |
| 108 |
kdfeoobpFsuAcxaQfWU1J8map82N8McdVOYMkEkC3HJP33TeymZKSUcU2/iMxr1W |
| 109 |
+kU0C2L7A9oXzWwkmiQ9WxAQ5KTYzyh5AzaaN45pju+QlFc2T4P4AZ4oqy+Zzzi2 |
| 110 |
CH2JZBPdv9+jMQLcwhpVoDsOPbbLGrx/aJEARwKvgs2fF+WYllraOu7uMPnaoYWw |
| 111 |
wNK9JYyhncx2pfG23PG7Uo3RtN8Aks0tptsHosOmB9ZArtnhYL2SxlotAoef+9X7 |
| 112 |
2qxFm59D9koW5FF0arnzzF/ul2zzos7NZRwJ1bwR1gYocxvN6Yqfg0zeX6uC+sDW |
| 113 |
dSukBOrVEuyftf+KurL9 |
| 114 |
=harh |
| 115 |
-----END PGP SIGNATURE----- |
Archives