On 16/11/2017 14:36, Joel Sherrill wrote:
> On Wed, Nov 15, 2017 at 4:35 PM, Chris Johns <[email protected]
> <mailto:[email protected]>> wrote:
>     I have been considering the default of building the RTEMS kernel when 
> building
>     the tools as a result of what you have been reporting so thank you for 
> taking
>     the time to do this.
> 
>     In theory it is a good idea however some archs have too many BSP and the 
> time
>     and resources it consumes makes this problematic. It may be simpler to 
> not build
>     the RTEMS kernel and to make sure the documentation and Quick Start in the
>     release details how to build the kernel. This way the tools get built and
>     installed as a step and then the kernel can be built as a second step. At 
> the
>     moment any failure means no tools and kernel and that is not great or user
>     friendly.
> 
> I think building RTEMS by default is a bad idea for a few reasons. 
> 
> + First, you are building everything which takes a long time and a lot of disk
> space.
> It is common to run out of disk space while building BSPs you don't care 
> about.
> I am teaching a class this week and at least one person ran during the RSB 
> build
> and another forgot --enable-rtemsbsp and ran out building RTEMS. It also took 
> a
> long time.
> 
> + Second, it may not be the configuration the end user wants even if the BSP
> they want is in the list.
> 
> + Finally, if they are going to develop a new or work on an existing custom
> BSP, then they are interested in building none of the existing BSPs in the
> tree.
> 
> I just don't see the need to build all the BSPs when you build the tools. 
>  

Can you please raise tickets on 4.11 and 5 for this?

Chris
_______________________________________________
users mailing list
[email protected]
http://lists.rtems.org/mailman/listinfo/users

Reply via email to