DebForMeeGo

(Why switch to rpm: LSB)
(Why keep deb: Mer was never going to be used anyway! End users comment is entirely subjective not factual. Licencing was never an issue! Multiarch is not packaging format specific!)
Line 9: Line 9:
== Why keep deb ==
== Why keep deb ==
* No porting for Maemo packages needed
* No porting for Maemo packages needed
-
* Maemo has the bigger community
+
* Maemo has the bigger community [CITATION NEEDED]
-
* community infrastructure is based on deb - would be thrown away
+
* allows syncing from Debian/ Ubuntu
* allows syncing from Debian/ Ubuntu
-
* is used for distributions targetting end users
+
**Moblin promotes pushing back to upstream - Debian/Ubuntu rarely do! There is no need to be "synced" when you work with upstream. This "pro" deb point is actually a minus!
-
** packages contain lots of related fixes
+
* not to throw away experience gained with maemo which is reflected in the packaging
* not to throw away experience gained with maemo which is reflected in the packaging
-
* deb package can contain license information into the XBS-License field of the control file.
+
**These changes should have been pushed back upstream. Why is Maemo the only benefactor? The larger community should benefit from Maemo experience and then the package format wouldn't matter.
-
* will be the first packaging system to allow multiarch.
+
-
* seems to be already tested in multi-arch (even optimized versions per arch using multiple architectures), while RPM seems to be immature in that area (RPM expert's comment needed)
+
* tools for verifying (e.g. lintian) available
* tools for verifying (e.g. lintian) available

Revision as of 03:38, 19 February 2010

This page collects all arguments for and against keeping the *.deb package format in MeeGo.

If you want to support this Idea you can vote at Brainstorm.

Contents

Scope

the aim is at least to keep the *.deb package format in the MeeGo distributions aiming at Nokia Phones/ Handhelds. Although using the .deb format throughout the MeeGo project is desireable, not arguments here apply to the same extend looking at the whole project.

Why keep deb

  • No porting for Maemo packages needed
  • Maemo has the bigger community [CITATION NEEDED]
  • allows syncing from Debian/ Ubuntu
    • Moblin promotes pushing back to upstream - Debian/Ubuntu rarely do! There is no need to be "synced" when you work with upstream. This "pro" deb point is actually a minus!
  • not to throw away experience gained with maemo which is reflected in the packaging
    • These changes should have been pushed back upstream. Why is Maemo the only benefactor? The larger community should benefit from Maemo experience and then the package format wouldn't matter.
  • tools for verifying (e.g. lintian) available

Why switch to rpm

  • No porting for Moblin packages needed
  • RPM is LSB-compilant. (standartiztion issue with deb though)
  • Moblin build infrastructure has more capabilities already in place, such as imaging.
  • RPMs can contain license information too
  • rpmlint can be used to check for a package quality (like lintian)

Alternative Solution

  • Keep a Debian base system, and provide third-party applications in an LSB package formet (which is RPM inside). LSB-compliant RPM should install flawlessly, and that way the frontier between system packages (DEB) and third-party applications (RPM) is well defined.
  • Add the imaging feature to the Deb package.