[2.5.70-bk-20030603] oldconfig always asking for machine type (x86)

[2.5.70-bk-20030603] oldconfig always asking for machine type (x86)

[Witold Krecicki]

On today snap of 2.5.70 from linux-2.5 bk tree while making make
oldconfig I always get:
Subarchitecture Type
> 1. PC-compatible (X86_PC)
  2. Voyager (NCR) (X86_VOYAGER)
  3. NUMAQ (IBM/Sequent) (X86_NUMAQ)
  4. SGI 320/540 (Visual Workstation) (X86_VISWS)
choice[1-4]:
Even if I'm doing this once after another, this a lil' bit complicates
creating binary packages with kernel, as building it should not wait for
any user input/interaction.
WK

Re: [2.5.70-bk-20030603] oldconfig always asking for machine type (x86)

[Roman Zippel]

Hi,

On 3 Jun 2003, Witold Krecicki wrote:

> choice[1-4]:
> Even if I'm doing this once after another, this a lil' bit complicates
> creating binary packages with kernel, as building it should not wait for
> any user input/interaction.

Could you send me your .config?
Thanks.

bye, Roman

Re: [2.5.70-bk-20030603] oldconfig always asking for machine type (x86)

[Roman Zippel]

Hi,

On 3 Jun 2003, Witold Krecicki wrote:

> Even if I'm doing this once after another, this a lil' bit complicates
> creating binary packages with kernel, as building it should not wait for
> any user input/interaction.

I think I found the problem, could you try the patch below?

bye, Roman

--- linux-2.5.70/scripts/kconfig/confdata.c	2003-05-28 22:47:05.000000000 +0200
+++ linux-2.5.70/scripts/kconfig/confdata.c	2003-06-04 02:13:38.000000000 +0200
@@ -243,7 +243,8 @@ int conf_read(const char *name)
 		prop = sym_get_choice_prop(sym);
 		sym->flags &= ~SYMBOL_NEW;
 		for (e = prop->expr; e; e = e->left.expr)
-			sym->flags |= e->right.sym->flags & SYMBOL_NEW;
+			if (e->right.sym->visible != no)
+				sym->flags |= e->right.sym->flags & SYMBOL_NEW;
 	}
 
 	sym_change_count = 1;

Re: [2.5.70-bk-20030603] oldconfig always asking for machine type (x86)

[Witold Krecicki]

W liście z ?ro, 04-06-2003, godz. 02:20, Roman Zippel pisze: 
> Hi,
> 
> On 3 Jun 2003, Witold Krecicki wrote:
> 
> > Even if I'm doing this once after another, this a lil' bit complicates
> > creating binary packages with kernel, as building it should not wait for
> > any user input/interaction.
> 
> I think I found the problem, could you try the patch below?
OK, it's working now :) 
WK