| 1 |
-----BEGIN PGP SIGNED MESSAGE----- |
| 2 |
Hash: SHA1 |
| 3 |
|
| 4 |
Hi. |
| 5 |
|
| 6 |
As decided in the last council's meeting, following the recent |
| 7 |
discussions about .la files removal, I'm sending an email to this list |
| 8 |
with a proposal for a plan to address this issue. |
| 9 |
|
| 10 |
My proposal for the plan is the following: |
| 11 |
|
| 12 |
1. Add a function to eutils to deal with the removal of the .la files. |
| 13 |
|
| 14 |
delete_libtool_archives() { find "${@:$D}" -name '*.la' -delete } |
| 15 |
|
| 16 |
That function was suggested by Diego, but Arfrever has argued that we |
| 17 |
should replace : with - as '"${@:$D}" expands to a subarray containing |
| 18 |
elements starting with element with index $D (where element 0 is $0)'. |
| 19 |
The point in having this function in eutils is to ensure we use a |
| 20 |
consistent way to address the .la files. This will also make it much |
| 21 |
easier to adapt or review this function if needed. |
| 22 |
|
| 23 |
2. Get a consensus about the wording of when it's appropriate and how |
| 24 |
it's appropriate to remove the .la files |
| 25 |
|
| 26 |
As agreed in the meeting, as a first draft, we have that "the motion is |
| 27 |
to drop la files, when appropriate, through the use of a function in |
| 28 |
eutils that will only be called if the static-libs use flag is not set |
| 29 |
or unless the package relies on pkg-config". |
| 30 |
In the meantime there were some concerns raised about some prefix arches |
| 31 |
and therefore there's a suggestion we use a hidden variable (not to be |
| 32 |
set by users) to control the removal of the files so that we can "mask" |
| 33 |
running the function on any profile where we can't drop the .la files. |
| 34 |
|
| 35 |
3. Add a page to the QA project space (unless they're not interested) |
| 36 |
about la files and how to deal with them |
| 37 |
|
| 38 |
I think we can ask Diego to use large parts of his blog posts and charts |
| 39 |
about .la files[1], [2], [3], [4] as a source for the document. We can |
| 40 |
also add some basic info from the autobook[5], [6]. The goal would be to |
| 41 |
have a document similar to our own --as-needed guide [7] and to other |
| 42 |
distributions[8], [9]. |
| 43 |
Anyone wishes to volunteer for this task? |
| 44 |
QA would such a document be welcomed in your project space? |
| 45 |
|
| 46 |
4. Add the news item about pending .la files removal, pointing to 3 and |
| 47 |
with instructions to run lafilefixer and explaining how portage-2.1.9* |
| 48 |
addresses this issue. |
| 49 |
|
| 50 |
We can base the news item in the existing proposal for this[10]. |
| 51 |
|
| 52 |
5. Get portage-2.1.9* marked stable. |
| 53 |
|
| 54 |
As this portage version will fix the contents of the .la files when |
| 55 |
installing them, it will prevent the trouble caused by the hard-coded |
| 56 |
lib references - for all the .la files that will still need to be installed. |
| 57 |
One issue that remains open to me, is whether there will be any control |
| 58 |
over this feature. I was told the plan is to have it dependent on |
| 59 |
FEATURES, which to me doesn't make much sense. |
| 60 |
|
| 61 |
6. allow maintainers to work on .la files removal as they see fit |
| 62 |
|
| 63 |
After we provide maintainers consistent tools to remove .la files, |
| 64 |
create some good documentation about this, get a stable version of |
| 65 |
portage that helps addressing the issue for newly installed packages |
| 66 |
and warn users, maintainers are free to work on this as they can. |
| 67 |
|
| 68 |
|
| 69 |
[1] - http://blog.flameeyes.eu/2008/04/14/whatabout-those-la-files |
| 70 |
[2] - |
| 71 |
http://blog.flameeyes.eu/2008/07/02/again-about-la-files-or-why-should-they-be-killed-off-sooner-rather-than-later |
| 72 |
[3] - |
| 73 |
http://blog.flameeyes.eu/2009/07/06/identifying-pointless-la-files-for-plugins |
| 74 |
[4] - |
| 75 |
http://blog.flameeyes.eu/2009/09/28/removing-la-files-for-dum-w-uncertain-people |
| 76 |
[5] - http://sources.redhat.com/autobook/autobook/autobook_11.html#SEC11 |
| 77 |
[6] - http://sources.redhat.com/autobook/autobook/autobook_68.html#SEC68 |
| 78 |
[7] - http://www.gentoo.org/proj/en/qa/asneeded.xml |
| 79 |
[8] - http://wiki.mandriva.com/en/Libtool_archives#shared_build |
| 80 |
[9] - http://wiki.mandriva.com/en/Overlinking |
| 81 |
[10] - |
| 82 |
http://archives.gentoo.org/gentoo-dev/msg_c65fd55a78f1418178689df822a25638.xml |
| 83 |
|
| 84 |
|
| 85 |
Please reply if you have any suggestions or warnings about this plan and |
| 86 |
or any of the points. |
| 87 |
I'm going to send one email for each point in this plan. Please send any |
| 88 |
replies about the plan to this thread and to each of the points to the |
| 89 |
respective threads. |
| 90 |
|
| 91 |
- -- |
| 92 |
Regards, |
| 93 |
|
| 94 |
Jorge Vicetto (jmbsvicetto) - jmbsvicetto at gentoo dot org |
| 95 |
Gentoo- forums / Userrel / Devrel / KDE / Elections / RelEng |
| 96 |
-----BEGIN PGP SIGNATURE----- |
| 97 |
Version: GnuPG v2.0.16 (GNU/Linux) |
| 98 |
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ |
| 99 |
|
| 100 |
iQIcBAEBAgAGBQJMzO/BAAoJEC8ZTXQF1qEPIdkP/j7H/y0qQVJe3LlTwuePYuWU |
| 101 |
4OuUWvPDKrf+ymZpUxDFCbsRt9AOEl+IStknaNWSJe82oJ6L1N/RUWSwzS6QN8gF |
| 102 |
dvF//pWO4VMJVYgYZMnvq/dww3kOXPRvcQP/qPeBrfto2VZ6v+W2FpFJOKCS+LVj |
| 103 |
JUF0Q9WhJencdLNq9TwGZLfUKFOg0w3Yf75ru7EvgGTaJEONIIA5/wIEI61oNOM0 |
| 104 |
NJGLusM/SM3UhMN4mnOhg3pl93XKrF960vY46mDkH9j/I32qbMtu7zH79JjZWXGy |
| 105 |
F9wU66BbglCVl4FcvCQyvuli7yXZA1rhC9QEZ0yVvxP6HMPYGevH/8uav1M4AXMl |
| 106 |
KMbKsUqch+bpd2MwMnDS7Z0eyvKYj417CNJK1kZQcKTW95AXw5fH7RN8WkXQK7Iy |
| 107 |
C8mDd+INLUboTqVsk5vjYzfz96lm1LwTlfMLOsjSp48PQ1OQJ0Ba7raUpCRMK5Uc |
| 108 |
awoFMZ40nMdCg6slXZTJX6l2cUlaTwLX8mt7bdp4xPmzAUuqTH81uAPL+Pk0yX3s |
| 109 |
ryaoVM7tZD3gG0cVEjtFRQQINyGGxeLO0rEdISmMCOyX7A+jgrcSjQUBF9dWvyTI |
| 110 |
oDDss6fEQT5AfV/hdG2G2PARtRa+8/aEiTzV5x5G+JSdqRRvTxXlXpxlU4cDpW3S |
| 111 |
GhxPqVOOFhH7AGbi2gYV |
| 112 |
=ATUx |
| 113 |
-----END PGP SIGNATURE----- |
Archives