* Frans Pop <[EMAIL PROTECTED]> [2008-07-10 18:06]: > 1) Patches are normally submitted through the BTS. Discussion/questions > and maybe review of rough, proposed patches beforehand is done on the > mailing list.
The patch was discussed off-list (not ideal, but certainly better than how most patches in d-i are handled, i.e. committed to SVN directly without any comments or review at all). > 2) We don't just add module udebs without also seeing something of > the code that will use it (or at the very least a discussion of that > code). Per will post more patches soon. (Although it seems the jffs2 patch is obsolete because of changes in the plans how to handle Kurobox Pro support). > 3) Module udebs are normally defined in kernel-wedge and then used > in packages for specific architectures. That's what Per did originally, but I recommended against it because it's only needed on a single sub-arch. I said that it could be added to kernel-wedge once other arches actually need it. YMMV. > 4) Normally at least some effort is made to see if a change is also > usable for other architectures (and if not, some explanation is > given). It's not. Per didn't include a changelog to make this obvious but it would have been something like "Add jffs2 modules on Orion because oldsys-preseed needs to access a jffs2 fs on the Kurobox". However, as I indicated above, this seems no longer to be true, so this patch is withdrawn for now. Per will post other patches for review soon. -- Martin Michlmayr http://www.cyrius.com/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
