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

Bug#435888: A few problems



OK, I've taken a quick look at this package and there are a few
problems.

(The "Ubuntu" package mentioned in the RFP is a .deb but it is not
part of Ubuntu and there are serious errors in the packaging of
that .deb) :

1. Nearly all the files in the upstream tarball are executable which is
a PITA and requires a new tarball to be released with only essential
scripts set to executable. (Probably only ./configure.)

2. The package includes loads of XSL that appears to have already been
packaged for Debian and some of that has been modified compared to
upstream. There should be an option supported by ./configure to omit
these files from the 'make install' so that Debian does not have to
carry duplicate XSL files. xmlcopyeditor would then depend on
docbook-xsl and other similar packages that are maintained separately.
If any XSL files remain after these dependencies have been identified,
they can be packaged separately. This may require changes in the
upstream source to use /etc/catalog.xml to locate the correct
stylesheets. (See the datafreedom-qsfxsl package for an example.)
http://packages.debian.org/sid/datafreedom-qsfxsl

3. There are over TWENTY licences claimed within the source! Despite
that, NONE of the source files have any copyright information or any
licence information:
$ sudo apt-get install devscripts
$ licensecheck -r . | grep -c UNKNOWN
111
$ licensecheck -r . | grep -v UNKNOWN
./ltmain.sh: GPL (v2 or later) GENERATED FILE
./src/xmlcopyeditor_private.h: GENERATED FILE

I'm guessing that separating out the XSL from the C++ will mean that
the C++ is under a single licence but that needs to be made explicit,
preferably by specifying the typical copyright and licence header in
every .cpp and .h file:

/***************************************************************************
 *            filename.c
 *
 *  timestamp
 *  Copyright  year  name email
 ****************************************************************************/
/*
    This package is free software; you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
    the Free Software Foundation; either version 3 of the License, or
    (at your option) any later version.

    This program is distributed in the hope that it will be useful,
    but WITHOUT ANY WARRANTY; without even the implied warranty of
    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
    GNU General Public License for more details.

    You should have received a copy of the GNU General Public License
    along with this program.  If not, see <http://www.gnu.org/licenses/
>. */
(GPL3 in this example from one of my upstream C projects.)

4. The supported translations are packaged in a non-standard way -
there is no po/ directory, instead arbitrary directories are used to
package the prebuilt messages.mo files. This is a very bad idea as it
prevents translators within Debian from contributing to the translation
effort. The package should use gettext macros and a po/ directory and
package xmlcopyeditor.pot so that translations can be updated and
committed back upstream.

The upstream package is therefore not suitable for Debian at this time.

The final problem is that it requires wxwidgets2.8 when only 2.6 is
currently available in Debian unstable because of various issues - one
of which is the fragility of wxwidgets versions:

> We can't by any sensible measure just keep accruing wx releases in the
> distro.  There are apps still happily using 2.4 because an 'upgrade' to
> 2.6 will break them, though the demise of gtk1 is going to put an end
> to that period of bliss eventually whether they are working or not...
> 
> wx2.6 was rushed out prematurely, because book deadlines apparently
> trump good release management, 2.8 followed fairly shortly after for
> reasons which should be obvious, but not quickly enough to introduce
> another transition into the etch cycle.  Its the release we probably
> should have had in etch, but only very few people I know of had tested
> it with apps that were candidates for the distro, and though there were
> some apps that already would no longer work with 2.6 in their latest
> form, the needs of the many ultimately had to win the hand for etch.
> 
> I don't know of any app that missed inclusion totally from etch for
> this reason.
> 
> So the million dollar question now, is which version do we want to ship
> with Lenny.  2.8 will almost certainly be obsolete by then, and it looks
> like 2.9 is going to totally break everybody's string handling yet again.
> How big a nightmare that proves to be is probably also going to depend
> on how many shortcuts it takes to make an intractable solution meet
> another contractual deadline...
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=403237

So I don't see xmlcopyeditor as a viable candidate for Debian, sorry.
(I don't know if a new release could work with 2.6.)

-- 


Neil Williams
--------------
codehelp@debian.org
codehelpgpg@users.sourceforge.net
=================================
http://www.data-freedom.org/
http://www.nosoftwarepatents.com/
http://www.linux.codehelp.co.uk/

Attachment: pgperDfItL33G.pgp
Description: PGP signature


Reply to: