On Mon, 2008-02-18 at 16:20 -0800, Michael Morrell wrote: > Thanks for the response. It would certainly help if there were some > real documentation on bitbake.
You've evidently seen the manual and that is currently what we have. Obviously improvements to it are welcome. > I tried playing around with setting the undocumented "deps" and > "postdeps" flags (which is what "addtask" seems to modify as well as > _BBTASKS), but I couldn't get it to delete a task. Yes, removing a task isn't really supported. > Part of the problem stems from OpenEmbedded's stealing the base > bbclass. This bbclass is always inherited (you can't seem to avoid > that) and the version supplied by bitbake itself is fairly simple and > one would be hardpressed to find a reason not to inherit it, but the > one that comes with OpenEmbedded is much more complicatated and adds > several tasks that are not right for everyone. I guess my best > solution is to rename OpenEmbedded's base bbclass as something like > oebase.bbclass and let Bitbake's base.bbclass be the one used as the > true base. I'd accept a patch to bitbake which sets the name of the base class to use in a variable (defaulting to "base" for backwards compatibility). > By the way, I see recdeptask described in the manual, but can't find > any reference to it in the code. I think the manual is out of date and recdeptask was dropped since in practise if wasn't found to be useful and the code to implement it was painful. > I also see uses of a bbdepcmd flag for tasks in OpenEmbedded's > base.bbclass. What, if anything, does this do? It used to be the way certain task dependencies were made but its been replaced (I think with deptask) and doesn't do anything anymore. They were left for compatibility during the transition of OE from bitbake 1.6 to 1.8. I'll go through OE and delete those references though, thanks for reminding me. Regards, Richard _______________________________________________ Bitbake-dev mailing list [email protected] https://lists.berlios.de/mailman/listinfo/bitbake-dev
