| Summary: | sci-geosciences/grass-6.4.2 fails to build | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Patrick Lauer <patrick> |
| Component: | New packages | Assignee: | Sam Jorna (wraeth) (RETIRED) <wraeth> |
| Status: | RESOLVED FIXED | ||
| Severity: | normal | CC: | moult, proxy-maint, thomas.bettler |
| Priority: | Normal | ||
| Version: | unspecified | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
| Bug Depends on: | 529530 | ||
| Bug Blocks: | |||
| Attachments: | an ebuild for grass 6.4.4 | ||
|
Description
Patrick Lauer
2014-06-23 02:48:34 UTC
I am also unable to build this for a long long time, looking to the bug list shows this is broken in many different ways. Maybe this should be treecleaned finally + 29 Mar 2015; Pacho Ramos <pacho@gentoo.org> qgis-1.8.0.ebuild: + Disable grass as it will be removed (#514514) + Pacho, can I confirm to remove DEPEND="!<sci-geosciences/grass-6.4.1-r1" in sys-processes/parallel? Also, if I package.mask this, I assume I should also mask sci-geosciences/gdal-grass? (In reply to Dion Moult from comment #3) > Pacho, can I confirm to remove DEPEND="!<sci-geosciences/grass-6.4.1-r1" in > sys-processes/parallel? I think the general policy is to leave the blockers (the "!<" deps I mean) near forever to ensure we don't break systems that hasn't been updated for a long long time... then, it's better to keep the blocker there, it won't cause any issues for this. > > Also, if I package.mask this, I assume I should also mask > sci-geosciences/gdal-grass? Please remember to grep the tree for finding any ebuilds that could still need this :) I have just dropped the old qgis ebuild that still needed this: [master 46283e2] sci-geosciences/qgis: Drop old 2 files changed, 101 deletions(-) delete mode 100644 sci-geosciences/qgis/qgis-1.7.0.ebuild And, yeah, the mask should be also for sci-geosciences/gdal-grass in the same chunk with grass Thanks a lot for your help and best regards :) # Dion Moult <moult@gentoo.org> (27 Sep 2015) # Masked for removal in 30 days. Bug #514514 # Fails to build for ages. sci-geosciences/grass sci-geosciences/gdal-grass 6;4.4 builds fine with current stable x64 gcc & package version translated ebuild from 6.4.2 to 6.4.4. Loads fine too. Although I can not say whether it is enough functional to deserve survival, since I don't know how to use it ;-). I'll keep it to my /usr/local/portage anyhow. Moreover, GRASS seems at version 7 now. Maybe GENTOO will treeclean this for other reason that those stated? Please see the work-in-progress at #541622. I hope it could be bumped instead of being removed. This package mask only applies to 6.4.2. If/when a working ebuild for >6.4.2 (e.g., 6.4.4 or 7 as you have suggested), then grass can and should definitely come back :) However, seeing that the sci-geosciences herd is empty, this means that there is no responsible dev for greenlighting a bumped ebuild. I have sent out a message asking for devs interested in sci-geosciences. If nobody responds, I will revbump this package myself. My current understanding is that we should directly to version 7 (obviously I don't use the package myself) - is that correct? (In reply to Dion Moult from comment #8) > This package mask only applies to 6.4.2. If/when a working ebuild for >6.4.2 > (e.g., 6.4.4 or 7 as you have suggested), then grass can and should > definitely come back :) > > However, seeing that the sci-geosciences herd is empty, this means that > there is no responsible dev for greenlighting a bumped ebuild. > > I have sent out a message asking for devs interested in sci-geosciences. If > nobody responds, I will revbump this package myself. My current > understanding is that we should directly to version 7 (obviously I don't use > the package myself) - is that correct? grass-7.0.0 seems to compile fine. See https://bugs.gentoo.org/show_bug.cgi?id=541622 Regards Juergen (In reply to jannis from comment #7) > Please see the work-in-progress at #541622. I hope it could be bumped > instead of being removed. Seconded. No reason to remove the package if there's a more recent version that can be made to work. I just posted an ebuild of 7.0.1 to #541622. (In reply to Dion Moult from comment #8) > This package mask only applies to 6.4.2. If/when a working ebuild for >6.4.2 > (e.g., 6.4.4 or 7 as you have suggested), then grass can and should > definitely come back :) Hi, could you please mask only =sci-geosciences/grass-6.4.2 rather than the whole sci-geosciences/grass ? with the mask change you made, you are preventing ANY version to build, regardless if it's only the ebuild that don't build or an ebuild from an other overlay or with local changes ! As soon as nobody is FIXING the ton of pending bugs of the package in the tree: https://bugs.gentoo.org/buglist.cgi?quicksearch=sci-geosciences%2Fgrass&list_id=2929314 The mask for removal will remain and it will be punted. People should either proxy maintain it and finally get the version in the tree fixed or rely in overlays Created attachment 413588 [details] an ebuild for grass 6.4.4 Builds fine here with https://grass.osgeo.org/grass64/source/grass-6.4.4.tar.gz, with FLAGS "cairo gmath opengl postgres python wxwidgets" set in portage use. (In reply to Pacho Ramos from comment #12) > The mask for removal will remain and it will be punted. People should either > proxy maintain it and finally get the version in the tree fixed or rely in > overlays Well, there are now two ebuilds for two more recent versions: a 6.4.4 ebuild in this bug, and a 7.0.1 ebuild in #541622. What do we need to do to get those into the tree? I will CC proxy-maintainers then: https://wiki.gentoo.org/wiki/Project:Proxy_Maintainers Just to repeat my previous comment: (from comment #8) > This package mask only applies to 6.4.2. If/when a working ebuild for >6.4.2 > (e.g., 6.4.4 or 7 as you have suggested), then grass can and should > definitely come back :) > > However, seeing that the sci-geosciences herd is empty, this means that > there is no responsible dev for greenlighting a bumped ebuild. > > I have sent out a message asking for devs interested in sci-geosciences. If > nobody responds, I will revbump this package myself. My current > understanding is that we should directly to version 7 (obviously I don't use > the package myself) - is that correct? To reiterate: 1. Proxy maintainership is fine, except that there is a more fundamental issue of the herd not existing. I have sent out a message. Give it another week in case someone responds. If not, then we can decide what to do. 2. Given that 6.x branch seems to have a lot of issues, and 7.x is a big jump ahead, I see no reason to include 6.x any more. 3. The ebuilds out there for 7.x have some QA issues. Not only did a quick inspection reveal commented out code, it should refer to the proper mysql virtual, as well as update the python eclass. So in conclusion :) 1. Wait a week more. If nobody responds, start talking to sci herd. 2. Clean up the 7.x ebuild. 3. Whoever is maintaining it in the end gets listed, and 7.x gets committed. Just bear in mind this is a big ask for a proxy user maintainer to take onn commit 2fa849db86f415ee6eca0a7fb965c88606ace3e6 Author: Ian Delaney <idella4@gentoo.org> Date: Wed Oct 14 23:11:05 2015 +0800 sci-geosciences/grass: bump to -7.0.1 Inherit python-single-r1 eclass, three USE flags added and two removed to update new configure options; set slot operators on required deps, patches to fix build issues; install desktop file via make_desktop_entry, tidy EAPI style vars set new proxy-maintainer 'wraeth', add proxy-maintainers herd in metadata along with new use flags Gentoo bug: #514514 |