[Qgis-psc] A plea for backporting

Markus Neteler neteler at osgeo.org
Fri Feb 11 13:59:09 PST 2011


Fellow PSC members,

as a long term QGIS user and an even longer FOSS4G activist I would like
to bring the following wish to your kind attention: please backport bugfixes.

Unlike other FOSS projects (GDAL, GRASS, ...) the QGIS project does not
offer backports of even important bugfixes to the last stable release
branch(es).

Looking at the changelog at http://trac.osgeo.org/qgis/browser/branches :

Branch    rev   age     Last Change
 Release-1_3_0  11661 	17 months 	macho: translation update: uk by sergey
 Release-1_4_0  12730 	13 months 	timlinux: Removed duplicate entry in
index.html and regenerated application default …
 Release-1_5_0  14255 	5 months 	macho: correct some typos in fr by jean roc
 Release-1_6_0 14615 	3 months 	kyngchaos: fix release name

... we observe that branches get essentially abandoned once a release is done.
Looking at http://www.qgis.org/wiki/Release_Roadmap I understand that QGIS 1.4
should be updated. But the last change was done more than one year ago.

Comparing that to GDAL:
Branch    rev   age     Last Change
 1.6	21472 	4 weeks 	warmerdam: Fixed order of arguments to VSIRename.
 1.7	21609 	12 days 	rouault: NITF: correctly assign hemisphere for a
ICORDS='U' NITF file with …

or GRASS:
Branch    rev   age     Last Change
 releasebranch_6_3	45008 	4 weeks 	neteler: formula correction
 releasebranch_6_4	45380 	22 minutes 	msieczka: Set svn:mime-type to
'text/x-plain; charset=utf-8 to fix UTF-8 rendering …

we see that here even the last but one stable release branch received bugfixes.

So, why are backports of (important) bugfixes needed?

While I support the "release-often" paradigm, this does not exclude
backports to be
done. Sure, it is a pain for the developers but it is an excellent
service for long term
users. Example: I am running these three software packages on several machines,
some of them are used by many users. Naturally I don't want to break things and
risk that they flood me with questions "why is all different" [as it
may happen in
a new release]. Moreover I prefer to merge in bugfixes and that's it.
Easily done
for GRASS and GDAL. Maybe once a year I update to the next stable release.
This concept works well for me for many years.

Imagine the IT folks of a public administration. For sure they will
not change the
installation every 6 month. But exchanging a DLL or whatever file, maybe yes.
Today, at the Trento GFOSS conference we saw presentations with QGIS 1.4
in action - of course still used.

Please consider to backport at least important fixes.
Your users will be grateful.

Kind regards,
Markus Neteler

-- 
Markus Neteler, PhD
Fondazione Edmund Mach (FEM) - Research and Innovation Centre
Department of Biodiversity and Molecular Ecology
Head of GIS and Remote Sensing Unit
Via E. Mach, 1 - 38010 S. Michele all'Adige (TN), Italy
Web:   http://gis.cri.fmach.it  -   http://grass.osgeo.org



More information about the Qgis-psc mailing list