On 21 January 2017 at 08:45, Alessandro Di Federico
<ale+q...@clearmind.me> wrote:
> A quite large part of {linux,bsd}-user/qemu.h is shared, this patch
> introduces qemu-user-common.h which factors it out. This shared part is
> also the bare minimum required to build a linux-user-like target, and,
> in particular, it will be useful for the libtcg targets.
> ---
> bsd-user/qemu.h | 193
> +----------------------------------------------------
> linux-user/qemu.h | 176 +-----------------------------------------------
> qemu-user-common.h | 180 +++++++++++++++++++++++++++++++++++++++++++++++++
> 3 files changed, 182 insertions(+), 367 deletions(-)
> create mode 100644 qemu-user-common.h
So my issue with this is that currently linux-user is pretty
well maintained and tested, but bsd-user is basically unmaintained
and not even compile tested. So it's convenient to have bsd-user
code be basically its own distinct standalone thing, because
then when we fix linux-user we don't have to care too much
about whether bsd-user might be accidentally broken by our
changes.
If we had more active testing of bsd-user I'd be a bit keener
about removing some of the code duplication.
thanks
-- PMM
