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

Re: Cloudy package ready for review



Hi Roger,

thanks; now the package creation process finishes, and I have a few more
comments:

* The CPPFLAGS are not propagated. They are usually set to
  "-D_FORTIFY_SOURCE=2", which enables some hardening against buffer
  overflows and such.

* The description of the -data and -doc packages is too short (say
  lintian, but I agree here :-) )
  I would just copy the same description of the main package, and add a
  short paragraph "This package contains the platform independent
  data/documentation".

* cloudy should at least "Suggest" cloudy-doc

* You could add a "doc-base" for the documentation, see
  https://www.debian.org/doc/manuals/maint-guide/dother.en.html#doc-base
  This helps to register the documentation within Debian.

* You should not use the short name "custom" twice in debian/copyright;
  they need to be unique.

Everything else looks OK for me, so the only major remaining issue is
the copyright problem.

Cheers

Ole

Roger Wesson <rw@nebulousresearch.org> writes:
> Hi Ole,
>
> Ah, I had created my base.tgz inadvertently with the wrong distribution.
>  Now I corrected that, I reproduced the build failure and fixed it -
> turns out there's an upstream patch for this which moves a class
> declaration.  So I've incorporated that and successfully built the
> package, locally and also on launchpad for xenial and yakkety.  So,
> hopefully just the copyright issues remaining to be dealt with, for
> which I await information.
>
> Cheers,
> Roger
>
> On 11/06/16 09:16, Ole Streicher wrote:
>> Hi Roger,
>> 
>> Roger Wesson <rw@nebulousresearch.org> writes:
>>> I've simplified the install files.  As for the package not compiling, I
>>> forgot to mention before that I could not reproduce that. I built the
>>> package successfully on my own machine and on launchpad.
>> 
>> The problem here seems to be that you use Ubuntu 14.04 (trusty) for
>> both, which is quite old. If you try it on 16.04 (xenial) or the
>> upcoming 17.10 (yaakkety), you will observer the same failures:
>> 
>> * xenial:
>> https://launchpad.net/~olebole/+archive/ubuntu/astro-xenial/+build/9900284/+files/buildlog_ubuntu-xenial-amd64.cloudy_13.03-0.utest1_BUILDING.txt.gz
>> * yakkety
>> https://launchpad.net/~olebole/+archive/ubuntu/astro-yakkety/+build/9900444/+files/buildlog_ubuntu-yakkety-amd64.cloudy_13.03-0.utest1_BUILDING.txt.gz
>> 
>> Generally, it must be built under the latest Debian unstable/testing. To
>> ensure this, I would strongly recommend to install "pbuilder". This
>> allows to build the package in a clean, actual Debian unstable
>> environment independently of what Distribution you actually use.
>> 
>>> Looking at the log of the failed build, it seems that pbuilder did not
>>> try to resolve the dependencies and did not install libc etc for the
>>> build, which therefore failed.
>> 
>> The libc (and the C compiler) are already installed by default. There is
>> no need to select them manually.
>> 
>> The first error in the log file is
>> 
>> g++ -g -O2 -fstack-protector-strong -Wformat -Werror=format-security -ansi -O3 -
>> In file included from /usr/include/c++/5/valarray:90:0,
>>                  from cddefines.h:59,
>>                  from maincl.cpp:6:
>> /usr/include/c++/5/bits/valarray_before.h: In instantiation of 'struct std::__fu
>> /usr/include/c++/5/valarray:121:50:   required from 'struct std::valarray<molezo
>> /usr/include/c++/5/valarray:1073:96:   required from 'class std::valarray<molezo
>> mole.h:355:27:   required from here
>> /usr/include/c++/5/bits/valarray_before.h:350:12: error: invalid use of incomple
>>      struct __fun : __fun_with_valarray<_Tp>
>>             ^
>> which for me looks like some forward declaration, which older C++
>> compilers may have allowed but the current one doesn't (no real idea; I
>> am not a C++ expert). You could try to find out yourself how to fix that
>> (and submit to upstream), or discuss it with upstream whether he has a fix.
>> 
>> Best regards
>> 
>> Ole
>> 


Reply to: