Re: xen-unstable stubdom build-failure when debug=n

[Sander Eikelenboom <linux@eikelenboom.it>]

Monday, July 28, 2014, 11:09:35 AM, you wrote:

> On Sat, 2014-07-26 at 17:14 +0200, Sander Eikelenboom wrote:
>> Just tested some more with a strange outcome:
>> 
>> When doing:
>> ~# git checkout  RELEASE-4.4.0
>> ~# make mrproper && ./configure && make debug=n
>> 
>> It also fails with the error in dhcp i mentioned before.
>> 
>> However if i do:
>> ~# git checkout  RELEASE-4.4.0
>> ~# make mrproper && ./configure && make
>> 
>> It builds fine .. (and release-4.4.0 should have the implicit debug=n)
>> So what is different when explicitly giving debug=n ?

> The only thing I can thing of would be that specifying debug= on the
> make command line makes it enter the environment or MAKEFLAGS or
> something, whereas setting it in the Makefile or .config etc doesn't.
> (Perhaps e.g. "export debug" in Config.mk would make these two behave
> the same?)

> Anyway, that having happened the debug could then filter down and get
> picked up by the configure or build system of some sub-component in a
> way we aren't expecting.

> That's all 100% speculation though.

Well just setting it only in the Config.mk works fine.

Could be due to the mixing, i was used to not changing my Config.mk and just 
specifying "make debug=(y|n)" on the commandline.
So Config.mk had debug=y, while the make had debug=n.
However some googling didn't show much if that was even ever supported,
although it used to work fine ;-)

Should there just be a ./configure --debug=y ?

> Ian.