Le 20/09/2014 00:08, Aron Xu a écrit :
>>> -------
>>> -> fcitx:
>>
>>> * BLOCKER: presagege is missing as a build-dep, do you mind either not
>> building with it (optional dep) or add it and check it matches the MIR
>> criterias?
>>
>> Added and checked presage and libtinyxml (b-d/dep of presage) to MIR.
>> Live without presage is doable, but better avoid that.
>>
>> -> Thanks, looking to it.
>>
>>> * BLOCKER: there are some LGPL files and no COPYING.LGPL file, please
>> ship it in the upstream tarball.
>>
>> https://github.com/fcitx/fcitx/issues/169
>>
>> I can commit but usually don't do that directly.
>>
>> -> Same than for brise, we need a release with this change before being
>> able to promote to main.
>>
>
> Uploaded to Debian, will sync to Ubuntu after being published.
Ack, keep me posted.

>
>
>>
>>> * NEED INFO/FIXING: fcitx-libs-gclient (and a lot of other binary packages): package-name-doesnt-match-sonames. Do you know why we don't follow upstream soname? They are at 0, so package name should be fcitx-libs-gclient0 as it's a public library. Otherwise, it should be in a private path.
>>> * NEED FIXING: I noted the remark on fcitx-libs in debian/README.source, and even if I'm not a fan of the current approach of the 3 libraries in the same package, I think we can live with it. However, we need to have a soname in the end, so that the binary package name can be bumped for any soname breaking in one of the 3 libraries (as some rdepends are not frmothe same source pacakage, like kde-config-fcitx or youker-assistant).
>>
>> The reason is that we have done this for fcitx-libs before, so that
>> following the same convention. We plan to change that when next major
>> release of Fcitx is out, and for now the API/ABI is pretty stable for
>> a while so there is no real pressure to rush for that.
>>
>> -> Still, this is a main criteria definition. So either have it
>> committed to the packaging github and I can accept without the change
>> being in ubuntu yet (please plan for the first library bump), either
>> ship with the soname change now to get it acked to main.
>>
>
> This is done in Debian git:
> http://anonscm.debian.org/cgit/pkg-ime/fcitx.git/commit/?id=07455aef1c1f19d85d5b68acaaa47ef3f6f464e7

Any upload planned with the fix?
>
>
>>
>>> * NEED FIXING: even if it's wrap under a sh call, better to be coherent and get the script executable: W: fcitx-bin: script-not-executable usr/share/fcitx/data/env_setup.sh
>>
>> This is intentional done by upstream, so after some communication I
>> don't see strong reason to maintain a diff.
>>
>> -> Ack.
>>
>>> * NEED FIXING: debian/copyright:
>>> - fcitx-4.2.8.4/src/frontend/qt/keyserver_x11.h: missing copyright to 2004 Jaroslaw Staniek <email address hidden>
>>> - fcitx-4.2.8.4/src/frontend/xim/lib/XimProto.h: missing the copyright (1992, 1993, 1994 by FUJITSU LIMITED)
>>> - src/ui/kimpanel/kimpanel.c. You set "Copyright: 2008-2010, Zealot.Hoi", but it misses "2010~2012 by CSSlayer *" (files: * doesn't cover it as it's overriden)
>>> - Copyright: 2003-2009, Troy D. Hanson     http://uthash.sourceforge.net => it's actually 2003-2010
>>> - Files: src/lib/fcitx-utils/qsort_r.c. Please mention the short name of the license "MIT/X11 (BSD like)"
>>> - fcitx-4.2.8.4/src/lib/fcitx-qt/fcitxqtkeysequencewidget.h and fcitx-4.2.8.4/src/lib/fcitx-qt/fcitxqtkeysequencewidget_p.h: GPL (v2 or later) LGPL (v2 or later)
>>>   Missing: 2007 Andreas Hartmetz <email address hidden> / 2001, 2002 Ellis Whitehead <email address hidden>
>>> - fcitx-4.2.8.4/src/module/lua/luawrap.c/h, fcitx-4.2.8.4/src/module/lua/luamod.c/h:: GPL (v2 or later). [Copyright: 2012~2012 by xubin *] (needs mention)
>>
>> http://anonscm.debian.org/cgit/pkg-
>> ime/fcitx.git/commit/?id=348b6dae534bb2d08e2f3728b345bb3ed18fa98a
>>
>> -> Thanks!
>>
>>> * MINOR: debian/patches/ubuntu-disable-fcitx-xkb.patch:
>>> do you mind using DEP3 for annotating the patch? (http://dep.debian.net/deps/dep3/)
>>
>> That's a quick `quilt header -e`, I'll update it before next upload if
>> it does not get dropped (attente may don't need it now).
>>
>> -> Can we get a clarification then? Or just commit the DEP3 header to
>> the packaging git so that this comment doesn't go in limbo if you keep
>> the patch it. Please envision to use DEP3 for all your future patches,
>> so easier to track multiple months afterwards.
>>
>
> The patch is now dropped.
>
>>> * MINOR: I think you should add a head to debian/control (after you cp
>> the .in to the control file) to tell to not edit that file as
>> debian/rules, through gencontrol.mk is going to edit and sed it.
>>
>> http://anonscm.debian.org/cgit/pkg-
>> ime/fcitx.git/commit/?id=7669d5769d0e5b7de4ce011a3a3031d82f8adab5
>>
>> -> Excellent!
>>
>>> * NEEDS INFO: I see that we are running the unit tests at build time.
>> Is there any integration tests that we can run, installing the various
>> modules? (like -anthy, -chewing… and so on…)
>>
>> Nope, the testing is limited to some core functions of the framework
>> itself, and there is no usable testing facility for other modules in
>> Fcitx 4.8 series.
>>
>> -> Ack.
>>
>>> -----
>>>
>>> -> fcitx-anthy:
>>> * NEEDS FIXING: the package is multi-arch and should be marked as such in debian/control
>>
>> It makes little sense to make sure all the modules are co-installable
>> since only one copy can be used, so that the work isn't actively
>> pushed. Changing:
>> http://anonscm.debian.org/cgit/pkg-ime/fcitx-anthy.git/commit/?id=0be99f31f8c8dc81afb09696d86c2095a01cc9ae
>>
>> -> Thanks. Any upload with the change planned?
>
> No upload planned, it's not really useful because of the dependency
> chain, will be incorporated with more changes.
Ack.
>
>
>>
>>> * NEED FIXING: debian/copyright: doesn't contain any other copyright holder. Please fix it.
>>> For instance: src/factory.h: GPL (v2 or later)
>>>   [Copyright: 2004 Hiroyuki Ikezoe / 2004 Takuro Ashie / 2012 CSSlayer / 2004 James Su <email address hidden>]
>>> src/imengine.cpp: GPL (v2 or later)
>>>   [Copyright: 2004 - 2005 Hiroyuki Ikezoe <email address hidden> / 2012 CSSlayer / 2004 - 2005 Takuro Ashie <email address hidden> / 2004 James Su <email address hidden>]
>>
>> http://anonscm.debian.org/cgit/pkg-ime/fcitx-anthy.git/commit/?id=5d5249cdf62b9a21b92ceedeeb4ef0d16f247619
>> http://anonscm.debian.org/cgit/pkg-ime/fcitx-anthy.git/commit/?id=d74a427fb6f53502d3c80e2b97d1b4925fe27669
>>
>> -> Thanks.
>>
>>> ------
>>> -> ficitx-cloudpinyin:
>>> * NEEDS FIXING: the package is multi-arch and should be marked as such in debian/control
>>
>> http://anonscm.debian.org/cgit/pkg-ime/fcitx-
>> cloudpinyin.git/commit/?id=f47d60ab904aba6029e631274c899fada1758d95
>>
>> -> Thanks. Any upload with the change planned?
>
> No upload planned, same as fcitx-anthy.
Ack.
>
>
>>
>>> * MINOR: debian/copyright:
>>> Copyright should be 2011-2012, not only 2011
>>
>> http://anonscm.debian.org/cgit/pkg-ime/fcitx-
>> cloudpinyin.git/commit/?id=52442b5c129411cf21c4b25887af980e8b54857b
>>
>> -> Thanks.
>>
>>> * NITPICK: typo in debian/control: "to retrive result candidates" -> retrieve
>>>
>>
>> http://anonscm.debian.org/cgit/pkg-ime/fcitx-
>> cloudpinyin.git/commit/?id=82eabd3d4ee8c06e5d903a3f7fffbc5cce43f615
>>
>> -> Thanks.
>>
>>> ------
>>> -> fcitx-configtool
>>> * BLOCKER: debian/copyright is not well-formated:
>>> line 5: missing Files: line.
>>> The License text should be in a separate stanza as well.
>>
>> http://anonscm.debian.org/cgit/pkg-ime/fcitx-
>> configtool.git/commit/?id=a948aa5caaf4c3cdc57bb51f642276bb6754e0ce
>>
>> -> Thanks. Any upload with the change planned?
>
> Uploaded to Debian and synced to Ubuntu.
Thanks, keep me posted once available

>
>
>>
>>> * NITPICK: debian/copyright
>>> Depends: line for fcitx-config-gtk is too long, please wrap it.
>>>
>>
>> http://anonscm.debian.org/cgit/pkg-ime/fcitx-
>> configtool.git/commit/?id=a133111096e95794fa1d221012237c4e9082e104
>>
>> -> Thanks.
>>
>>> ------
>>> -> fcitx-googlepinyin
>>> * BLOCKER: same issue in debian/copyright
>>
>> http://anonscm.debian.org/cgit/pkg-ime/fcitx-
>> googlepinyin.git/commit/?id=a8dda275395884afaa59d47f69c6c5dbca4f2b78
>>
>> -> Thanks. Any upload with the change planned?
>>
>>> * BLOCKER: build-dep on libgooglepinyin0-dev which is in universe.
>> Please drop that plugin from the MIR or add the dependency after
>> checking it's suiting the MIR requirements.
>>
>> Added and checked.
>>
>> -> Ack, will look at it.
>>
>>> * NEEDS FIXING: the package is multi-arch and should be marked as such in debian/control
>>>
>>
>> http://anonscm.debian.org/cgit/pkg-ime/fcitx-
>> googlepinyin.git/commit/?id=7dee3104f9725637d42da95424b5fbf59ad9e120
>>
>> -> Thanks. Any upload with the change planned?
>>
>
> No upload planned, same as fcitx-anthy.

It would be needed to upload it for the first BLOCKER above (misformed debian/copyright)

>
>
>>> ------
>>> -> fcitx-hangul
>>> * NEEDS FIXING: the package is multi-arch and should be marked as such in debian/control
>>
>> http://anonscm.debian.org/cgit/pkg-ime/fcitx-
>> hangul.git/commit/?id=0d4418f3a35e1346f6c4461d3eb5d97016e4f002
>>
>> -> Thanks. Any upload with the change planned?
>>
>
> No upload planned, same as fcitx-anthy.
Ack.
>
>
>>> * NEEDS INFO: fcitx-hangul: hardening-no-stackprotector usr/lib/x86_64
>> -linux-gnu/fcitx/fcitx-hangul.so. Any way we can fix this? Not having
>> stackprotector leads the potentially code that can be attacked.
>>
>> I don't see such problem, nor do lintian.d.o:
>> https://<email address hidden>#fcitx-hangul
>>
>> -> It was a false positive in lintian utopic that has been fixed.
>>
>>> * MINOR: debian/copyright:
>>> Copyright: 2012 CSSlayer <email address hidden> -> should be 2010-2012
>>>
>>
>> http://anonscm.debian.org/cgit/pkg-ime/fcitx-
>> hangul.git/commit/?id=648c746466c92d6b9a86a5d4bcade691e57b2f86
>>
>> -> Thanks!
>>
>>> ------
>>> -> fcitx-m17n
>>> * NEEDS FIXING: the package is multi-arch and should be marked as such in debian/control
>>
>> http://anonscm.debian.org/cgit/pkg-ime/fcitx-
>> m17n.git/commit/?id=15108dfe41f329f4bb041ca96f2dbebda117ec24
>>
>> -> Thanks. Any upload with the change planned?
>>
>
> No upload planned, same as fcitx-anthy.
Ack.
>
>
>>> * MINOR: debian/copyright:
>>> missing   Copyright: 1995-1997 Peter Mattis, Spencer Kimball and Josh MacDonald in im/keysymname.c
>>>
>>
>> http://anonscm.debian.org/cgit/pkg-ime/fcitx-
>> m17n.git/commit/?id=9d981e902b048bc3cdb74d94b15ab896ae8029da
>>
>> -> Thanks!
>>
>>> ------
>>> -> fcitx-qimpanel:
>>> * BLOCKER: there is no COPYING for GPL2 and GPL3. Both are needed to be shipped upstream.
>>
>> https://github.com/lenky0401/fcitx-qimpanel/issues/26
>>
>> -> do they plan to do a release soon? We can't promote it without a
>> release containing those files.
>>
>
> Yes, in archive now.
Great, please update the status.
>
>
>>> * NEEDS FIXING: the package is multi-arch and should be marked as such
>> in debian/control
>>
>> I prefer not doing it at this moment because I need some time to
>> verify the behavior.
>>
>> -> In case of doubt, only mark it as multi-arch: same.
>>
>
> It's not a library, and M-A: same isn't right. We need test to see if
> it can work as M-A: foreign.

Is this testing planned? The first remark and the need testing is more than a month old, which may me wonder if this will be tested at all?
>
>
>>
>>> * NEEDS INFO: I guess the autostart binary shouldn't be in usr/bin, but more in an exec path like libexec for instance.
>>>
>>
>> Will talk with relevant people to see if we can move that.
>>
>> -> did you get any info from it?
>>
>
> Not being dealt in time of last release, but upstream agrees for the
> idea.
>
>>> ------
>>> -> fcitx-qimpanel-configtool:
>>> * BLOCKER: there is no COPYING for GPL2 and GPL3. Both are needed to be shipped upstream. There is only an empty LICENSE file.
>>>
>>
>> This package is merged to fcitx-qimpanel upstream for next release, so
>> there is no need to open an issue again.
>>
>> -> When is the release planned?
>>
>
> This happened already, removing the task from MIR.
Ack.
>
>
>>> ------
>>> -> fcitx-qt5
>>> * BLOCKER: there is no COPYING for GPL2 nor LGPL2. Both are needed to be shipped upstream.
>>
>> https://github.com/fcitx/fcitx-qt5/issues/4
>>
>> -> do they plan to do a release soon? We can't promote it without a
>> release containing those files.
>>
>>> * NEEDS FIXING: fcitx-libs-qt5: package-name-doesnt-match-sonames
>> libfcitx-qt5-0. Please rename the binary package + add an override,
>> there is only one shared lib and it's better to follow debian packaging
>> convention for Qt libs.
>>
>> As expained for fcitx-libs-gclient and fcitx-libs.
>>
>> -> Same remark and question than fcitx-libs then :)
>>
>
> Still prefer not doing that right now, and wait for next major
> release, if possible.

I don't think we can do without a release (for the COPYING fixes) anyway. I don't think it's acceptable to let that package entering main without following the standard Qt convention, as there is a high chance that this will be forgotten and never dealt afterwards.

>
>
>>> * NEEDS FIXING:
>>> debian/copyright:
>>> missing src/keyserver_x11.h: LGPL (v2 or later)
>>>   [Copyright: 2001 Ellis Whitehead <email address hidden> / 2004 Jaroslaw Staniek <email address hidden>]
>>
>> http://anonscm.debian.org/cgit/pkg-ime/fcitx-
>> hangul.git/commit/?id=648c746466c92d6b9a86a5d4bcade691e57b2f86
>>
>> -> Thanks!
>>
>>> * MINOR: it would be nice to have a .symbol file.
>>>
>>
>> Will deal with it a bit later.
>>
>> -> It will be good to invest some time on this. Not sure why I put that
>> as MINOR, should be rather NEEDS FIXING actually. Please do work on it
>> or ask if you need help (as we'll need a release + explanation on the
>> soname scheme)

Any feedback on that one? As mentioned, this should be fixed for entering main.
>>
>>
>>> ------
>>> -> fcitx-rime:
>>> * BLOCKER: COPYING is GPL2, so we can assume the only file is src/fcitx-rime.c is under that license. debian/copyright mention GPL3+. Needs to be fixed in accordance.
>>
>> http://anonscm.debian.org/cgit/pkg-ime/fcitx-
>> rime.git/commit/?id=2b0beb3d640500eb533b7fa0755a5c33c7182fc1
>>
>> -> Thanks!
>>
>>> * BLOCKER: build-deps on librime which is in universe. Please add that
>> one to the MIR if it matches the MIR criterias or drop that one.
>>
>> Already in the report.
>>
>> -> Will look at it.
>>
>>> * BLOCKER: this package doesn't build on armhf which is one of our
>> supported arch. It needs to be buildable on that arch (librime-data
>> can't be installed)
>>
>> Need to retry the build, old librime does not build on that much
>> architectures so no librime-data, but current one does.
>>
>> -> Thanks for retrying the build, seems it passed now!
>>
>>
>>> * NEEDS FIXING: the package is multi-arch and should be marked as such in debian/control
>>
>> This makes little sense here, as librime-data isn't M-A'd, which is a
>> Depends of this package.
>>
>> -> Ack.
>>
>>> * MINOR: The source in debian/copyright is fcitx one, not fcitx-rime
>>>
>>
>> Upstream-Name shows fcitx-rime, and it's exactly maintained by fcitx
>> upstream. Actually it's quite normal that active modules get merged to
>> Fcitx's umbrella project, and Fcitx maintainers ensure APIs are synced
>> once there's any change.
>>
>> -> Ack.

I see this one is still listed on the bug report. However all the build-depends and runtime deps (rime and so on…) have been removed. Did you forget to remove it from the bug?

>>
>>
>>> ------
>>> -> fcitx-sunpinyin:
>>> * BLOCKER: COPYING is GPL2 as the files from the project. debian/copyright mention GPL3+. Needs to be fixed in accordance.
>>
>> http://anonscm.debian.org/cgit/pkg-ime/fcitx-
>> sunpinyin.git/commit/?id=214a48b57b3104f96168f6dd794d2fc692e46cd6
>>
>> -> Thanks!
>>
>>> * BLOCKER: debian/copyright is not well-formated (same issue than in
>> fcitx-configtool)
>>
>> http://anonscm.debian.org/cgit/pkg-ime/fcitx-
>> sunpinyin.git/commit/?id=bb05161922d11fb98ee21d6154e74d91bd5a2e54
>>
>> -> Thanks!
>>
>>> * NEEDS FIXING: the package is multi-arch and should be marked as such in debian/control
>>>
>>
>> This makes little sense as sunpinyin-data isn't M-A ready, which is a
>> Depends of fcitx-sunpinyin.
>>
>> -> Ok.
>>
>>> ------
>>> -> fcitx-table-extra:
>>> * BLOCKER: tables/scj6.txt is GPL3, there is no COPYING file mentionning GPL3. You need upstream to ship it.
>>
>> https://github.com/fcitx/fcitx/issues/170
>>
>> -> Same than the rest, we need a release with the change to be able to
>> distribute it.
>>
>
> Progress is this is applied to upstream git, but not released yet.
Ok, keep me posted once you have a release ready.

>
>
>>> * NEEDS FIXING:
>>> W: fcitx-table-cantonese: icon-size-and-directory-name-mismatch usr/share/icons/hicolor/64x64/apps/fcitx-cantonese.png 48x48
>>> W: fcitx-table-stroke5: icon-size-and-directory-name-mismatch usr/share/icons/hicolor/64x64/apps/fcitx-stroke5.png 48x48
>>> W: fcitx-table-zhengma: icon-size-and-directory-name-mismatch usr/share/icons/hicolor/48x48/apps/fcitx-zhengma.png 64x64
>>
>> Will get a patch for upstream soon and apply to Debian/Ubuntu.
>>
>> -> Great!
>>
>>> * NEEDS INFO: debian/rules:
>>> override_dh_install:
>>>     find . -name *.mo -delete
>>> why stripping the generated .mo file? They aren't shown anyway in any UI?
>>
>> No, they are not shown atm.
>>
>> -> Mind adding a small comment in debian/rules about it?
>>
>>>
>>> ------
>>> -> fcitx-table-other:
>>> * NEEDS INFO: debian/rules:
>>> override_dh_install:
>>>     find . -name *.mo -delete
>>> why stripping the generated .mo file? They aren't shown anyway in any UI?
>>>
>>
>> No, they are not shown atm.
>>
>> -> Mind adding a small comment in debian/rules about it?

Any feedback on that one?

>>
>>
>>> ------
>>> -> librime:
>>> * BLOCKER: libkyotocabinet-dev, libmarisa-dev and libyaml-cpp-dev build-deps are in universe. Please add that one to the MIR if it matches the MIR criterias or drop those.
>>
>> Added and checked.
>>
>> -> Will check them soon.
>>
>>> Note that libboost-signals-dev is in universe, but the source is in
>> main, so we can promote it as the same time.
>>
>>> * MINOR: would be nice to multi-arch the library.
>>> * MINOR: it would be nice to have a .symbol file (not sure why there is an override no-symbols-control-file)
>>
>> It makes little sense since the dependent data files aren't identical
>> across different architectures and upstream does not maintain it at
>> the moment.
>>
>> -> Ok. As long as the binary package finishes with 0, it's fine as it
>> denotes unstable ABI. Do we know more about their plan for ABI
>> stability? (it's something that we encourage, especially for main).
>>
>>> * MINOR: debian/copyright:
>>> Files: thirdparty/src/glog/*
>>>        thirdparty/src/gtest/*
>>> Copyright: 2008, Google Inc.
>>> It should be 1999-2008
>>
>> http://anonscm.debian.org/cgit/pkg-
>> ime/librime.git/commit/?id=cac4f2a7fdfec25a572d22649195c8a76809ce81
>>
>> -> Thanks!
>>
>>> * MINOR/NEED INFO: any idea why the source has a priority extra?
>>>
>>
>> Same reason as brise, changed anyway:
>> http://anonscm.debian.org/cgit/pkg-ime/librime.git/commit/?id=42ad5837d156057344559ee89635132a1b71dd3e
>>
>> -> Yeah, see my explanation in the first stenza. Thanks for changing it!
>>
>> ------
>> Packages added recently:
>> ------
>> -> libgooglepinyin:
>> * NEEDS FIXING: any chance for a .symbols file (seems though that the ABI is unstable, isn't it?)?
>>
>> * MINOR: debian/compat is set to 9, where the build-dep on debhelper is
>> (>= 8.1.3~). Please fix accordingly.
>>
>> ------
>> -> libtinyxml/tinyxml:
>> * NEEDS INFO: We already have multiple xml libraries in main. Any chance to switch the build-dep on that one to another one?
>> * BLOCKER: not published in ubuntu (did you subscribe the wrong component, how did you check the content?)? https://launchpad.net/ubuntu/+source/libtinyxml. I guess you meant tinyxml (and I analyzed this one). Please retarget the bug report if so.
>> * NEEDS FIXING: as upstream seems to maintain their ABI, would be nice to have some symbols files.
>>
>


As a general rule, I would really ask you to clean up this MIR, remove all components that you dropped (like there is still fcitx-rime above), revert the tasks to "confirmed" for those which are ready in your opinion.
Also, it would be great to have the answers even on the "MINOR" remarks :) (I classified the importance of issues to help you prioritizing, not to have the MINOR ones ignored :)).

Thanks!