The dependency on curses has been removed in commit c7a856b42e403e2b
("target/unicore32: Prefer qemu_semihosting_log_out() over curses").
So we can remove the related lines in the Makefile now, too.

Signed-off-by: Thomas Huth <thuth@redhat.com>
---
 target/unicore32/Makefile.objs | 4 ----
 1 file changed, 4 deletions(-)

diff --git a/target/unicore32/Makefile.objs b/target/unicore32/Makefile.objs
index 35d8bf530d..6b41b1e9ef 100644
--- a/target/unicore32/Makefile.objs
+++ b/target/unicore32/Makefile.objs
@@ -2,7 +2,3 @@ obj-y += translate.o op_helper.o helper.o cpu.o
 obj-y += ucf64_helper.o
 
 obj-$(CONFIG_SOFTMMU) += softmmu.o
-
-# Huh? Uses curses directly instead of using ui/console.h interfaces ...
-helper.o-cflags := $(CURSES_CFLAGS)
-helper.o-libs := $(CURSES_LIBS)
-- 
2.18.1

On Thu, 23 Jul 2020 at 14:23, Thomas Huth <thuth@redhat.com> wrote:
>
> The dependency on curses has been removed in commit c7a856b42e403e2b
> ("target/unicore32: Prefer qemu_semihosting_log_out() over curses").
> So we can remove the related lines in the Makefile now, too.
>
> Signed-off-by: Thomas Huth <thuth@redhat.com>

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>

Another, unrelated, CONFIG_CURSES oddity:
hw/input/virtio-input-hid.c has an "#undef CONFIG_CURSES"
before it includes ui/console.h. But that header file
doesn't do anything different if CONFIG_CURSES is defined,
so I think the undef is now useless and can be removed.

thanks
-- PMM

On 23/07/2020 16.50, Peter Maydell wrote:
> On Thu, 23 Jul 2020 at 14:23, Thomas Huth <thuth@redhat.com> wrote:
>>
>> The dependency on curses has been removed in commit c7a856b42e403e2b
>> ("target/unicore32: Prefer qemu_semihosting_log_out() over curses").
>> So we can remove the related lines in the Makefile now, too.
>>
>> Signed-off-by: Thomas Huth <thuth@redhat.com>
> 
> Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
> 
> Another, unrelated, CONFIG_CURSES oddity:
> hw/input/virtio-input-hid.c has an "#undef CONFIG_CURSES"
> before it includes ui/console.h. But that header file
> doesn't do anything different if CONFIG_CURSES is defined,
> so I think the undef is now useless and can be removed.

There used to be some curses stuff in that header, but it has been
removed more than two years ago (see commit b0766612d16da18). So yes, I
think you're right, that #undef CONFIG_CURSES can be removed now. Care
to send a patch?

 Thomas

That's OK for unicore32 target.

Acked-by: Guan Xuetao <gxt@pku.edu.cn>


> -----原始邮件-----
> 发件人: "Thomas Huth" <thuth@redhat.com>
> 发送时间: 2020-07-23 21:22:19 (星期四)
> 收件人: "Guan Xuetao" <gxt@mprc.pku.edu.cn>, qemu-devel@nongnu.org
> 抄送: "Philippe Mathieu-Daudé" <f4bug@amsat.org>, qemu-trivial@nongnu.org, "Richard Henderson" <richard.henderson@linaro.org>
> 主题: [PATCH] target/unicore32: Remove CURSES stuff from the Makefile.objs
> 
> The dependency on curses has been removed in commit c7a856b42e403e2b
> ("target/unicore32: Prefer qemu_semihosting_log_out() over curses").
> So we can remove the related lines in the Makefile now, too.
> 
> Signed-off-by: Thomas Huth <thuth@redhat.com>
> ---
>  target/unicore32/Makefile.objs | 4 ----
>  1 file changed, 4 deletions(-)
> 
> diff --git a/target/unicore32/Makefile.objs b/target/unicore32/Makefile.objs
> index 35d8bf530d..6b41b1e9ef 100644
> --- a/target/unicore32/Makefile.objs
> +++ b/target/unicore32/Makefile.objs
> @@ -2,7 +2,3 @@ obj-y += translate.o op_helper.o helper.o cpu.o
>  obj-y += ucf64_helper.o
>  
>  obj-$(CONFIG_SOFTMMU) += softmmu.o
> -
> -# Huh? Uses curses directly instead of using ui/console.h interfaces ...
> -helper.o-cflags := $(CURSES_CFLAGS)
> -helper.o-libs := $(CURSES_LIBS)
> -- 
> 2.18.1

On 7/23/20 3:22 PM, Thomas Huth wrote:
> The dependency on curses has been removed in commit c7a856b42e403e2b
> ("target/unicore32: Prefer qemu_semihosting_log_out() over curses").
> So we can remove the related lines in the Makefile now, too.

Don't we also need to remove all uses of CONFIG_CURSES to
be complete? Eventually display a warning with --(en|dis)able-curses
is used.

> 
> Signed-off-by: Thomas Huth <thuth@redhat.com>
> ---
>  target/unicore32/Makefile.objs | 4 ----
>  1 file changed, 4 deletions(-)
> 
> diff --git a/target/unicore32/Makefile.objs b/target/unicore32/Makefile.objs
> index 35d8bf530d..6b41b1e9ef 100644
> --- a/target/unicore32/Makefile.objs
> +++ b/target/unicore32/Makefile.objs
> @@ -2,7 +2,3 @@ obj-y += translate.o op_helper.o helper.o cpu.o
>  obj-y += ucf64_helper.o
>  
>  obj-$(CONFIG_SOFTMMU) += softmmu.o
> -
> -# Huh? Uses curses directly instead of using ui/console.h interfaces ...
> -helper.o-cflags := $(CURSES_CFLAGS)
> -helper.o-libs := $(CURSES_LIBS)
> 

On 23/07/2020 15.42, Philippe Mathieu-Daudé wrote:
> On 7/23/20 3:22 PM, Thomas Huth wrote:
>> The dependency on curses has been removed in commit c7a856b42e403e2b
>> ("target/unicore32: Prefer qemu_semihosting_log_out() over curses").
>> So we can remove the related lines in the Makefile now, too.
> 
> Don't we also need to remove all uses of CONFIG_CURSES to
> be complete? Eventually display a warning with --(en|dis)able-curses
> is used.

Which ones do you mean? I can't find any other CONFIG_CURSES anymore in
the target/ folder...

Or do you also plan to get rid of ui/curses.c ??? ... that's certainly
another story.

 Thomas

On 7/23/20 3:45 PM, Thomas Huth wrote:
> On 23/07/2020 15.42, Philippe Mathieu-Daudé wrote:
>> On 7/23/20 3:22 PM, Thomas Huth wrote:
>>> The dependency on curses has been removed in commit c7a856b42e403e2b
>>> ("target/unicore32: Prefer qemu_semihosting_log_out() over curses").
>>> So we can remove the related lines in the Makefile now, too.
>>
>> Don't we also need to remove all uses of CONFIG_CURSES to
>> be complete? Eventually display a warning with --(en|dis)able-curses
>> is used.
> 
> Which ones do you mean? I can't find any other CONFIG_CURSES anymore in
> the target/ folder...
> 
> Or do you also plan to get rid of ui/curses.c ??? ... that's certainly
> another story.

Ah indeed.

Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>