diff mbox

git build from clean broken?

Message ID 530FB326.9080602@redhat.com
State New
Headers show

Commit Message

Paolo Bonzini Feb. 27, 2014, 9:50 p.m. UTC
Il 27/02/2014 20:57, Stefan Weil ha scritto:
> Yes. I saw that problem, too. This patch for Makefile fixed it for me:
> 
> index ccab967,a28a3c8..8ec3a99
> @@@ -170,6 -159,6 +170,7 @@@ qemu-options.def: $(SRC_PATH)/qemu-opti
>   SUBDIR_RULES=$(patsubst %,subdir-%, $(TARGET_DIRS))
>   SOFTMMU_SUBDIR_RULES=$(filter %-softmmu,$(SUBDIR_RULES))
> 
> ++$(SOFTMMU_SUBDIR_RULES): $(block-obj-y)
>   $(SOFTMMU_SUBDIR_RULES): config-all-devices.mak

This is correct.  Alternatively:

------------------ 8< --------------------
From: Paolo Bonzini <pbonzini@redhat.com>
Subject: [PATCH] build: build block-obj-y explicitly before recursing

block-obj-y is not anymore part of common-obj-y, because "nesting"
variables is complicated and requires specifying the correct
ordering in the calls to unnest-vars.  However, because of this
we need to specify block-obj-y in the dependencies of the
target subdirectories.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Comments

Peter Maydell Feb. 27, 2014, 10:17 p.m. UTC | #1
On 27 February 2014 21:50, Paolo Bonzini <pbonzini@redhat.com> wrote:
> Il 27/02/2014 20:57, Stefan Weil ha scritto:
>> Yes. I saw that problem, too. This patch for Makefile fixed it for me:
>>
>> index ccab967,a28a3c8..8ec3a99
>> @@@ -170,6 -159,6 +170,7 @@@ qemu-options.def: $(SRC_PATH)/qemu-opti
>>   SUBDIR_RULES=$(patsubst %,subdir-%, $(TARGET_DIRS))
>>   SOFTMMU_SUBDIR_RULES=$(filter %-softmmu,$(SUBDIR_RULES))
>>
>> ++$(SOFTMMU_SUBDIR_RULES): $(block-obj-y)
>>   $(SOFTMMU_SUBDIR_RULES): config-all-devices.mak
>
> This is correct.  Alternatively:

Can you either post that to the list as a proper patch or alternatively
just add a reviewed-by tag to the patch Stefan sent earlier, depending
which you think is the better fix? Then I'll apply it...

thanks
-- PMM
Paolo Bonzini Feb. 27, 2014, 10:23 p.m. UTC | #2
Il 27/02/2014 23:17, Peter Maydell ha scritto:
> On 27 February 2014 21:50, Paolo Bonzini <pbonzini@redhat.com> wrote:
>> Il 27/02/2014 20:57, Stefan Weil ha scritto:
>>> Yes. I saw that problem, too. This patch for Makefile fixed it for me:
>>>
>>> index ccab967,a28a3c8..8ec3a99
>>> @@@ -170,6 -159,6 +170,7 @@@ qemu-options.def: $(SRC_PATH)/qemu-opti
>>>   SUBDIR_RULES=$(patsubst %,subdir-%, $(TARGET_DIRS))
>>>   SOFTMMU_SUBDIR_RULES=$(filter %-softmmu,$(SUBDIR_RULES))
>>>
>>> ++$(SOFTMMU_SUBDIR_RULES): $(block-obj-y)
>>>   $(SOFTMMU_SUBDIR_RULES): config-all-devices.mak
>>
>> This is correct.  Alternatively:
>
> Can you either post that to the list as a proper patch or alternatively
> just add a reviewed-by tag to the patch Stefan sent earlier, depending
> which you think is the better fix? Then I'll apply it...

Done, thanks.

Can you apply also Fam's patch

http://article.gmane.org/gmane.comp.emulators.qemu/258469/raw

?

Paolo
Peter Maydell Feb. 27, 2014, 10:55 p.m. UTC | #3
On 27 February 2014 22:23, Paolo Bonzini <pbonzini@redhat.com> wrote:
> Can you apply also Fam's patch
>
> http://article.gmane.org/gmane.comp.emulators.qemu/258469/raw
>
> ?

Sure. I'd misread Markus' reply as being opposed to it, but
rereading I see he was just making a suggestion for the future.

thanks
-- PMM
Peter Maydell March 3, 2014, 11:50 p.m. UTC | #4
On 27 February 2014 22:17, Peter Maydell <peter.maydell@linaro.org> wrote:

> Can you either post that to the list as a proper patch or alternatively
> just add a reviewed-by tag to the patch Stefan sent earlier, depending
> which you think is the better fix? Then I'll apply it...

Do we have a definite version of this fix which everybody is happy
with yet? I've got confused with the various versions and review
comments on them. I'd like to apply the buildfix before I apply
any further pull requests...

thanks
-- PMM
Stefan Weil March 4, 2014, 6:57 a.m. UTC | #5
Am 04.03.2014 00:50, schrieb Peter Maydell:
> 
> Do we have a definite version of this fix which everybody is happy
> with yet? I've got confused with the various versions and review
> comments on them. I'd like to apply the buildfix before I apply
> any further pull requests...
> 
> thanks
> -- PMM

Simply take the version without confusing review comments:
http://patchwork.ozlabs.org/patch/324909/. :-)

To be honest: the other patch fixes one build regression and introduces
a new one, so it would not improve the situation.

Paolo noticed that Makefile is a mess in some parts, and I agree. But
that is something which can be handled separately.

Regards
Stefan
Paolo Bonzini March 4, 2014, 7:10 a.m. UTC | #6
Il 04/03/2014 07:57, Stefan Weil ha scritto:
> Simply take the version without confusing review comments:
> http://patchwork.ozlabs.org/patch/324909/. :-)
>
> To be honest: the other patch fixes one build regression and introduces
> a new one, so it would not improve the situation.
>
> Paolo noticed that Makefile is a mess in some parts, and I agree. But
> that is something which can be handled separately.

Agree.

Paolo
diff mbox

Patch

diff --git a/Makefile b/Makefile
index a443cd4..3938b6c 100644
--- a/Makefile
+++ b/Makefile
@@ -167,7 +167,7 @@  subdir-dtc:dtc/libfdt dtc/tests
 dtc/%:
 	mkdir -p $@
 
-$(SUBDIR_RULES): libqemuutil.a libqemustub.a $(common-obj-y)
+$(SUBDIR_RULES): libqemuutil.a libqemustub.a $(common-obj-y) $(block-obj-y)
 
 ROMSUBDIR_RULES=$(patsubst %,romsubdir-%, $(ROMS))
 romsubdir-%: