Hi, On Fri, Oct 29, 2010 at 9:27 PM, Steven Rostedt <rostedt@xxxxxxxxxxx> wrote: > On Fri, 2010-10-29 at 20:26 -0400, Arnaud Lacombe wrote: >> Hi, > >> Technically, it is _not_ an environment variable; it is a kconfig >> symbol's name. In this particular case, both name collide and the >> kconfig symbol happens to get its default value from the environment. >> The real fixes, from my point of view, would be to have a perl binding >> of the kconfig backend and get rid of read_kconfig(). That said, your >> patch is broken if a symbol happen not to match an environment >> variable (both in name and value). > > That sounds like a proper fix. Unfortunately, I don't have the time to > work on this now, and although this fix works by coincidence, it > never-the-less works, as suppose to just having localmodconfig crash. > yes, sure. > But you are right. After KS and Plumbers, I'll work on doing something > like that. That can probably solve some of the other issues I'm having > (still keeping too many modules enabled). > I have some work in progress in this area: cleanup, better split between the kconfig backend and frontend, and eventually a perl binding. That would all be an interesting goal for .38 :-) Currently I got the parser working within perl, but I'm not really really sure of what I can currently do with it and how the whole thing interface to one another :) - Arnaud -- To unsubscribe from this list: send the line "unsubscribe linux-kbuild" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
