[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Fwd: Saga 2.1.1 - debhelper9 - multiarch



Sebastiaan,

Thanks for the extensive review!
Most things seem easy to fix though I think will only fix the spelling
errors in trunk.

Since afaik no other software than saga itself relies on libsaga
upstream is actually quite loose regarding api/abi changes, and so it
changes every release.

We originally chose to use libsaga instead of the actual full library
names to avoid having to pass through the new queue everytime.

Considering saga-depends: perhaps Francesco can explain better why he
chose to do it like that. I should add that the system was not
perfectly working: the file was not updated most previous releases.

Johan

On Fri, Jan 17, 2014 at 4:13 AM, Sebastiaan Couwenberg
<sebastic@xs4all.nl> wrote:
> On 01/13/2014 10:14 AM, Johan Van de Wauw wrote:
>> Is anyone actually reading the pkg-grass-devel mailing list or should
>> I rather send my questions here?
>
> The pkg-grass-devel is most used for bug reports, commit logs. While
> sometimes feedback is given in response to mails to pkg-grass-devel,
> general discussion is better visible here on debian-gis.
>
>> Anyway:
>> I've updated saga to the last upstream release (2.1.1), a bugfix release.
>>
>> The major change in packaging is that I switched to debhelper9 and
>> multiarch. I'd like to hear from someone more knowledgeable whether I
>> did the right thing for the saga package itself.
>
> In general your package looks good. I do have some comments.
>
>
> The libsaga binary package still needs a Pre-Depends on
> ${misc:Pre-Depends} for debhelpers automatic Multi-Arch related
> dependencies, see:
>
> https://wiki.debian.org/Multiarch/Implementation#Recipes_for_converting_packages
>
>
> The libsaga-dev binary package installs /usr/bin/saga-depends, is that
> script really needed? Handling dependencies on the saga libraries should
> be done with symbols or shlibs files.
>
> The policy section about shared libraries was rewritten in version
> 3.9.4, so the packages doesn't really implement the policy version it
> claims or explains why it diverges.
>
> http://www.debian.org/doc/debian-policy/ch-sharedlibs.html#s-sharedlibs-depends
>
> Since saga libraries are in C++ their symbols are likely to change
> often, you may not want to use symbols file for saga because of this.
> But using the pkg-kde symbolshelper it's relatively easy to maintain the
> symbols files for C++ libraries.
>
> http://pkg-kde.alioth.debian.org/symbolfiles.html
>
> C++ symbols files remain a bit problematic, so if you opt out of this
> please add a lintian override with a comment to explain this.
>
>
> Not specifically related to your Multi-Arch changes, when you run
> lintian with -I --show-overrides --pedantic there are 3 warning, 20 info
> and 1 pedantic tags.
>
> You can use a lintian hook for pbuilder/cowbuilder to run lintian after
> every build to help keep your package lintian clean. An example is
> included in the pbuilder package on disk at:
>  /usr/share/doc/pbuilder/examples/B90lintian
> or online at:
>  http://sources.debian.net/src/pbuilder/0.215/examples/B90lintian
>
> I'd like to walk you through the issues lintian reported after my build
> of your package. The lintian tags are pretty well documented, their
> extended information usually provides enough information to devise a fix.
>
>
> W: python-saga: postrm-has-useless-call-to-ldconfig
> W: python-saga: postinst-has-useless-call-to-ldconfig
>
> http://lintian.debian.org/tags/postrm-has-useless-call-to-ldconfig.html
> http://www.debian.org/doc/debian-policy/ch-sharedlibs.html#s-ldconfig
>
> Because the python-saga package contains .so libraries, debhelper adds
> the ldconfig call to the post{inst,rm} maintainer scripts even though
> they are not in one of the default library directories of the dynamic
> linker.
>
> You can override dh_makeshlibs to exclude the python-saga package:
>
> override_dh_makeshlibs:
>         dh_makeshlibs -Xpython-saga
>
>
> W: libsaga: package-name-doesnt-match-sonames libsaga-api-2.1.1
> libsaga-gdi-2.1.1
> I: libsaga: no-symbols-control-file
>
> http://lintian.debian.org/tags/package-name-doesnt-match-sonames.html
> http://www.debian.org/doc/debian-policy/ch-sharedlibs.html#s-sharedlibs-runtime
> http://www.netfort.gr.jp/~dancer/column/libpkg-guide/libpkg-guide.html
>
> You can split the libraries into their own binary package and more
> easily maintain symbols for each. If you opt out of the split, please
> add a lintian override with a comment to explain this.
>
>
> I: saga: hardening-no-fortify-functions
>
> http://lintian.debian.org/tags/hardening-no-fortify-functions.html
> https://wiki.debian.org/HardeningWalkthrough
>
> Debhelper 9 automatically enables the hardening buildflags, but not all
> build systems use CPPFLAGS besides CFLAGS, so you may need to add
> CPPFLAGS to CFLAGS in debian/rules:
>
> CFLAGS+=$(CPPFLAGS)
>
> lintian uses hardening-check to verify the hardening flags, and this
> check has some false-positives. If that's the case you can add a lintian
> override with the comment containing the hardening-check verbose output,
> e.g.:
>
> # -D_FORTIFY_SOURCE=2 is used during build, but hardening-check reports:
> # Fortify Source functions: yes (some protected functions found)
> #        unprotected: memset
> #        unprotected: memmove
> #        unprotected: wmemcpy
> #        unprotected: wcscpy
> #        unprotected: memcpy
> #        unprotected: fread
> #        protected: vfwprintf
> #        protected: vwprintf
> #        protected: sprintf
>
>
> I: saga source: duplicate-short-description libsaga libsaga-dev
> I: python-saga: capitalization-error-in-description python Python
>
> http://lintian.debian.org/tags/duplicate-short-description.html
> http://lintian.debian.org/tags/capitalization-error-in-description.html
> http://www.debian.org/doc/debian-policy/ch-binary.html#s-descriptions
>
> The short description for libsaga is easily fixed by using:
>
> Description: SAGA GIS shared libraries
>
> The capitalization error is also easily fixed:
>
> Description: SAGA GIS Python bindings
>  This package contains the Python bindings to SAGA.
>
>
> I: saga: spelling-error-in-binary
>
> http://lintian.debian.org/tags/spelling-error-in-binary.html
>
> Lintian loves spelling errors, add patches to fix the errors and send
> them upstream to get rid of them. You can reword the 'Allows to'
> sentences using "Allowing' instead of the suggested 'Allows one to' that
> usually makes more sense.
>
> Don't blindly sed -i the source files to fix the spellings errors, the
> result can be more awkward to read than leaving the typo. Rewriting the
> sentence is the way to go then.
>
>
> I: saga source: vcs-field-not-canonical
>
> http://lintian.debian.org/tags/vcs-field-not-canonical.html
>
> This used to be more relevant when Alioth was split between two machines
> with the SCM repos separated from the other hosting. Alioth was recently
> moved to a new more powerful machine after catastrophic hardware
> failure, so it's all hosted on the same system again.
>
> The proposed Debian GIS Policy documents the URLs to use, as does the
> package template:
>
> Vcs-Browser: http://anonscm.debian.org/gitweb/?p=pkg-grass/saga.git
> Vcs-Git: git://anonscm.debian.org/pkg-grass/saga.git
>
> http://anonscm.debian.org/gitweb/?p=pkg-grass/package_template.git;a=blob;f=debian/control
>
>
> I: saga: desktop-entry-lacks-keywords-entry
>
> http://lintian.debian.org/tags/desktop-entry-lacks-keywords-entry.html
>
> Add some keyword to assist searching for programs such as saga, things
> that come to mind:
>
> Keywords=gis;analysis;gui;spatial
>
>
> I hope you're not to overwhelmed by the amount of information, so take
> it easy and don't worry about fixing all the issues immediately, the low
> severity tags are not displayed by default for a reason :)
>
>
> Kind Regards,
>
> Bas
>
> --
> GnuPG: 0xE88D4AF1 (new) / 0x77A975AD (old)
>
>
> --
> To UNSUBSCRIBE, email to debian-gis-request@lists.debian.org
> with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
> Archive: [🔎] 52D89FF0.6020007@xs4all.nl">http://lists.debian.org/[🔎] 52D89FF0.6020007@xs4all.nl
>


Reply to: