On Sun, Nov 06, 2005 at 09:05:49PM +0100, Jonas Smedegaard wrote: > > > > Martin's solution : instead of looking at > > > > only /lib/modules/<version> to see if modules where installed, > > > > actually do a find /lib/modules/<version> -name \*.ko, and search > > > > for real modules. > > > > > > As I understand it Martin actually agreed with Manoj in his latest > > > email ([EMAIL PROTECTED]). > > > > Yeah, but he was not aware that this breaks the plan for out-of-tree > > module packages, and i believe will be more bristle than what we have > > now. > > I have heard this mentioned several times now, but without any details. > Could you please point me to some summary of that plan? Or if missing > then post a draft to our wiki?
Well, i am not a wiki guy, so i will post here, and you can add it somewhere ? Basically, the plan is the folowing : each debian packaged module should build binary modules for each debian kernel flavour. non debian packaged modules should be encouraged to do the same, like some are trying to do with vmware for example. To build those, the module packages will build-depend on linux-headers-2.6.14, which in turn depends on all per-flavour headers for that arch. the /usr/src/linux-headers-2.6.14-1/flavour file bastian recently removed should contain the list of flavours to build, so module packages have everything needed to build a control file, and build the packages. Each modules should build packages using the KSRC=/lib/modules/<version>-<abi>-<flavour>/build, which is the default KSRC recomended by upstream and which every module out there use. > > > I also believe this to be the most sane approach. > > > > Ah, can you provide any kind of arguments for it ? I mean apart from > > the purely esthetical onesi, which are subjective and of dubious > > relevance here ? > > Uh, so opinions based on existing arguments do not count. > > Frans: Sorry, I was unaware that you requested only specific input. Bah, i am not sure. Friendly, Sven Luther -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
