On Tue, 2003-07-15 at 16:25, Gary Lawrence Murphy wrote: > >>>>> "B" == Buchan Milne <[EMAIL PROTECTED]> writes: > > B> Gary Lawrence Murphy wrote: > >> I'm hoping I missed something, but the apm-tools (apmd) tell me > >> the 9.1 distro kernel > > B> Of course, this is the wrong list for this question ... > > Even if we're building cooker with options incompatible with laptops? > > Maybe I'm mistaken. I thought the purpose of cooker was to ensure a > consistent and robust distribution through massively parallel > debugging under wide user circumstances -- unless there's good reason > to believe this subsystem has changed, aren't the lessons of > installing the just-prior release in unusual circumstance suitable > topics for discussion? > > For example, does the cooker list not care if the pcmcia_cs *.conf > files mistakenly assign prism2 wireless cards to hermes chipsets or > that the prism2_cs module, according to Google search results on > "mandrake prism2_cs" has _never_ been successfully configured? If you > ask me, we'll never get a distro 'perfect', but it still makes us look > bad if we include dysfunctional parts
To add an on-list reply to the off-list replies I've sent Gary: using orinoco_cs by default is, IMHO, correct. It *does* support prism2 cards (it works with mine) and it's compliant with the kernel wireless specification, which wlan-ng isn't. It's quite simple to use wlan-ng, I've done it to compare with orinoco_cs. You just need to be sure to have the prism2-tools package installed. > Is it inappropriate to the future release to ensure we distribute > packages compatible with the kernel we're including? In my case here, > if APM _is_ disabled in the kernel, that's ok, but then the apmd kit > being shipped with the cooker should be dropped because it's > incompatible: If people can recompile their own kernel to gain APM > support, you can be pretty certain they know how to download the apmd > software; if APM _is_ included in the kernel, then maybe this issue > points to an essential incompatibility with the kernel settings. > > The thing is, unless someone reports it, how will we know? Well, you've reported two essentially invalid bugs, which isn't helping anyone much :). APM isn't disabled except when you enable ACPI, and this isn't just Mandrake, it's standard kernel behaviour. Plus the default on installation is for ACPI to be disabled (and thus for APM to be enabled). You can switch between the two either by adding and removing acpi=off to lilo.conf or, I believe, through one of the drak* tools. Associating prism2 cards with orinoco_cs by default is correct behaviour given the qualities of orinoco_cs vs. wlan-ng, and besides, merely comes from the stock pcmcia-cs config file. -- adamw
