Re: RFS: egroupware (fixes critical bug)
Hi Thomas, George, Laurent,
thank you for your replies. I'd like to comment on some of the questions
raised. First let me note, that I see two basic kinds of objections:
- Objections related to the packages themselves:
Those mainly arise from the fact, that Peter had invested *a lot* of
work into the packaging process, some of which I might not have
perfectly understood from the beginning. Thus for example I removed the
watch file - not knowing, that lintian will complain in --pedantic mode.
Also I acknowledged Peter's work, when I ignored most of the lintian
warnings originating from his packages. "If they were included before" -
so I thought - "they can't be that important". Partially I still believe
that if they weren't a big show stopper before, they shouldn't be now.
However I intend to fix them over the time of continuous maintenance.
- Objections related to the package's content: The egroupware suite
Here I disagree with those claiming egroupware was not suitable for
inclusion as such.
I am not a egroupware developer and I do not intent to become one. Also
I didn't take part in the conflict involving the discussion whether to
rewrite or continuously improve the codebase years ago.
In my opinion every project should be free to chose whether they prefer
a continuous development process or whether they take the steps of a
complete refactoring all at once. Also every developer should be free to
chose the project to work on and of course every user is free to chose
the project to use and work with. Therefore I don't think the egroupware
way of doing things should in any form conflict with the debian way of
doing things.
Of course I also talked to Ralf Becker, one of the egroupware
developers, who underlined the numerous changes to the codebase during
the last 4 years. You can find the projects subversion repository at
http://svn.egroupware.org/egroupware/ . If you'd like to get a quick
overview of the changes, have a look at the changelog at
http://www.egroupware.org/changelog .
In addition to those two main fields of objection, I think what really
qualifies egroupware for inclusion into debian could be condensed to the
following:
The project has many users. It has been included into debian for many
years. There's a new guy (me) taking care of the packages and getting
the bug fixes from upstream included.
If you have any further questions go ahead and ask. Also I'd be glad if
I could get some/more comments on my previous mail regarding the
technical issues of the packages itself, some of which I already fixed.
With kind regards,
Lars
Am 19.06.2010 14:15, schrieb Lars Volker:
> Hi Michal,
>
> thanks for your reply. I've added comments to the various lintian errors
> right below.
>
>
>>> The packages are based on previous work by Peter and were only
>>> changed to include the latest upstream version.
>
>> It would be good to fix some Lintian warnings as well:
>
>> P: egroupware-egw-pear source:
>> direct-changes-in-diff-but-no-patch-system
>> egw-pear/HTTP/WebDAV/Server.php and 5 more
> egroupware-egw-pear contains copies of pear modules, which are needed
> for egroupware to run. Some of those modules have been touched by Peter.
> However I don't think they will be updated to newer upstream versions
> anywhere soon. Therefore I added an override for this.
>
>> W: egroupware-egw-pear source: out-of-date-standards-version 3.8.2
>> (current is 3.8.4)
> Fixed.
>> W: egroupware source: out-of-date-standards-version 3.8.2 (current is
>> 3.8.4)
> Fixed.
>
>> I: egroupware-core:
>> possible-documentation-but-no-doc-base-registration
>> I: egroupware-etemplate:
> possible-documentation-but-no-doc-base-registration
> Those errors were already present, but I don't know how important they
> are. Do I have to fix them?
>
>> I: egroupware-core: unused-debconf-template egroupware/configuration/note
>> I: egroupware-core: unused-debconf-template
> egroupware/header/password/mismatch
> It is used in egroupware-core.config, however this is not detected by
> lintian. Should I add an override or is this a bug in lintian?
>
>> I: egroupware-addressbook: debian-news-entry-without-blank-line line
>> 6
> Fixed.
>
>> P: egroupware-*: no-upstream-changelog
> I've read http://lintian.debian.org/tags/no-upstream-changelog.html and
> noticed the sentence about multiple binary packages produced from a
> single source. However in the upstream tarball several Changelog files
> are located troughout the source tree, some of which have not been
> touched for several years. I'll try and ask the egroupware developers to
> add a central changelog to their released tarballs, which I can then
> link/copy or place in the main egroupware-core package.
>
> As I did not change anything here, I assume that those errors were
> present before. I'd like to address them in future. Please let me know
> if you think they are important enough to prevent the package from being
> accepted.
>
>> P: egroupware-felamimail: copyright-refers-to-symlink-license
>> usr/share/common-licenses/LGPL
> The file in question is from the horde project, which linked to the LGPL
> at http://www.fsf.org/copyleft/lgpl.html . I think this justifies
> symlinking to the latest version of LGPL, but of course I can change it
> to whatever seems more appropriate. I didn't change this as well.
>
>> W: egroupware-core: embedded-php-library W: egroupware-core:
>> embedded-javascript-library W: egroupware-phpsysinfo:
>> embedded-php-library W: egroupware-projectmanager:
>> embedded-php-library W: egroupware-projectmanager:
>> duplicate-font-file
> All errors above were already present in the debian packages built by
> Peter, so I assumed, it might be OK to leave them untouched. Especially
> as fixing them would inflict a lot of work and I suspect would also
> cause version conflicts between the current versions in debian and the
> versions needed by egroupware (except maybe for the fonts). Shall I add
> them to overrides then?
>
>
> I hope my comments clear out some of the concerns. Where possible I
> added a fix and uploaded new versions of the packages. Regarding all
> other problems, I'm open for any comment or discussion. Thanks for your
> help.
>
> With kind regards,
>
> Lars
Reply to: