In the mainboard selection, the selected mainboard is printed twice in certain cases, this patch eliminates the second mention.
This change might have unintended side-effects, but seems to work okay
Signed-off-by: Patrick Georgi patrick.georgi@coresystems.de
On Tue, Aug 25, 2009 at 10:56 AM, Patrick Georgipatrick@georgi-clan.de wrote:
In the mainboard selection, the selected mainboard is printed twice in certain cases, this patch eliminates the second mention.
This change might have unintended side-effects, but seems to work okay
I'd rather fix the offending boards. Isn't it board-specific? Maybe I haven't seen all the cases.
Myles
Am Dienstag, den 25.08.2009, 11:10 -0600 schrieb Myles Watson:
On Tue, Aug 25, 2009 at 10:56 AM, Patrick Georgipatrick@georgi-clan.de wrote:
In the mainboard selection, the selected mainboard is printed twice in certain cases, this patch eliminates the second mention.
This change might have unintended side-effects, but seems to work okay
I'd rather fix the offending boards. Isn't it board-specific? Maybe I haven't seen all the cases.
No, it's specific to the Kconfig source layout:
choice config { THE BOARDS } config { THE BOARDS' DEPENDENCIES } endchoice
To fix that, we'd have to do choice config { THE BOARDS } endchoice config { THE BOARDS' DEPENDENCIES }
which doesn't work with the source statement
Patrick
On Tue, Aug 25, 2009 at 11:37 AM, Patrick Georgipatrick@georgi-clan.de wrote:
Am Dienstag, den 25.08.2009, 11:10 -0600 schrieb Myles Watson:
On Tue, Aug 25, 2009 at 10:56 AM, Patrick Georgipatrick@georgi-clan.de wrote:
In the mainboard selection, the selected mainboard is printed twice in certain cases, this patch eliminates the second mention.
This change might have unintended side-effects, but seems to work okay
I'd rather fix the offending boards. Isn't it board-specific? Maybe I haven't seen all the cases.
No, it's specific to the Kconfig source layout:
choice config { THE BOARDS } config { THE BOARDS' DEPENDENCIES } endchoice
To fix that, we'd have to do choice config { THE BOARDS } endchoice config { THE BOARDS' DEPENDENCIES }
which doesn't work with the source statement
How about this fix?
Signed-off-by: Myles Watson mylesgw@gmail.com
Thanks, Myles
Am Dienstag, den 25.08.2009, 14:13 -0600 schrieb Myles Watson:
How about this fix?
I thought about that, but it has the disadvantage that it spreads the configuration of the board across two files.
I'm not sure what the best fix is...
Patrick
On Tue, Aug 25, 2009 at 2:30 PM, Patrick Georgipatrick@georgi-clan.de wrote:
Am Dienstag, den 25.08.2009, 14:13 -0600 schrieb Myles Watson:
How about this fix?
I thought about that, but it has the disadvantage that it spreads the configuration of the board across two files.
Yes.
I'd like to think that much of the second file will disappear as CONFIG_USE_DCACHE_RAM and others are moved to their final resting places. Then it might make sense to move the last few things like the mainboard strings into the mainboard/<vendor>/Kconfig file.
I'm not sure what the best fix is...
I guess the third option is live with it being printed twice.
Thanks, Myles
On Tue, Aug 25, 2009 at 02:34:22PM -0600, Myles Watson wrote:
On Tue, Aug 25, 2009 at 2:30 PM, Patrick Georgipatrick@georgi-clan.de wrote:
Am Dienstag, den 25.08.2009, 14:13 -0600 schrieb Myles Watson:
How about this fix?
I thought about that, but it has the disadvantage that it spreads the configuration of the board across two files.
Yes.
Indeed, and I think we should avoid that. The small proposed patch is perfectly fine, IMHO.
Acked-by: Uwe Hermann uwe@hermann-uwe.de
I'd like to think that much of the second file will disappear as CONFIG_USE_DCACHE_RAM and others are moved to their final resting places. Then it might make sense to move the last few things like the mainboard strings into the mainboard/<vendor>/Kconfig file.
I doubt this will always work, there are many per-board options that will still be required, and they should really be in the board dir, IMHO.
I'm not sure what the best fix is...
I guess the third option is live with it being printed twice.
Nah, why should we? The fix is easy.
Uwe.
Nah, why should we? The fix is easy.
The point of Kconfig is that it is standard. We had a custom config tool already.
Thanks, Myles
On Wed, Aug 26, 2009 at 09:07:22AM -0600, Myles Watson wrote:
Nah, why should we? The fix is easy.
The point of Kconfig is that it is standard. We had a custom config tool already.
Sure, I agree, but this small fix doesn't significantly change the tool's usage in any way. Besides we already have small changes against upstream's kconfig to adapt it to our needs anyway.
Uwe.