I've found the problem. It's from your patch commited in December. http://lists.gnu.org/archive/html/commit-grub/2006-12/msg00002.html http://cvs.savannah.gnu.org/viewcvs/grub2/normal/main.c?cvsroot=grub&r1=1.18&r2=1.19
The code block from main.c in function read_config_file() is where things start: if (nested) grub_env_set_data_slot ("menu", newmenu); else newmenu = grub_env_get_data_slot ("menu"); The path taken is the "else" path. So it's doing grub_env_get_data_slot ("menu") .. well the problem is "menu" is never set anywhere. The calling order is: grub_normal_execute() grub_readconfig_file() grub_env_get_data_slot() Though I'm very confused of how to exactly fix this with the logic in the patch. On 2/17/07, Jerone Young <[EMAIL PROTECTED]> wrote:
I will try it out and get some output from it. On 2/12/07, Marco Gerards <[EMAIL PROTECTED]> wrote: > "Jerone Young" <[EMAIL PROTECTED]> writes: > > > Trying to get grub to read from grub.cfg file with the current code in > > CVS appears to be busted. Release 1.95 works fine and can read the > > grub.cfg file without a problem automatically. But with code in CVS it > > nolonger reads it automatically (meaning no nice bootup menu), and if > > you do read it in via the "configfile" command you get to the menu, > > BUT it will not load the kernel from the configuration (this config > > file works fine with 1.95 release). If you try and load from the menu > > you get "error: You need to load the kernel first.". Any ideas what > > may be causing this in CVS, is there active development on some pieces > > not yet done ? > > It is very likely my latest commit broke some things. I did change > code related to all this. > > Can you reproduce this somehow using grub-emu? > > -- > Marco > > > > _______________________________________________ > Grub-devel mailing list > Grub-devel@gnu.org > http://lists.gnu.org/mailman/listinfo/grub-devel >
_______________________________________________ Grub-devel mailing list Grub-devel@gnu.org http://lists.gnu.org/mailman/listinfo/grub-devel