Comment 30 for bug 219630

Martin-Eric,

the geode debdiff in comment 14 (http://launchpadlibrarian.net/15171834/debdiff_xf86-video-geode.txt) looks ok, and I already ack'ed this. Your new debdiff in comment 27 is huge again, and again changes a lot of unrelated things (build system, etc.) and also has an inappropriate changelog delta. Again: For SRUs, there should be *one* changelog entry with an appropriate description (readable by mere-mortal computer users) with an LP: #xxxxxx bug reference, and a minimal patch. So is the previous debdiff still valid?

nsc: still huge. Getting the new upstream version is ok for me (since it's a sharply focused device driver), and thus the autotools/libtool noise is ok, but changing all the packaging underneath it seems dangerous to me (debian/xsfbs/*.mk). How do these changes influence the build and the generated .debs?

To finally get this SRU done, I really advise you to prepare a proper SRU for hardy, instead of just keep preparing intrepid/unstable backports. Take the new upstream version, use the hardy packaging, and just apply the patches which are necessary, this will make it much easier to review, and much less prone to regressions.

Thank you!