Summary: | Allow use of DOCS="" as a way of resetting the PM supplied list (was: =sys-fs/eudev-1.2-r1: In /usr/libexec/paludis/utils/dodoc at line 41 : at least one argument needed) | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Franz Trischberger <franz.trischberger> |
Component: | [OLD] Core system | Assignee: | PMS/EAPI <pms> |
Status: | RESOLVED DUPLICATE | ||
Severity: | normal | CC: | alexanderyt, ciaran.mccreesh, esigra, franz.trischberger, hasufell, jdhore, multilib+disabled, ssuominen, ulm |
Priority: | Normal | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
See Also: |
https://bugs.gentoo.org/show_bug.cgi?id=481642 https://bugs.gentoo.org/show_bug.cgi?id=463736 https://bugs.gentoo.org/show_bug.cgi?id=487852 |
||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Attachments: |
build.log
Patch for multilib-minimal.eclass |
sys-apps/paludis doesn't have a default list of DOCS to install like sys-apps/portage? i'm asking, not stating. (In reply to Samuli Suominen from comment #1) > sys-apps/paludis doesn't have a default list of DOCS to install like > sys-apps/portage? i'm asking, not stating. I don't know about that. AFAIK it's the first time I hit such an error. But your question seems to imply this might be an issue with paludis. As I am on the way "back to the roots" (today was moving from systemd back to openrc ;)) I don't know how long I will have paludis running - probably next weekend I am transitioning back to portage, as paludis bit me quite some time in the past weeks and maintenance seems to be sub-par (at best, if there currently is any). (Just stating this, if a DEV wants to fix this and needs me using paludis) If this is because of empty DOCS="", then what is the correct way to tell default_src_install to ignore installation of default list of docs from the PM? app-dicts/canna-2ch/canna-2ch-20040519.ebuild:DOCS="" dev-python/icalendar/icalendar-3.2.ebuild:DOCS="" gnome-base/gvfs/gvfs-1.14.2.ebuild: DOCS="" media-fonts/ekushey-bangla-fonts/ekushey-bangla-fonts-20070420.ebuild:DOCS="" media-fonts/sil-padauk/sil-padauk-2.8.ebuild:DOCS="" media-fonts/thaifonts-scalable/thaifonts-scalable-0.4.13.ebuild:DOCS="" media-libs/jpeg/jpeg-6b-r11.ebuild:DOCS="" media-libs/tiff/tiff-3.9.7-r1.ebuild:DOCS="" net-irc/hexchat/hexchat-2.9.5.ebuild:DOCS="" net-irc/hexchat/hexchat-2.9.5_p20130525.ebuild:DOCS="" net-irc/hexchat/hexchat-2.9.5-r1.ebuild:DOCS="" net-mail/notmuch/notmuch-0.10.2-r3.ebuild: DOCS="" bindings python distutils_src_install net-mail/notmuch/notmuch-0.11.1-r3.ebuild: DOCS="" bindings python distutils_src_install net-mail/notmuch/notmuch-0.12.ebuild: DOCS="" bindings python distutils_src_install net-mail/notmuch/notmuch-0.13.1.ebuild: DOCS="" bindings python distutils_src_install net-mail/notmuch/notmuch-0.14-r1.ebuild: DOCS="" bindings python distutils_src_install net-mail/notmuch/notmuch-0.15.2.ebuild: DOCS="" bindings python distutils_src_install net-mail/notmuch/notmuch-0.16.ebuild: DOCS="" bindings python distutils-r1_src_install sys-fs/eudev/eudev-1.2-r1.ebuild:DOCS="" sys-fs/eudev/eudev-9999.ebuild:DOCS="" Paludis is actually compliant with PMS here: http://dev.gentoo.org/~ulm/pms/5/pms.html#x1-1030009.1.9 Seems the case that empty DOCS could be needed was overlooked when the default src_install was specified for EAPI 4. (In reply to Samuli Suominen from comment #3) > If this is because of empty DOCS="", then what is the correct way to tell > default_src_install to ignore installation of default list of docs from the > PM? > > app-dicts/canna-2ch/canna-2ch-20040519.ebuild:DOCS="" > dev-python/icalendar/icalendar-3.2.ebuild:DOCS="" > gnome-base/gvfs/gvfs-1.14.2.ebuild: DOCS="" > media-fonts/ekushey-bangla-fonts/ekushey-bangla-fonts-20070420.ebuild:DOCS="" > media-fonts/sil-padauk/sil-padauk-2.8.ebuild:DOCS="" > media-fonts/thaifonts-scalable/thaifonts-scalable-0.4.13.ebuild:DOCS="" > media-libs/jpeg/jpeg-6b-r11.ebuild:DOCS="" > media-libs/tiff/tiff-3.9.7-r1.ebuild:DOCS="" > net-irc/hexchat/hexchat-2.9.5.ebuild:DOCS="" > net-irc/hexchat/hexchat-2.9.5_p20130525.ebuild:DOCS="" > net-irc/hexchat/hexchat-2.9.5-r1.ebuild:DOCS="" > net-mail/notmuch/notmuch-0.10.2-r3.ebuild: DOCS="" bindings python > distutils_src_install > net-mail/notmuch/notmuch-0.11.1-r3.ebuild: DOCS="" bindings python > distutils_src_install > net-mail/notmuch/notmuch-0.12.ebuild: DOCS="" bindings python > distutils_src_install > net-mail/notmuch/notmuch-0.13.1.ebuild: DOCS="" bindings python > distutils_src_install > net-mail/notmuch/notmuch-0.14-r1.ebuild: DOCS="" bindings python > distutils_src_install > net-mail/notmuch/notmuch-0.15.2.ebuild: DOCS="" bindings python > distutils_src_install > net-mail/notmuch/notmuch-0.16.ebuild: DOCS="" bindings python > distutils-r1_src_install > sys-fs/eudev/eudev-1.2-r1.ebuild:DOCS="" > sys-fs/eudev/eudev-9999.ebuild:DOCS="" I also use notmuch and it installs fine. (Just tested again) (In reply to Franz Fellner from comment #5) > I also use notmuch and it installs fine. (Just tested again) That list might contain false-positives, so that might not proof anything. Actually, the only packages that don't look like false positives are media-libs/jpeg, media-libs/tiff, and sys-fs/eudev. And these are not using the package manager's default, but multilib-minimal_src_install(). + 13 Aug 2013; Ian Stakenvicius <axs@gentoo.org> eudev-1.2-r1.ebuild, + eudev-1.2.ebuild, eudev-9999.ebuild: + dropped DOCS='' as it is not PMS-legal (bug 480892) [..] ..to address the failure in eudev, for now. Once we figure out exactly what multilib-* eclass might be contributing we should reassign the bug. Thanks for the fix, installed just fine :) (In reply to Ian Stakenvicius from comment #8) > + 13 Aug 2013; Ian Stakenvicius <axs@gentoo.org> eudev-1.2-r1.ebuild, > + eudev-1.2.ebuild, eudev-9999.ebuild: > + dropped DOCS='' as it is not PMS-legal (bug 480892) [..] > > > ..to address the failure in eudev, for now. Once we figure out exactly what > multilib-* eclass might be contributing we should reassign the bug. Ian, did you just forget to close the bug because it looks fixed. (In reply to Anthony Basile from comment #10) > (In reply to Ian Stakenvicius from comment #8) > > + 13 Aug 2013; Ian Stakenvicius <axs@gentoo.org> eudev-1.2-r1.ebuild, > > + eudev-1.2.ebuild, eudev-9999.ebuild: > > + dropped DOCS='' as it is not PMS-legal (bug 480892) [..] > > > > > > ..to address the failure in eudev, for now. Once we figure out exactly what > > multilib-* eclass might be contributing we should reassign the bug. > > Ian, did you just forget to close the bug because it looks fixed. I don't think this is fixed at all, both PMS and the code in multilib-minimal.eclass remain broken as they should be allowing empty DOCS="" as a way of resetting the PM supplied list. The code for DOCS handling at the end of multilib-minimal.eclass looks broken too, as it ends up calling dodoc without arguments when it's set but empty? Or am I seeing wrong? CCing multilib-minimal eclass maintainer hasufell@ as well for that (In reply to Samuli Suominen from comment #12) > The code for DOCS handling at the end of multilib-minimal.eclass looks > broken too, as it ends up calling dodoc without arguments when it's set but > empty? Or am I seeing wrong? > CCing multilib-minimal eclass maintainer hasufell@ as well for that Please do tests before you call something broken. Empty DOCS works perfectly with portage. http://git.overlays.gentoo.org/gitweb/?p=proj/portage.git;a=blob;f=bin/ebuild-helpers/doins;h=4679e83a9a7140e100da733cb34faad384991a64;hb=HEAD#l10 (In reply to Julian Ospald (hasufell) from comment #13) > Empty DOCS works perfectly with portage. > > http://git.overlays.gentoo.org/gitweb/?p=proj/portage.git;a=blob;f=bin/ > ebuild-helpers/doins;h=4679e83a9a7140e100da733cb34faad384991a64;hb=HEAD#l10 Bug 481664 now. Created attachment 356468 [details, diff] Patch for multilib-minimal.eclass I suggest that we proceed as follows: - Fix multilib-minimal_src_install to allow for empty DOCS, see attached patch. - Fix the package manager's default_src_install accordingly in EAPI 6 (see bug 463736). - Add a QA warning to dodoc if called without arguments, in order to prevent further ebuilds from relying on Portage specific behaviour (bug 481664). - Fix the few (if any) ebuilds that are using the package manager's default_src_install with empty DOCS. (In reply to Ulrich Müller from comment #15) > Created attachment 356468 [details, diff] [details, diff] > Patch for multilib-minimal.eclass > > I suggest that we proceed as follows: > - Fix multilib-minimal_src_install to allow for empty DOCS, see attached > patch. But what's the point for that? IMO the inconsistency with EAPI behavior is not worth it. You can simple override multilib_src_install_all() since it does nothing more than installing docs AFAIR. (In reply to Michał Górny from comment #16) > > - Fix multilib-minimal_src_install to allow for empty DOCS, see attached > > patch. > > But what's the point for that? IMO the inconsistency with EAPI behavior is > not worth it. You can simple override multilib_src_install_all() since it > does nothing more than installing docs AFAIR. Huh? multilib-minimal_src_install() always processes DOCS, regardless of multilib_src_install_all() being defined or not. So there's no easy way to override the default. Oh my. Hasufell, kiddo, why did you do that? ;f (In reply to Michał Górny from comment #18) > Oh my. Hasufell, kiddo, why did you do that? ;f WTF? YOU requested it and YOU reviewed it. (In reply to Julian Ospald (hasufell) from comment #19) > (In reply to Michał Górny from comment #18) > > Oh my. Hasufell, kiddo, why did you do that? ;f > > WTF? YOU requested it and YOU reviewed it. Really? I don't recall telling you to do docs outta _all(). (In reply to Michał Górny from comment #20) > (In reply to Julian Ospald (hasufell) from comment #19) > > (In reply to Michał Górny from comment #18) > > > Oh my. Hasufell, kiddo, why did you do that? ;f > > > > WTF? YOU requested it and YOU reviewed it. > > Really? I don't recall telling you to do docs outta _all(). Then you probably missed that during your review and now blame me for not doing it the way you wanted? Anyway, I am open to suggestions and patches and have always been. *** This bug has been marked as a duplicate of bug 463736 *** |
Created attachment 355882 [details] build.log I tried to install eudev which failed. Enabling USE="doc" failed with the same error.