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

Re: [RFS] skim - smart common input method platform for KDE

Hash: SHA1

Hi Justin!

Justin Pryzby wrote:
> On Sat, Dec 03, 2005 at 04:09:22PM +0800, William J Beksi wrote:
>>I am looking for someone to sponsor skim. My package is on Debian
>>mentors (skim_1.4.3):
>>It is lintian clean but has the following linda warnings listed below.
>>Any comments would be greatly appreciated.
>>W: libskim0; The library libscim is not in a shlibs file.
>> The library shown above is not listed in a shlibs file. This means
>> that packages that depend on this one won't get ${shlibs:Depends}
>> correctly.
> This is a serious bug; see also #339056.  Since you provide libskim0
> and libskim-dev, it is gives the impression that you intend to
> provide public libraries for use by other programs.

I think this issue has been cleared up by Ming's response.

> Other comments:
> ./debian/control:
>   Section:
>   libs rather than utils, I think, if you are going to provide public
>   libs.  Maybe set the scim binary package to section utils.  AFAIK
>   binary packages inherit the section from their source package, by
>   default. 

skim is KDE frontend for scim, it provides a binary, I think it should
be in utils.

>   Build-Depends: libscim-dev
>   doesn't this mean that you have to bootstrap libscim-dev for all
>   architectures?  That is a a showstopper IMHO.  OTOH right now I
>   think your libscim-dev is Arch: all until you provide a static
>   library (see below).

Right, libscim-dev is provided by scim, I think this was also cleared up
 by Ming's response.

> ./debian/copyright:
>   Please include the years of copyright holding, and also the GPL
>   boilerplate *as found in the upstream source files*.


> ./debian/rules:
>   Please consider moving DH_COMPAT=4 to ./debian/compat; see the
>   rational at
>   <http://lists.debian.org/debian-mentors/2005/11/msg00247.html>.


>   This is being really picky, and maybe I don't understand, but:
>    	$(SCONS) -c && rm cache -fr 
>   doesn't make sense to me;  I'm guessing that scons -c creates
>   ./cache?  Why not $(SCONS) -c && rm cache?  This includes the
>   additional assertion that ./cache exists.  Assertive programming is
>   a Good Thing.  (Is cache a directory?  If so, then rm -r ./cache).

The cache directory is created after the first compilation by scons in
order to speed up successive compilations. Regardless of whether cache
exits or not, I want it removed the when the package is compiled. Since
I don't care to see the rm message about cache not being there the first
time I will keep the '-f'.

> +#	dh_install
>   Um, aren't you using the *.install files?  THen you need to
>   uncomment this..

I use dh_install on line 46 of ./debian/rules:
dh_install --sourcedir=debian/tmp --fail-missing

Maybe I should have removed the commented reference to dh_install on
line 58.

> ./debian/watch
>   Please consider using the qa.debian.org sourceforge redirect, to
>   increase the liklyhood that watchfiles actually work for more than
>   one person..


> ./debian/libskim-dev.install
>   Could you install a static archive, also?

I'm not sure which static archive you are referring to.
I have included ./debian/tmp/usr/lib/lib*.la in libskim-dev.install.

> I'm sorry, but I can't sponsor your package, yet..

Thank you for all of your helpful comments, new changes have been
uploaded to Debian mentors...still looking for a sponsor.

Does anyone have any comments on the linda warning about the binary
being linked with both version 5 and 6 of libstdc++?

Version: GnuPG v1.4.2 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org


Reply to: