Re: [PATCH] debug/gdb: --disable-sim for CT_GDB_CROSS

[Khem Raj <raj dot khem at gmail dot com>]

On Tue, Nov 22, 2011 at 5:48 PM, Michael Hope <michael.hope@linaro.org> wrote:
> On Wed, Nov 23, 2011 at 6:21 AM, Yann E. MORIN
> <yann.morin.1998@anciens.enib.fr> wrote:
>> Zhenqiang, All,
>>
>> On Tuesday 22 November 2011 12:19:43 Zhenqiang Chen wrote:
>>> # HG changeset patch
>>> # User Zhenqiang Chen <zhenqiang.chen@linaro.org>
>>> # Date 1321604572 -28800
>>> # Node ID 2d6b11f65d7cfa0802e2db9a23363d04be44eecd
>>> # Parent Âa6c68712024f406df7d2da4af377f51982615833
>>> debug/gdb: --disable-sim for CT_GDB_CROSS.
>>>
>>> sim was already disabled for CT_GDB_NATIVE.
>>>
>>> Reviewed-by: Michael Hope
>>> Signed-off-by: Zhenqiang Chen <zhenqiang.chen@linaro.org>
>>>
>>> diff --git a/scripts/build/debug/300-gdb.sh b/scripts/build/debug/300-gdb.
>>> --- a/scripts/build/debug/300-gdb.sh
>>> +++ b/scripts/build/debug/300-gdb.sh
>>> @@ -146,6 +146,7 @@
>>> Â Â Â Â Â Â Â --with-build-sysroot="${CT_SYSROOT_DIR}" Â Â\
>>> Â Â Â Â Â Â Â --with-sysroot="${CT_SYSROOT_DIR}" Â Â Â Â Â\
>>>        --with-expat=yes              Â\
>>> +      Â--disable-sim                \
>>>        --disable-werror              Â\
>>> Â Â Â Â Â Â Â "${cross_extra_config[@]}"
>>
>> Your mailer is mangling every patch, by adding a leading space to every
>> line that start with a space, so the patches do not apply cleanly. I will
>> be fixing that locally, no need to resend.
>>
>> Otherwise, I agree with Mike: Why is it needed to disable 'sim' in the
>> cross-gdb? If the build is broken for some archs, it should rather be
>> fixed. At least, disabling/enabling 'sim' should be a config option,
>> so users that want to try and/or fix it have an easy way to.
>
> I'd like to disable sim as it isn't useful when building for a GLIBC
> target and it's one more thing to support. ÂARM don't want to
> distribute a binary simulator due to IP reasons which I must admit
> I've never understood :)
>
> How about a GDB_CROSS_EXTRA_CONFIG_ARRAY instead? ÂDisabling the
> simulator is rare enough that it shouldn't be an option on it's own.

disabling it for glibc and uclibc toolchains is fine however it should
be left enabled for bare metal toolchains

>
> -- Michael
>
> --
> For unsubscribe information see http://sourceware.org/lists.html#faq
>
>

--
For unsubscribe information see http://sourceware.org/lists.html#faq